<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Blake Vanderhyde on Medium]]></title>
        <description><![CDATA[Stories by Blake Vanderhyde on Medium]]></description>
        <link>https://medium.com/@blakevanderhyde?source=rss-53e34460e4c0------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*WZcrRAtbHBF98Kco7AQ85w.jpeg</url>
            <title>Stories by Blake Vanderhyde on Medium</title>
            <link>https://medium.com/@blakevanderhyde?source=rss-53e34460e4c0------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Fri, 20 Nov 2026 07:29:22 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@blakevanderhyde/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[AI for Lawyers: Separating Signal from Noise]]></title>
            <link>https://medium.com/@blakevanderhyde/ai-for-lawyers-separating-signal-from-noise-585800666995?source=rss-53e34460e4c0------2</link>
            <guid isPermaLink="false">https://medium.com/p/585800666995</guid>
            <category><![CDATA[law]]></category>
            <category><![CDATA[legaltech]]></category>
            <category><![CDATA[law-firm-marketing]]></category>
            <category><![CDATA[legal-ethics]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <dc:creator><![CDATA[Blake Vanderhyde]]></dc:creator>
            <pubDate>Sat, 05 Sep 2026 15:06:04 GMT</pubDate>
            <atom:updated>2026-10-20T19:00:47.496Z</atom:updated>
            <content:encoded><![CDATA[<h4><em>A plain-language guide for the legal profession.</em></h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/796/1*MTHFwt0MGmgSlEWg66KXbg.png" /></figure><p>The material below was originally published in 2026 as a fourteen-part series on LinkedIn. It has been collected and edited into a single volume so it can be read straight through, shared inside a firm, and returned to when a question comes up.</p><p>The goal is simple: to be useful.</p><h3>About the Author</h3><p>Blake Vanderhyde is a plaintiff attorney with twenty-seven years in practice and twenty years in legal marketing. He has worked with some of the largest plaintiff firms in the country on case acquisition for many of the most significant mass tort and multidistrict litigations in U.S. history, with an emphasis on pharmaceutical and medical device injury claims and toxic exposure cases.</p><p>From 2020 to 2025 he worked alongside former U.S. Senator Rudy Boschwitz on a public health effort to address antimicrobial resistance, work that brought him into meetings with federal regulators, members of Congress, and Fortune 500 healthcare executives. He documented the journey in a twenty-seven part public investation of the crisis. It was during that work that he began using artificial intelligence seriously, every day, on problems that mattered.</p><p>He is the co-founder of Ardeo, an AI marketing platform built for plaintiff firms operated by A4A, Inc. He lives in Minnesota.</p><h3>Contents</h3><ul><li>Chapter 1: Introduction</li><li>Chapter 2: Why So Many Lawyers Are Hesitant About AI</li><li>Chapter 3: Why AI Hallucinates Legal Citations</li><li>Chapter 4: Why the AI Conversation in Law Needs Better Questions</li><li>Chapter 5: A Better Framework for Thinking About AI in a Law Firm</li><li>Chapter 6: You Do Not Need to Be an AI Expert, But You Do Need Strategic Understanding</li><li>Chapter 7: Confidentiality, Privilege, and Professional Judgment in the Age of AI</li><li>Chapter 8: Assessing the AI Vendor Landscape</li><li>Chapter 9: Where AI Can Actually Help a Law Firm Today</li><li>Chapter 10: How Law Firms Can Evaluate AI Tools Without Getting Lost in the Noise</li><li>Chapter 11: Why Marketing May Be One of the Most Overlooked Functions in the AI Conversation</li><li>Chapter 12: AI-Generated Synthetic Performers in Legal Advertising</li><li>Chapter 13: What Good Judgment Looks Like in an AI-Enabled Law Firm</li><li>Chapter 14: Final Thoughts: The Profession Needs You in This Conversation</li></ul><h3>Chapter 1: Introduction</h3><p>I recently finished publishing a 27-part series about antimicrobial resistance (AMR), a project I was deeply involved in for many years. If you followed that series, thank you. If you did not and would like to read it, you can find it here on Medium.</p><p>What many readers may not have fully seen during that journey was that the project changed the way I think and work.</p><p>To do that work well, I had to move far beyond my own training, experience, and comfort zone. I found myself working deeply with artificial intelligence, not casually, but as part of the real work. Quite frankly, I did not have much of a choice.</p><p>I used AI to help think through regulatory pathways, financial models, clinical trial design, market analysis, business strategy, and large bodies of scientific research.</p><p>In a very practical sense, it became a force multiplier. It saved me time and money. It expanded my ability to test ideas, run models, and pressure-test strategies at a level I could not have reached alone.</p><p>And as I worked with these tools more seriously, I began to see implications for my primary profession: the law.</p><p>I started asking myself how these tools might be used to better serve clients and help the legal industry as a whole.</p><p>So I called one of my oldest and closest friends, a technical builder with deep engineering skills, and we began building something new for lawyers.</p><p>That is not what this guide is about. I am not here to pitch a product.</p><p>I am here to be helpful. To apply what I have learned and try to bring clarity to a profession being hit by AI from every direction at once.</p><p>Right now, many lawyers are experiencing AI as noise.</p><p>New tools every week.</p><p>New jargon with very little explanation.</p><p>Vendors making promises they cannot keep.</p><p>Bar associations issuing guidance in real time.</p><p>Senior lawyers skeptical while younger lawyers are already using these tools daily.</p><p>For busy attorneys, the rational response is often to tune it out.</p><p>I understand that response. In many ways, it is a reasonable one.</p><p>Lawyers are right to worry about hallucinated cases, confidentiality, shifting ethical boundaries, and whether these tools can be trusted in serious professional settings. They are also right to be wary of hype and vague claims.</p><p>But I do not think tuning it out will serve lawyers or, more importantly, their clients.</p><p>So I want to offer a different approach.</p><p>Not hype. Not fear. Not exaggerated claims of expertise. Just a serious effort to think clearly, speak plainly, and separate signal from noise.</p><p>I do not have all the answers. Nobody does. This technology is moving too quickly, changing too constantly, and touching too many domains for any one person to fully master.</p><p>But I have spent significant time working with it, building, testing, making mistakes, and learning where it helps and where it does not.</p><p>In the chapters that follow, I share what I think lawyers should pay attention to, and where I believe caution is warranted.</p><p>My goal is simple: to be useful.</p><h3>Chapter 2: Why So Many Lawyers Are Hesitant About AI</h3><p>Much of the legal profession’s hesitation around artificial intelligence is not irrational resistance. It is a rational response to unclear rules, real professional risk, and a market full of noise.</p><p>That matters, because a great deal of commentary about AI in law fails to appreciate how legitimate many of these concerns really are.</p><p>Lawyers are trained to identify risk, protect confidentiality, work within ethical rules, and exercise judgment under pressure.</p><p>When new technology arrives surrounded by hype, conflicting claims, and rapidly changing guidance, caution is not a defect.</p><p>It is a sign of professional seriousness.</p><p>I understand that instinct. In many respects, it is the right one.</p><p>In my view, the hesitation many lawyers feel toward AI tends to fall into four broad categories: professional risk, economic resistance, cultural resistance, and cognitive overload.</p><ol><li>Professional risk</li></ol><p>This is the most obvious concern, and in many ways the most important.</p><p>Lawyers have seen peers sanctioned for filing briefs with hallucinated citations. Fabricated cases and overconfident outputs are not minor errors in this profession.</p><p>They can be career-threatening.</p><p>That fear is rational.</p><p>There are also ethical concerns.</p><p>Bar associations are still developing guidance. Rules vary by jurisdiction, and the boundaries are still evolving.</p><p>Then there is confidentiality.</p><p>Lawyers handle sensitive client information, litigation strategy, proprietary business material, health records, and privileged communications.</p><p>Questions about what can be entered into an AI system, how that data is processed, and whether privilege could be compromised are not signs of technophobia. They are part of the job.</p><ol><li>Economic resistance</li></ol><p>The second layer is economic.</p><p>For firms built on the billable hour, AI raises a direct question: if work becomes faster, what happens to the business model?</p><p>Even when firms see the upside, they still have to reconcile it with pricing, training, and client expectations.</p><p>For contingency fee firms, the economic concern looks different.</p><p>They are asking whether it will improve case acquisition, client communication, and operational efficiency without creating new risk.</p><p>They want to know whether it will help them sign better cases, reduce overhead, and improve execution, or whether it is simply another layer of vendor hype.</p><p>That hesitation is rational too.</p><p>There is also a practical reality.</p><p>Many lawyers do not see technology implementation or marketing systems as their role. They delegate those functions.</p><p>But even when execution is outsourced, strategic understanding still matters.</p><ol><li>Cultural resistance</li></ol><p>The third form of resistance is cultural.</p><p>There is often a generational divide inside firms. Younger lawyers may adopt AI tools more quickly, but they rarely control budgets or policy. Senior lawyers usually have that authority, and many formed their professional instincts in a very different technological environment.</p><p>Some remain unconvinced that AI will have lasting importance in legal practice. They may have tried an early version of ChatGPT in 2023, found it underwhelming, and moved on.</p><p>The difficulty is that they are making a long-term judgment based on an outdated snapshot of a rapidly changing technology.</p><p>There is also fatigue with hype.</p><p>Lawyers are constantly pitched by vendors who promise disruption but do not understand legal practice.</p><p>That does not build trust. It erodes it.</p><ol><li>Cognitive overload</li></ol><p>The final category is cognitive overload.</p><p>Most lawyers do not have the time to track this space closely. They are serving clients, trying cases, and running a business.</p><p>They do not have spare hours each week to study new AI models, token limits, and shifting terminology.</p><p>So they do what many professionals do when faced with constant and accelerating complexity.</p><p>They tune it out.</p><p>That response is understandable. But it creates a problem, because this is exactly when clarity matters most.</p><p>Not every lawyer needs to become technical. But more lawyers do need a trustworthy framework for thinking about what these tools are.</p><p>Where can they help, where they introduce risk, and how to approach them with sound judgment.</p><p>The hesitation is rational. Permanent avoidance is not.</p><p>Caution makes sense.</p><p>Blind adoption or blind rejection does not.</p><p>The better path is thoughtful engagement: learning enough to understand the risks, the opportunities, and the practical realities without getting swept up in the hype.</p><p>That is the spirit in which I wrote this guide.</p><p>Not to pressure adoption.</p><p>Not to dismiss skepticism.</p><p>Not to pretend the problems are solved.</p><p>Only to help separate signal from noise.</p><p>Lawyers who approach this technology with clarity and discipline will be better positioned to serve their clients than those who either chase the hype or ignore it entirely.</p><h3>Chapter 3: Why AI Hallucinates Legal Citations</h3><p>Let me start with the story every lawyer already knows.</p><p>A lawyer asks an AI tool to find case law supporting a legal argument. The tool returns several citations.</p><p>The formatting looks right. The reasoning sounds plausible. The case names are convincing.</p><p>But the cases do not exist, and the lawyer does not verify them.</p><p>The lawyer files the brief. The court discovers the fabrication. Sanctions follow. The lawyer’s reputation takes a hit that may never fully recover.</p><p>This has happened enough times now that the pattern is familiar. And the conclusion many attorneys have drawn is understandable: AI is unreliable and dangerous.</p><p>I think that conclusion is only partly right. And the part that is wrong is costing the profession dearly.</p><p>How the Models Work</p><p>To understand why AI tools fabricate legal citations, it helps to understand something basic about how they work.</p><p>Tools like ChatGPT, Claude, and Gemini are technically what the AI world calls large language models. For present purposes, the label matters less than the basic idea. These models learn by reading enormous amounts of text. Books, articles, websites, academic papers, and news reports all contribute to what they have read. In simple terms, the more robust and accessible the material on a subject, the more likely the model is to give you a useful answer about it.</p><p>That creates a distinct problem for legal research.</p><p>Much of American case law sits inside proprietary databases such as Westlaw and Lexis. That matters because if the model cannot read the underlying material, it cannot learn it in the same way. And because those databases are not openly available for these models to read in the way public websites, articles, and books are, they have not read the full universe of case law.</p><p>That gap matters.</p><p>It is a major part of the hallucination problem.</p><p>There is a second issue as well.</p><p>These models are built to give you an answer. They do not naturally respond by saying, “I do not have enough reliable information to answer that.” They tend to generate a response anyway.</p><p>When the information they have read is broad and accessible, that tendency can still produce something useful. But when the underlying material is thin, fragmented, or closed off, as it often is in legal research, the tool may still give you something that looks polished, complete, and authoritative.</p><p>In legal research, that is a serious risk.</p><p>Why the Quality Varies</p><p>Understanding that limitation is essential. If you do not understand what the model has read, what it has not read, and where the danger zones are, you can get into trouble quickly.</p><p>In some areas of legal research, these models perform surprisingly well.</p><p>Ask about a major Supreme Court case and the answer may be detailed and largely accurate. Why? Because landmark decisions generate extensive public analysis. Law review articles, news stories, blog posts, commentary, and academic discussion give the model a much stronger foundation to work from.</p><p>But ask about a niche appellate ruling, a procedural order in a mass tort MDL, or an obscure district court decision, and the quality can deteriorate quickly. The underlying material may be sparse or fragmented, and the answer often reflects that.</p><p>The result is something lawyers now recognize all too well: work that looks like legal research, but is not legal research.</p><p>From the lawyer’s perspective, the experience is disorienting. The model seems highly capable one minute and dangerously unreliable the next. Without understanding why, it is easy to dismiss the technology entirely.</p><p>That conclusion is understandable. But it is also incomplete.</p><p>The Distinction That Matters</p><p>Hallucinations are most likely when the model is being asked to produce precise, verifiable facts from a limited or closed body of information. Legal citation is close to a worst-case scenario.</p><p>But there is a wide range of work AI can do for a law firm that does not suffer from those same constraints.</p><p>Ask AI to draft marketing content, refine client communications, pressure-test arguments, or organize case strategy. These tasks rely less on retrieving exact records and more on reasoning, structure, synthesis, and language. The available material is broader, and the risk profile is different.</p><p>Understanding that distinction matters greatly.</p><p>Many lawyers encountered AI in one of its weakest use cases, saw it fail, and treated that failure as a verdict on the entire technology. They tested it in a constrained data environment and generalized the result.</p><p>But the technology is not simply reliable or unreliable. It is uneven. And understanding where it is uneven is not just a technical exercise. It is a strategic one.</p><p>None of this means hallucinations should be taken lightly. They should not. Citations must be verified. Answers require review. Professional judgment remains essential.</p><p>But dismissing the technology because it struggles in one particularly difficult context would also be a mistake.</p><p>The question is not whether AI hallucinates. It does.</p><p>The better question is where that risk is central, and where it is not.</p><p>That is the distinction that matters.</p><p>And the firms that understand it early will be in a much stronger position than those still treating one highly visible failure mode as the full story.</p><h3>Chapter 4: Why the AI Conversation in Law Needs Better Questions</h3><p>Open LinkedIn on almost any given day and you will find no shortage of posts about AI and the legal profession.</p><p>Most of them say some version of the same thing.</p><p>AI will save lawyers time. Here are the top ten AI tools every firm needs. AI is transforming the legal industry.</p><p>None of those statements is entirely wrong.</p><p>Some of them may even be helpful.</p><p>But many of them are still primarily slogans.</p><p>And slogans do not help a managing partner make a six-figure technology decision.</p><p>They do not help an associate understand when AI output can be trusted and when it must be verified.</p><p>They do not help a compliance officer determine whether an AI-generated advertisement satisfies the rules in New York as opposed to Texas.</p><p>The legal profession does not need more slogans.</p><p>It needs better questions.</p><p>Much of the content being produced about AI and law right now seems to fall into three broad categories.</p><p>Vendor Content</p><p>The first is vendor content presented as thought leadership.</p><p>A company builds an AI tool, then publishes articles explaining why the problem its tool solves is one of the most important problems in the profession. Not surprisingly, the analysis often points in the same direction: toward the product being offered. The framing is designed to sell more than to educate. And because it is often written by people who built the technology rather than people who practice law, it can miss the nuances that matter most to working attorneys.</p><p>Academic Commentary</p><p>The second is academic or policy-oriented commentary aimed largely at other thought leaders.</p><p>At a high level, much of it is thoughtful, important and necessary.</p><p>Questions about agentic AI, the future of the billable hour, and the broader restructuring of legal services do matter.</p><p>But that conversation can sit far above the day-to-day reality of a firm trying to decide whether to invest in a new platform or keep doing things the way it has always done them.</p><p>Entry-Level Tactical Advice</p><p>The third is entry-level tactical advice.</p><p>How to draft an email with AI.</p><p>How to summarize a document.</p><p>Ten prompts every lawyer should know.</p><p>This serves a purpose, especially for those lawyers who are new to the tools.</p><p>But it offers little guidance on the strategic decisions that determine whether AI creates value or risk.</p><p>What is often missing across all three is the right altitude</p><p>The first tends to focus on novelty and the product. The second often sits far above the firm. The third stays at the level of individual tasks. What many law firms need is something in between.</p><p>A More Useful Conversation.</p><p>The conversation firms need asks different questions.</p><p>It asks about the data environment.</p><p>Where is the model getting its information?</p><p>Is the training data relevant to the task?</p><p>Is it complete?</p><p>Is it current?</p><p>As I discussed in a previous article, one reason AI hallucinates legal citations is not simply that the technology is flawed.</p><p>It is that the data environment for legal research is fundamentally incomplete.</p><p>A surface-level analysis rarely gets to that point.</p><p>A more useful one starts there.</p><p>It asks about task design.</p><p>Not, “Can AI do this?” but, “What exactly am I asking it to do, and what are the consequences if the output is wrong?”</p><p>As we explored earlier, the risk profile of AI changes dramatically depending on the task. A lawyer who understands that evaluates each use case on its own merits. A lawyer who does not may either reject the technology entirely or adopt it without the right safeguards.</p><p>It asks about workflow.</p><p>How does this tool fit into existing systems?</p><p>Does it connect to case management software?</p><p>Does it create more work or less?</p><p>What does the review process actually look like?</p><p>Who is responsible for verifying output?</p><p>These are not glamorous questions. They do not usually drive engagement on LinkedIn. But they are often the questions that determine whether an AI tool becomes genuinely useful or just another expensive piece of software that sits underused.</p><p>And it asks about judgment.</p><p>Not the AI’s judgment.</p><p>Yours.</p><p>Where does the lawyer’s expertise remain essential?</p><p>Where does human review add real value, and where does it become unnecessary friction?</p><p>How do you build processes that preserve the benefits of the technology without losing the discipline the profession requires?</p><p>These are not the questions most likely to go viral.</p><p>But they are the questions that matter.</p><p>They separate firms making thoughtful decisions from firms simply reacting to the latest trend.</p><p>The Question of Balance</p><p>How deep do you go before you stop being useful to the reader?</p><p>It is not always an easy call.</p><p>In a previous article I wrote about general-purpose AI models hallucinating legal citations. Someone with deeper technical specialization could write a long white paper on retrieval augmented generation and why properly grounded systems hallucinate less. Someone deeper still could correct me by pointing out that large language models do not actually read text or learn facts the way humans do. They predict tokens probabilistically based on patterns in their training data.</p><p>Both points are valid, but may be the wrong altitude for a managing partner trying to decide whether to invest in a new platform next quarter.</p><p>That is the challenge.</p><p>Go too shallow and you get slogans. Go too far into the weeds and you get a technical paper. Neither helps a firm make a sound decision. The real challenge is finding the level where the conversation is deep enough to be useful, but not so deep that it becomes disconnected from the decision in front of the reader.</p><p>This series is my attempt to find that middle ground.</p><p>It is not always easy, but I will do my level best to be useful.</p><h3>Chapter 5: A Better Framework for Thinking About AI in a Law Firm</h3><p>When a vendor pitches your firm on AI, an associate asks about a new tool, or leadership is deciding whether to invest resources, do you have a clear framework for evaluating it? Or does the conversation collapse into a vague yes-or-no decision?</p><p>In the previous articles, I argued two things. First, AI should not be treated as a single verdict, but evaluated task by task. Second, much of the current conversation around AI in law is too abstract to support real decision-making.</p><p>This article brings those ideas together.</p><p>What follows is a practical framework for thinking about AI in your firm. It is not a product recommendation or a prediction about the future. It is simply a way to organize your thinking so that you have a clearer mental model when evaluating whether AI belongs in your practice.</p><p>A useful starting point is to break AI use into six broad categories of legal work. Each has distinct strengths, weaknesses, and risk profiles. Understanding those differences may be the single most useful thing a firm leader can do right now.</p><p>Category 1: Retrieval Tasks</p><p>This is where AI finds and returns information: legal research, case citations, statutes, and regulations.</p><p>Strengths: AI can process enormous volumes of text and identify potentially relevant information faster than any human research team.</p><p>Weaknesses: As I discussed in an earlier chapter, general-purpose AI models do not reliably have access to the full universe of case law. Much of it sits behind paywalls. When the model lacks reliable data, it may produce results that look correct but are not. This is where hallucinations occur.</p><p>Risk profile: High. Every output must be independently verified against authoritative source. No exceptions. This is not the place to save time. This is the place to be disciplined.</p><p>Practical guidance: Use AI to generate starting points for research, not conclusions. Treat every citation as unverified until confirmed. The tool can point you in a direction. You still have to do the work.</p><p>Category 2: Reasoning Tasks</p><p>This is where AI helps think through a problem: analyzing the strengths and weaknesses of a legal argument, identifying counterarguments, evaluating litigation strategy, and pressure-testing assumptions.</p><p>Strengths: Large language models are often very good at structured reasoning. They can hold multiple variables at once, identify inconsistencies, and surface considerations that a single attorney might miss. They do not get tired, and if prompted properly, they will challenge your assumptions rather than simply reinforce them.</p><p>Weaknesses: AI lacks context. It does not know your client, your judge, your procedural posture, or your relationship with opposing counsel. It reasons in the abstract. Applying that reasoning requires human judgment.</p><p>Risk profile: Medium. The output is analytical, not authoritative. It is useful as a thinking partner, not as a decision-maker.</p><p>Practical guidance: Use AI to stress-test your own thinking. Ask it to argue the other side. Ask it to identify the weakest points in your position. Then apply your own judgment to decide what matters. This is one of the highest-value uses of AI in legal practice, and one of the least discussed.</p><p>Category 3: Drafting Tasks</p><p>This is where AI generates text: first drafts of briefs, motions, memos, contracts, and client communications.</p><p>Strengths: Speed. AI can produce a structurally sound first draft in minutes. For routine documents with predictable formats, the efficiency gain can be significant.</p><p>Weaknesses: AI drafts tend toward the generic. They often lack the specificity, nuance, and voice that distinguish strong legal writing from merely adequate legal writing. They may also include inaccurate statements of law, especially when citations are involved. Tone is another issue. A letter that is slightly too aggressive, or a client communication that is slightly too casual, can create problems the attorney then has to unwind.</p><p>Risk profile: Medium to high, depending on the document. An internal memo carries less risk than a court filing. A first draft carefully reviewed by an experienced attorney is very different from a first draft sent or filed as-is.</p><p>Practical guidance: AI is your associate, not your partner. It produces a draft. You turn it into something worth filing or sending. Speed is useful, but it does not replace judgment.</p><p>Category 4: Communication Tasks</p><p>This is where AI supports client and internal communication: intake follow-ups, scheduling, status updates, FAQs, and onboarding.</p><p>Strengths: Consistency and speed. AI can help ensure that every new lead receives a prompt, professional response. It can maintain communication cadence during high- volume periods when your team is stretched thin. For firms handling large campaigns, this directly affects case acquisition.</p><p>Weaknesses: AI cannot fully interpret emotional nuance. A message that works for a corporate client may feel tone-deaf to someone dealing with loss or distress. AI can approximate empathy. It cannot replicate it.</p><p>Risk profile: Low to medium. Routine communications carry minimal risk. Sensitive client interactions require human involvement.</p><p>Practical guidance: Use AI for the first touch and routine follow-up. Reserve human attention for the moments that require genuine empathy, strategic sensitivity, or</p><p>professional judgment. Build the workflow so the system knows when to hand it off to a person.</p><p>Category 5: Workflow and Automation Tasks</p><p>This is where AI helps manage processes: document organization, calendar management, deadline tracking, data extraction from intake forms, routing leads to the right practice group, generating internal reports.</p><p>Strengths: AI is tireless, consistent, and fast at repetitive operational tasks. It is good at tracking deadlines, locating documents, and keeping leads moving through the intake pipeline.</p><p>Weaknesses: Workflow automation is only as good as the system it operates within. If your case management software does not integrate well with the tool, you may create more work rather than less. The technology itself may function perfectly well. Implementation is where firms often stumble.</p><p>Risk profile: Low. These are internal operational tasks. The consequences of an error are generally correctable.</p><p>Practical guidance: Start here. Workflow automation is often the lowest risk and highest return entry point. It builds familiarity and confidence before moving into higher- stakes applications.</p><p>Category 6: Marketing and Business Development Tasks</p><p>This includes content creation, advertising, SEO, generative and answer engine optimization, audience analysis, campaign management, and more.</p><p>Strengths: High potential value. AI is trained on vast amounts of publicly available marketing and persuasion data. It can generate content, analyze performance, and optimize messaging at scale.</p><p>Weaknesses: Compliance. AI does not reliably adhere to attorney advertising rules. It may generate claims that imply guarantees, overstate results, or violate jurisdictional standards. Human compliance review is not optional. It is part of the job.</p><p>Risk profile: Low to medium. The efficiency gains can be substantial, but only if the content passes through a meaningful compliance and brand review process.</p><p>Practical guidance: AI can do a large portion of the work in legal marketing. Compliance review, strategic judgment, and brand voice calibration remain human work. That is where attorney oversight and marketing experience still earn their keep.</p><p>Across all categories, client confidentiality and data security obligations still apply. Tool selection and workflow design must reflect that.</p><p>The question is not whether to adopt AI. The question is how.</p><p>This framework is not exhaustive. It does not cover every possible use case. But it is far more useful than asking a generic question like, “Should our firm adopt AI?”</p><p>The better questions are these: Where should we start? Where is the risk? Where is the value? What oversight is required before we move forward?</p><p>In most cases, the path forward is straightforward.</p><p>Start with workflow and automation. Build confidence. Move into marketing and business development. Capture efficiency. Then work gradually into the higher-risk categories with the review structures necessary to use the technology responsibly.</p><p>That is how you adopt AI like a lawyer rather than like a consumer.</p><p>The earlier chapters of this guide lay the groundwork for this framework. For firm leaders working through these questions, the full series may provide additional context.</p><h3>Chapter 6: You Do Not Need to Be an AI Expert, But You Do Need Strategic Understanding</h3><p>I hear some version of this regularly from attorneys:</p><p>“I didn’t go to law school to learn about technology.”</p><p>Fair enough. Neither did I.</p><p>There is a natural and understandable resistance among many attorneys, particularly senior partners and firm owners, to engaging with AI at anything beyond a surface level. They built successful practices on legal judgment, client relationships, experience, and hard work. Technology was something other people handled: IT departments, vendors, consultants, and marketing agencies.</p><p>And for most of the history of legal technology, that division of labor made sense.</p><p>You did not need to understand how your email server worked in order to send an email. You did not need to understand how Westlaw’s search algorithm functioned in order to use it effectively. The tools were mature, relatively predictable, and the consequences of not understanding the underlying technology were usually limited.</p><p>AI is different, and the difference matters.</p><p>The reason AI requires a different level of engagement is because it produces outputs that resemble judgment. Your email server never drafted a motion. Your CMS never wrote a client letter. Traditional legal technology was largely passive. You instructed the system, and the output was generally predictable.</p><p>Generative AI is not passive. You provide a prompt, and the system creates something new: a draft brief, a marketing strategy, a case evaluation, a client communication, or a research summary.</p><p>Sometimes the output is excellent. Sometimes it is adequate. Sometimes it is dangerously wrong.</p><p>The difference between those outcomes often turns on:</p><p>• the quality of the instructions,</p><p>• whether the task was appropriate for AI in the first place, and</p><p>• the rigor of the human review process.</p><p>Those are not merely technical questions. They are strategic and managerial questions. They require judgment. And they cannot be fully delegated to an associate, an IT department, or an outside vendor.</p><p>Importantly, this does not mean lawyers need to become AI experts.</p><p>You do not need to understand neural network training. You do not need to know what transformer architecture means. You do not need to study the mathematics behind large language models. But firm leadership does need to understand several things clearly.</p><p>First: What AI Is Good at and What It Is Not</p><p>This is the framework I discussed in earlier chapters.</p><p>Retrieval tasks can carry substantial risk because of hallucinations and data limitations. Drafting tasks can create useful starting points, but they do not eliminate the need for attorney review and judgment. Reasoning tasks may provide value, but they still require careful supervision. Marketing and business development tasks often present some of the highest upside with relatively manageable risk.</p><p>A senior partner does not need to personally use every AI tool.</p><p>But leadership does need enough understanding to set intelligent policy, allocate resources appropriately, and establish realistic expectations inside the firm.</p><p>Second: Where The Data Goes</p><p>When your firm uses AI tools, client information, case details, internal strategy, and work product may be processed by systems outside your control.</p><p>Is the data stored? Is it encrypted? Is it retained? Is it used for model training? Who has access to it?</p><p>These are governance questions.</p><p>And they are leadership responsibilities.</p><p>The answers also vary significantly from vendor to vendor. A partner who cannot ask these questions intelligently is not in a position to adequately protect the firm or its clients.</p><p>Third: The Compliance Landscape</p><p>Attorney advertising rules, disclosure obligations, confidentiality requirements, and ethical duties around competence and supervision are evolving quickly.</p><p>No managing partner needs to memorize every ethics opinion or proposed rule change.</p><p>But firm leadership does need to understand that the landscape is changing and ensure that someone inside the organization is actively monitoring those developments.</p><p>Ignoring the issue is not a strategy.</p><p>Fourth: Evaluating Vendors Realistically</p><p>The AI vendor market is crowded, noisy, and filled with inflated claims.</p><p>Many vendors have never practiced law or worked inside a law firm. They often do not understand legal workflows, compliance obligations, privilege concerns, or the operational realities of legal practice.</p><p>That gap matters.</p><p>There is an analogy here that most attorneys will immediately recognize.</p><p>You do not need to be an accountant to run a law firm. But you do need to understand your financial statements well enough to make informed decisions. You need to understand the difference between revenue and profit. You need to know whether your overhead is sustainable.</p><p>AI is similar.</p><p>You do not need to build the systems. You do not need to configure the infrastructure. You do not need to write code.</p><p>But you do need enough strategic understanding to govern the technology responsibly, evaluate risk intelligently, and ask the right questions of the people handling the technical work.</p><p>If you have been sitting out the AI conversation because it felt like someone else’s problem, I understand. The noise surrounding AI is exhausting. The jargon can be impenetrable. And most attorneys are already operating at capacity.</p><p>But this is not going away.</p><p>The firms that thrive over the next decade will not necessarily be the firms with the most AI tools.</p><p>They will be the firms whose leaders understood enough, early enough, to make sound strategic decisions.</p><p>Not perfect decisions. Strategic ones.</p><h3>Chapter 7: Confidentiality, Privilege, and Professional Judgment in the Age of AI</h3><p>Everything in this guide so far has focused on reducing fear and increasing clarity. This one addresses where the fear is justified.</p><p>Confidentiality and attorney-client privilege are not abstract concepts. They are foundational obligations, among the first things we learn in law school and the last things we should ever compromise. AI introduces legitimate new risks to both, and they deserve serious treatment.</p><p>I want to be direct about what this chapter is and what it is not.</p><p>This is not legal advice. I am offering general observations and identifying questions that I believe every firm should be considering. Ethical rules vary by jurisdiction. Privilege analysis is often fact-specific. Technology platforms regularly change their terms, data practices, and security frameworks. Any firm evaluating AI tools should consult qualified</p><p>counsel about the rules that apply in its jurisdiction and the specific technology under consideration.</p><p>With that understood, here is what I believe matters most.</p><p>Where Does the Data Go?</p><p>When an attorney types a client’s case details into an AI tool, that information leaves the attorney’s immediate control. It travels to an external system, is processed by software the attorney did not build, and is maintained by a company the attorney may know very little about.</p><p>That should prompt obvious questions. Where is the data stored? Who has access to it? How long is it retained? Is it encrypted in transit and at rest? Is it used to train or improve the model?</p><p>If a firm’s proprietary work product, legal strategies, client information, or case analysis is being used to improve a model that serves other firms, including potentially adverse parties, the implications are serious. A firm may be exposing some of its most valuable intellectual assets without fully realizing it.</p><p>Not every AI tool works this way. Some vendors offer clear contractual commitments that client data will not be used for model training. Some provide segregated or isolated environments to protect firm information. Others are less transparent.</p><p>The point is simple: The terms matter, the architecture matters, and the contract matters. And in my experience, many attorneys do not read them carefully enough.</p><p>Could This Affect Privilege?</p><p>This is an area where the law is still evolving, and it is important not to overstate what is settled. At the same time, the underlying concern is not merely theoretical.</p><p>Attorney-client privilege generally protects confidential communications made for the purpose of obtaining or providing legal advice, while the work product doctrine protects materials prepared in anticipation of litigation. Both protections, however, may be at risk if protected information is disclosed to a third party without appropriate safeguards.</p><p>Whether inputting privileged information into a third-party AI tool constitutes such a disclosure is a serious question with far-reaching implications. The answer may depend on a range of factors, including the specific tool at issue, the contractual terms governing its use, the platform’s technical design and data practices, and how a given jurisdiction approaches waiver.</p><p>In light of this uncertainty, practices vary. I am not going to tell anyone which approach is correct. That is a judgment each firm must make for itself. But any firm using AI without having considered this question is taking an unnecessary risk.</p><p>Competence and Supervision.</p><p>Some emerging professional guidance has emphasized that lawyers using AI need a reasonable understanding of the technology’s capabilities and limitations, and that supervisory duties extend to AI-generated work product.</p><p>That should not be controversial.</p><p>A senior attorney who assigns a task to an AI tool bears essentially the same supervisory obligation as one who assigns a task to a junior associate. The output must be reviewed. The analysis must be verified. The final product must meet the same standard the attorney would apply to any work bearing the attorney’s name.</p><p>The only thing that has changed is the nature of the delegate.</p><p>You Need an Internal Policy</p><p>Does your firm have a written policy governing AI use? If not, it should.</p><p>A responsible AI policy does not need to be long or complicated. But it should answer a few questions clearly. Which tools are approved for use? What types of information may and may not be entered into those tools? Who is responsible for reviewing AI-generated output? What disclosure obligations apply when AI is used in work product submitted to a court? How are vendor relationships evaluated and monitored over time?</p><p>The firms that will navigate this transition most effectively are the ones that establish clear boundaries before problems arise, not after. An ad hoc approach, where each lawyer decides independently what to input and how to use AI, is the riskiest path of all.</p><p>Closing Thought</p><p>I want to close with something that often gets lost in the noise around AI ethics.</p><p>The attorneys who are most cautious about AI are often the ones taking their professional obligations most seriously. Responsible skepticism is not resistance to progress; it is the foundation on which responsible progress must be built.</p><p>The goal is not to move fast.</p><p>The goal is to move wisely. To adopt AI in ways that enhance practice without compromising the trust clients have placed in us.</p><p>That trust can take years to build. No technology, however powerful, is worth treating it lightly.</p><h3>Chapter 8: Assessing the AI Vendor Landscape</h3><p>If you are a managing partner trying to evaluate AI vendors right now, the market can feel noisy. Large established companies. Smaller upstarts. Specialized tools. General platforms. Strong claims from every direction.</p><p>The challenge is not simply identifying what is available. It is deciding how to evaluate the field clearly.</p><p>If you have been leading a firm for a long time, you have probably written many large checks over the years to established legal technology companies. These are organizations with thousands of employees, decades of history, and well-known names in the profession. When you buy from them, there is comfort in that. You assume there will be strong infrastructure, solid support, and a company that will still be standing next year.</p><p>That instinct has served the profession well for a long time. And there are good reasons for it.</p><p>But in the AI era, that same instinct may also cause some firms to overlook products that fit them especially well.</p><p>Where Large Vendors Have Advantages</p><p>Let me begin with what the large vendors do well.</p><p>They have scale. They can invest in security infrastructure, compliance processes, customer support, and enterprise-level operations in ways that small teams often cannot. They have brand recognition that makes procurement decisions easier to justify internally. Those are real advantages, and I do not pretend otherwise.</p><p>But there is also a structural pattern worth noticing.</p><p>Where Large Vendors Can Struggle</p><p>Many of the large legal technology vendors built their core products years, and in some cases decades, before the current generation of AI arrived. Their platforms were designed for a different era: search-based legal research, document management, e- discovery, and workflow systems built around older assumptions. When AI arrived, many</p><p>of these companies did what large companies often do. They layered the new technology onto existing products.</p><p>The result is often useful, but not always transformative. The AI features can feel like additions rather than foundations. The interface is still organized around workflows that predate the technology. The product may do more than it once did, but it does not necessarily do things differently.</p><p>That is not really a criticism. It is a structural reality.</p><p>Large companies with established customer bases, revenue streams, and product architectures often face strong internal pressure to protect what already works. Radical redesign creates risk. It can disrupt existing users, confuse existing buyers, and undermine existing revenue. The incentive structure tends to favor incremental improvement over fundamental rethinking.</p><p>Where Smaller Teams Have Advantages</p><p>Now consider what is happening at the other end of the market.</p><p>Some of the most interesting AI products being built right now are coming from very small teams. That is not an accident. It is, in part, a consequence of what AI itself makes possible.</p><p>AI is a powerful force multiplier. A small number of highly capable builders can design, test, and deploy software at a pace that would have required a much larger team not long ago. And when that small team understands a particular market deeply, it can often move faster than a large company that first has to align committees, budgets, product roadmaps, and internal politics before doing anything substantial.</p><p>The advantage of small teams is not simply that they are leaner. It is that they can be more focused.</p><p>A small team building AI tools specifically for a particular type of firm is thinking about a narrower set of problems, a narrower set of workflows, and a narrower set of compliance considerations. Every feature decision is filtered through the same question: does this actually serve the lawyers it was built for?</p><p>A large vendor serving many practice types across many firm sizes usually cannot operate with that same level of specificity. Its product has to be broad enough to serve many constituencies at once. The result may be a tool that works reasonably well for many users, but not exceptionally well for the firm with more specialized needs.</p><p>Where Smaller Teams Raise Legitimate Questions</p><p>Many newer companies do not yet have brand recognition, and getting noticed in a crowded market is difficult. A product may be excellent and still struggle to gain traction.</p><p>Smaller vendors also may not have established customer bases or predictable revenue streams, which can raise legitimate questions about continuity, support depth, and long- term stability. Those concerns deserve serious attention as well.</p><p>The Real Question is Product Fit</p><p>That leads to an important question: how well does the product actually fit your workflow, as opposed to how much your team has adapted itself to it? When was the last time the vendor introduced something that materially changed how your firm operates?</p><p>Size is not a proxy for quality. Legacy is not a proxy for innovation. And brand familiarity is not a proxy for product fit.</p><p>Why Domain Understanding Matters</p><p>What often makes the difference is not just technical talent, but strong domain understanding. Teams that understand how law firms actually operate may be better positioned to build products that fit those firms well. The rhythms and practical constraints of legal practice are not always easy to appreciate from the outside, and products tend to be stronger when that understanding is built in from the beginning.</p><p>That kind of domain knowledge can be difficult to build into a product after the fact. And it may be one reason certain smaller teams are able to build tools that feel more aligned with the lawyers they serve.</p><p>I am not suggesting that every small AI vendor is better than every large one. Of course not. There are excellent products from large companies and weak products from small ones.</p><p>The point is simply that the right evaluation criteria are fit, expertise, judgment, and execution, not just headcount or name recognition. For many firms, the best answer may be a thoughtful mix of large, established vendors and smaller, more specialized teams.</p><p>No matter the size of the vendor, ask the hard questions. Demand specificity. And look closely at the product.</p><h3>Chapter 9: Where AI Can Actually Help a Law Firm Today</h3><p>The earlier chapters built a foundation: how AI works, where it fails, how to evaluate risk, how to think about vendors, why confidentiality matters, and where the current conversation often misses the level where real decisions get made.</p><p>Now it is time to get practical.</p><p>This article is about where AI can create real value for a law firm today. Not next year. Not after another round of model improvements. Today. With tools that already exist and workflows that a reasonably motivated firm could begin implementing without hiring a data scientist.</p><p>What follows are nine areas where AI can already be useful. For each one, the real question is not simply what the technology can do, but why it matters and where human judgment still belongs.</p><p>Not every firm will start in the same place. But most firms will find opportunities in at least some of the categories below.</p><ol><li>Intake communication</li></ol><p>When a potential client contacts your firm, the clock starts. A prospective client who fills out a contact form at 9:00 p.m. and hears nothing until the next morning may already have moved on.</p><p>AI can help close that gap.</p><p>It can respond immediately, not with a generic autoresponder, but with an intelligent first touch that acknowledges the inquiry, asks relevant follow-up questions, sets expectations, and captures the context your intake team will need.</p><p>What to watch for: tone. AI can approximate empathy, but it cannot truly replicate it. Sensitive inquiries, especially those involving trauma, grief, or urgency, should be routed quickly to a human being.</p><ol><li>Follow-up systems</li></ol><p>Many firms do not lose cases because they failed to generate leads. They lose cases because they failed to follow up consistently.</p><p>A lead comes in, the team is busy, the follow-up slips to tomorrow, then to the next day, and eventually disappears.</p><p>AI can help manage that cadence: scheduled messages, reminders, status tracking, and escalation when a lead goes cold. This is not glamorous work. But it often makes the difference between a signed retainer and a lost opportunity.</p><p>What to watch for: over-automation. There is a point where persistence stops feeling helpful and starts feeling mechanical or intrusive. Someone should still own the cadence.</p><ol><li>Marketing content</li></ol><p>AI is very good at producing first drafts of marketing content: social copy, newsletter drafts, website pages, video scripts, and similar materials.</p><p>The available training material for persuasion, narrative structure, and audience engagement is broad and accessible, which makes this a much better use case than legal citation generation.</p><p>A firm that publishes inconsistently can often increase its output significantly if AI handles the first draft and a human handles review, compliance, and brand voice.</p><p>What to watch for: attorney advertising compliance. AI does not know your jurisdiction’s rules the way a lawyer or experienced legal marketer must know them. Every piece of content still requires review before publication.</p><ol><li>Message testing</li></ol><p>Before committing to a campaign theme, a tagline, or an advertising angle, AI can help sharpen your thinking. You can test multiple versions of a message, ask which is more likely to resonate with a particular audience, and surface possible compliance concerns before spending money on production.</p><p>This kind of work once required focus groups or expensive outside research. AI does not eliminate those entirely, but it can provide a fast and inexpensive first pass.</p><p>What to watch for: overconfidence. AI message testing is directional, not definitive. It is useful for generating hypotheses, not proving them with certainty.</p><ol><li>Internal knowledge organization</li></ol><p>Every law firm has valuable institutional knowledge that often lives in the heads of senior lawyers. Strategies that worked. Arguments that failed. Lessons learned about specific judges, jurisdictions, or opposing counsel.</p><p>Too often, that knowledge disappears when people leave.</p><p>AI can help capture, organize, and retrieve that information through internal knowledge bases, searchable case debriefs, and natural-language querying. In the right environment, institutional memory can become more durable and more usable.</p><p>What to watch for: confidentiality. This is not a consumer-tool use case. Internal knowledge systems require secure, private infrastructure and careful governance.</p><ol><li>Argument pressure-testing</li></ol><p>This remains one of the highest-value uses of AI in legal practice.</p><p>Give the tool the structure of your argument. Ask it to identify the weakest points. Ask it to construct the strongest counterargument. Ask it what opposing counsel is likely to say.</p><p>It will not always be right. But it will often surface issues you might miss when you have lived inside a case for too long.</p><p>What to watch for: false assumptions. AI often reasons better than it knows facts. If you use case-specific facts to pressure-test an argument, do so only in a manner consistent with your confidentiality obligations, your firm’s policies, and the safeguards of the system you are using. Then apply your own judgment to what comes back.</p><ol><li>Research assistance with verification</li></ol><p>I have been clear throughout this guide that AI-generated citations must be independently verified. That remains true. But that does not mean AI has no role in research.</p><p>AI can help identify relevant areas of law, generate research outlines, summarize public commentary, and suggest lines of inquiry worth exploring. Used properly, it functions like a research assistant who is very fast, widely read, and occasionally wrong.</p><p>What to watch for: fabricated citations. Always. Every time. Without exception.</p><ol><li>Workflow automation</li></ol><p>Document organization. Calendar management. Deadline tracking. Intake form processing. Lead routing. Internal reporting.</p><p>These are repetitive operational tasks that consume staff time every week and often suffer from the limits of human attention.</p><p>AI can often handle this work faster and more consistently. This is one of the best places for most firms to begin. The risk is relatively low, the gains can be immediate, and the process helps build internal comfort before moving into higher-stakes applications.</p><p>What to watch for: integration. A tool that does not connect well to your case management system may create a second layer of work rather than remove one.</p><ol><li>Competitor and market analysis</li></ol><p>What are other firms in your practice area doing? What messages are they running? Which campaigns appear to be gaining traction? What case categories are becoming crowded, and which are still emerging?</p><p>AI can scan, synthesize, and organize this kind of competitive intelligence at a scale that would be difficult to replicate manually. Used well, that can help a firm allocate resources more intelligently and see opportunities earlier.</p><p>What to watch for: incomplete inputs. AI analysis is only as good as the data you give it and the data it can access. Use it to inform judgment, not replace it.</p><p>The point is disciplined usefulness.</p><p>All nine of these use cases are available today. None requires exotic infrastructure. None depends on waiting for some future version of the technology to arrive. What they do</p><p>require is discipline about where AI adds value and where human judgment remains essential.</p><p>That discipline is the difference between AI as a real competitive advantage and AI as an expensive experiment that never quite delivers.</p><p>Start where the risk is lowest and the value is clearest. Build from there.</p><h3>Chapter 10: How Law Firms Can Evaluate AI Tools Without Getting Lost in the Noise</h3><p>At some point, the research ends and the decision begins.</p><p>You have read the articles. You have heard the pitches. You have seen the demos. Someone at your firm, whether it is you, a younger associate, or an outside consultant, has recommended investing in an AI tool. The question is no longer whether to engage with AI. It is which tool, for what purpose, and on what terms.</p><p>This is where many firms get lost. Not because they lack intelligence or judgment, but because the market is built to overwhelm. Vendors compete for attention with polished demos, impressive statistics, and language designed to create urgency. The number of options makes comparison difficult. And the technology is unfamiliar enough that even experienced decision-makers can feel uncertain about what questions to ask.</p><p>This article is a framework you can use before you start evaluating AI tools. It is not exhaustive. But it covers the questions that matter most.</p><ol><li>What problem are we actually trying to solve?</li></ol><p>Start here. Not with the tool. Not with the vendor. With the problem.</p><p>It sounds obvious, but firms often adopt technology because it is impressive rather than because it addresses a clearly defined need. A tool that generates polished marketing content is of limited value if your real bottleneck is intake response time. A research assistant may be the wrong investment if your research workflow is already strong and your marketing is where you are losing opportunities.</p><p>Define the problem in concrete terms. We are losing leads because response times are too slow. Our content output is inconsistent. Compliance review is manual and inefficient.</p><p>We have no reliable way to track campaign performance. If the problem cannot be articulated clearly, the evaluation process is already off course.</p><ol><li>What task category does the tool actually address?</li></ol><p>This connects directly to the framework from the previous articles. Is this a retrieval task, a reasoning task, a drafting task, a communication task, a workflow task, or a marketing task? Each carries a different risk profile and requires a different level of human oversight.</p><p>A workflow tool designed for low-risk automation should not be evaluated the same way as a tool used for legal research or court-facing drafting. The questions you ask, the safeguards you require, and the standards you apply should match the nature of the task.</p><p>If a vendor cannot explain clearly what category of work its product is meant to perform, that is not a small issue. It is a warning sign.</p><ol><li>What data is entering the system?</li></ol><p>This question belongs in every evaluation conversation. What information will your firm be putting into the tool? Client names? Case details? Legal strategy? Financial data? Marketing performance data?</p><p>For each category of information, you need clear answers. Where is it stored? Is it encrypted? Is it used to improve the model? Can it be accessed by other customers? What happens to it if the relationship ends?</p><p>If the answers are vague, evasive, or buried in terms of service that no reasonable person would understand on a first read, that should give you pause.</p><ol><li>What happens when the tool gets something wrong?</li></ol><p>Every AI tool produces errors. The relevant question is not whether errors will occur, but what happens when they do.</p><p>If a workflow tool routes something incorrectly, the consequence may be an inconvenience. If a research tool produces a fabricated citation that ends up in a court filing, the consequence may be serious and lasting.</p><p>Map the likelihood of error against the consequence of error. High-consequence tasks require more rigorous review, stronger internal controls, and a slower adoption curve. Lower-consequence tasks can usually be adopted more quickly with lighter oversight.</p><p>And yes, it is worth asking the vendor directly about known limitations and failure modes. A vendor who suggests the tool does not really make mistakes is not giving you useful information.</p><ol><li>Where does human review enter the process?</li></ol><p>Every AI-assisted workflow needs a defined review point before the output reaches a client, a court, or the public. The question is where that review happens and who is responsible for it.</p><p>For marketing content, review may need to cover both accuracy and advertising compliance. For client communications, tone and strategic appropriateness matter. For legal research, independent verification remains essential.</p><p>The review structure should be designed before the tool is adopted, not after. If no one can explain who reviews what and when, the system is already vulnerable.</p><ol><li>Who inside the firm owns the workflow?</li></ol><p>This is a question firms often overlook. Once an AI tool is adopted, who is responsible for it? Who monitors performance? Who refines the workflow when something is not working? Who communicates with the vendor?</p><p>AI tools are rarely “set it and forget it” products. They require ongoing attention. Workflows change. Outputs drift. Compliance expectations evolve. If no one owns the system, it will gradually move from useful to neglected.</p><p>Ownership should be assigned before the contract is signed.</p><ol><li>How will we know whether this is actually creating value?</li></ol><p>Before investing, define what success looks like in measurable terms. More signed cases. Faster response times. Lower acquisition costs. Greater content output without sacrificing quality. Less staff time consumed by repetitive administrative work.</p><p>Then measure against a baseline. Regularly.</p><p>Too many firms adopt AI with vague expectations and no clear way to determine whether the investment is paying off. When that happens, the firm often concludes that “AI did not work,” when the real problem is that no one ever defined what success was supposed to look like.</p><ol><li>What is vendor credibility actually based on?</li></ol><p>Not on the website. Not on the demo. Not on the logos.</p><p>Demos are designed to impress. They show the product under ideal conditions, with ideal inputs, and ideal outputs. That tells you something, but not enough.</p><p>Credibility comes from specificity. Can the vendor show verified results from firms like yours? Not just testimonials, but measured improvements with context. Does the vendor understand your practice area? A tool built for BigLaw contract review may be excellent for that purpose and still be irrelevant to a plaintiff firm running mass tort campaigns.</p><p>Does the vendor understand your compliance environment? Can they speak intelligently about the constraints your firm actually operates under?</p><p>And finally, are they candid about limitations? Vendors worth listening to usually know where their product works well and where it does not. That kind of honesty matters.</p><p>The goal is not to be impressed. It is to make a sound decision.</p><p>That is the framework. Eight questions. None of them requires deep technical expertise. All of them require something lawyers already use every day: disciplined judgment.</p><p>AI is not magic. It is a tool. And like every other tool a firm has ever adopted, its value depends on whether it was chosen for the right problem, implemented with the right safeguards, and measured with enough honesty to know whether it is actually working.</p><p>Choose carefully. Measure honestly. Adjust as you learn.</p><p>That is how a law firm adopts AI without getting lost in the hype.</p><h3>Chapter 11: Why Marketing May Be One of the Most Overlooked Functions in the AI Conversation</h3><p>When many attorneys think about AI in legal practice, they think about research, document review, and other work that begins after a client has already walked through the door.</p><p>Far less attention is paid to the work that gets the client there in the first place.</p><p>For plaintiff firms, that may be one of the biggest blind spots in the AI conversation.</p><p>Legal marketing has long been treated as secondary. It is funded, sometimes reluctantly, and often delegated to marketing teams or outside vendors. Rarely is it viewed as a core competency alongside trial work or legal analysis.</p><p>That is understandable.</p><p>Lawyers are trained to practice law, not to run campaigns. The profession values substance over promotion, and there is a natural belief that good work should speak for itself.</p><p>But good work cannot speak if it is difficult to hear.</p><p>In plaintiff practice, competition for cases is intense. Every new mass tort or MDL attracts multiple firms pursuing the same pool of potential clients. The firms that sign the</p><p>most cases are not always those with the strongest trial reputations. They are often the firms that reach the right people first, communicate clearly, and convert interest into signed retainers.</p><p>That is not a judgment. It is simply how the market works.</p><p>A brilliant lawyer without clients has one major problem: no pipeline. Marketing fills that gap, and AI is changing how that work can be done.</p><p>Today, AI can generate high-quality content across formats, personalize messaging for different audiences and case types, optimize campaigns in real time, and accelerate intake. Work that once required large teams or outside agencies can now be done faster and at lower cost.</p><p>And speed matters. In plaintiff practice, it often determines whether a lead becomes a client.</p><p>Of course, compliance with attorney advertising rules remains essential. But those obligations can still be met with appropriate supervision and safeguards.</p><p>So why does marketing remain overlooked?</p><p>Because research and analysis feel more like traditional lawyering. They align more naturally with the profession’s identity. Marketing does not carry the same prestige, so it is often deprioritized when firms consider where AI belongs.</p><p>For plaintiff firms, that ordering may be backwards.</p><p>Research and document review are post-acquisition functions. Marketing is a pre- acquisition function. It determines whether a client enters the system at all.</p><p>Improving research speed by 30 percent refines an internal process. Improving case acquisition can change the trajectory of a firm.</p><p>This is not an argument against using AI in substantive legal work. It is an argument that overlooking AI in marketing may carry a higher opportunity cost, particularly in contingency-fee practices where revenue depends on a steady pipeline of cases.</p><p>There is also a human dimension.</p><p>Plaintiff clients are often in crisis. They may be injured, uncertain, and searching for help. Many do not know whether they have a claim or where to find the right lawyer. They may be looking late at night, overwhelmed, and trying to make sense of what happened.</p><p>Seen this way, AI-powered marketing is not merely a business tool. It is also a way to reach people who might otherwise never find help. A well-designed campaign, combined with timely follow-up, can connect someone with counsel when it matters most.</p><p>When marketing is separated from the work of serving clients, it is worth considering who gets missed: the people who never found a lawyer, never received a callback, or never realized they had a viable claim.</p><p>Those missed connections have real consequences. AI gives firms new tools to close those gaps.</p><p>Closing those gaps is not just a business advantage. It is often the difference between someone finding help or facing a crisis alone.</p><h3>Chapter 12: AI-Generated Synthetic Performers in Legal Advertising</h3><p>Artificial intelligence has become one of the most significant developments in legal advertising. Much of the conversation focused on the risks. Will consumers be misled? Should synthetic performers be disclosed? How should regulators respond?</p><p>Those are important questions, and the legal landscape is beginning to take shape. In New York, for example, disclosure requirements governing synthetic performers in advertising took effect on June 9, 2026. As AI becomes more common, similar conversations will undoubtedly continue across the country.</p><p>But amid the debate over technology, an important point may be overlooked.</p><p>Synthetic performers are not introducing an entirely new concept to legal advertising. They are expanding one that has existed for decades.</p><p>The Longstanding Role of Dramatizations</p><p>Law firms have relied on dramatizations for many years to explain legal issues and communicate the experiences of injured people.</p><p>Sometimes a dramatization illustrates a legal concept using a fictional scenario. Other times it may combine elements from multiple cases into a single representative narrative.</p><p>There are also situations in which protecting a client’s privacy is essential. For example, survivors of childhood sexual abuse, victims of sexual assault, and survivors of human trafficking have compelling reasons not to appear publicly in an advertisement.</p><p>Regardless of the reason, the underlying principle remains the same.</p><p>A dramatization is a storytelling tool. It is intended to educate and illustrate, not to present documentary footage. Appropriate disclosures make clear that viewers are watching a dramatization rather than an actual event.</p><p>That model has served the legal profession well for decades.</p><p>Synthetic Performers Fit Within That Framework</p><p>Synthetic performers accomplish the same objective through a different method of production.</p><p>Rather than hiring a professional actor, a law firm can use artificial intelligence to generate a synthetic performer. The audience is still watching a dramatization. The message remains illustrative rather than documentary. The ethical obligation to communicate truthfully and disclose the nature of the presentation does not change.</p><p>In other words, although the technology changes how a dramatization is created, it does not change why dramatizations exist in the first place.</p><p>That distinction matters.</p><p>What AI really changes is accessibility.</p><p>Producing a professional dramatization with actors requires casting, production crews, filming locations, scheduling, insurance, editing, and substantial coordination. It is a process that often requires significant time and financial investment.</p><p>Those investments are often worthwhile, but they also create real constraints.</p><p>The cost and complexity of traditional production limit how often educational campaigns can be created, updated, translated, or expanded into new practice areas. They also limit how many firms can participate, and how many victims are ultimately reached.</p><p>That reality has consequences beyond advertising budgets.</p><p>Every educational campaign delayed by production timelines is another period during which someone suffering from abuse, neglect, discrimination, or another serious injury may remain unaware of their legal rights.</p><p>Every campaign that is never produced because of prohibitive costs is one less opportunity to inform someone that legal help may exist.</p><p>Viewed in that light, efficiency is not simply a market incentive.</p><p>It can also become a public benefit.</p><p>If synthetic performers allow law firms to create accurate, responsible, and clearly disclosed educational content more quickly and at lower cost, more people may encounter information that encourages them to seek legal advice.</p><p>The technology has not replaced the ethical model established by dramatizations. It has made that model more accessible.</p><p>Transparency Is Essential</p><p>None of this diminishes the importance of disclosure.</p><p>Transparency distinguishes a responsible dramatization from a misleading one, regardless of whether the performer is human or synthetic.</p><p>Viewers should understand when they are watching a synthetic performer. That disclosure builds trust, satisfies evolving legal requirements, and reinforces the profession’s longstanding commitment to truthful communication.</p><p>Professional judgment remains the safeguard.</p><p>Looking Ahead</p><p>The legal profession should evaluate synthetic performers for what they are: an evolution of an established practice.</p><p>When used transparently, synthetic performers allow law firms to produce responsible educational content using a different production method. In many cases, that method is faster, less expensive, and easier to update as laws, investigations, or public awareness efforts evolve.</p><p>Those efficiencies ultimately serve a larger purpose.</p><p>Artificial intelligence will change how legal advertising is produced.</p><p>It should not change why it exists.</p><p>At its best, legal advertising helps people recognize that they may have rights, remedies, and options they did not know existed. If synthetic performers allow that message to reach more people while preserving the profession’s commitment to truthfulness and transparency, then they represent something far more important than a new technology.</p><p>They represent another way to ensure that access to legal information evolves alongside the tools used to communicate it.</p><h3>Chapter 13: What Good Judgment Looks Like in an AI-Enabled Law Firm</h3><p>Over the course of this guide, I have tried to be useful: to explain artificial intelligence to practicing attorneys in a way that is honest, practical, and respectful of the profession.</p><p>I have written about where AI succeeds and where it fails. I have offered frameworks for evaluating risk, selecting tools, and building effective workflows. I have argued that</p><p>marketing deserves more strategic attention than many firms give it. Throughout, I have tried to be direct about what I know, transparent about what I do not, and careful not to overstate what the technology can deliver.</p><p>This final article is not really about technology.</p><p>It is about judgment.</p><p>The legal profession is pulling in different directions.</p><p>Some firms remain so skeptical of AI that they have stepped away from the conversation. In some cases, that skepticism comes from an early bad experience. In others, it reflects legitimate concerns about ethics, confidentiality, or reliability. Much of that caution is justified. But when skepticism becomes disengagement, it can become a strategic mistake.</p><p>Other firms have embraced AI with more enthusiasm than discipline. They adopt tools without written policies, enter sensitive information into systems they do not fully understand, publish content without adequate review, and move faster than their safeguards can support. That is not innovation. It is unnecessary exposure.</p><p>The firms most likely to navigate this transition successfully will take a different path.</p><p>A quieter one. A more disciplined one. A more sustainable one.</p><p>In an AI-enabled law firm, good judgment means discipline without paralysis.</p><p>It means establishing clear policies that define approved tools, what information may be entered, and who reviews AI-generated work. Not because the technology should be feared, but because professional responsibility requires thoughtful boundaries.</p><p>It means starting with lower-risk applications and building from there: workflow automation, marketing content with compliance review, or client intake communications with human oversight. A firm does not need to transform overnight. It needs to begin wisely, learn from experience, and expand deliberately.</p><p>It means assigning ownership. Someone should be responsible for AI governance, not as an afterthought, but as a defined role. That person monitors tools, evaluates new developments, adjusts workflows, and keeps firm leadership informed.</p><p>It means measuring results honestly. Success should be defined in advance, evaluated over time, and based on evidence rather than vendor claims or assumptions.</p><p>It also requires humility.</p><p>The legal and regulatory landscape surrounding AI is still evolving. Professional guidance continues to develop. Bar associations and courts are learning, and jurisdictions are not</p><p>moving in lockstep. Anyone claiming to have a complete map of where all of this is headed is overstating the case.</p><p>That uncertainty is uncomfortable for lawyers. We are trained to work within established rules, standards, and precedent.</p><p>But waiting for perfect clarity carries its own risk. Firms that delay until every question has been answered may find themselves years behind those that moved forward carefully, thoughtfully, and within the bounds of reasonable professional judgment. The goal is not perfection. It is reasonableness, caution, and informed judgement.</p><p>It also means respecting what came before.</p><p>AI cannot replace the qualities that make a good lawyer. It cannot read a jury, recognize when a witness is being evasive, exercise sound judgment under pressure, or build trust with a client through careful communication and genuine concern.</p><p>Those remain deeply human strengths. AI should support them, not replace them. At its best, it handles work that does not require those strengths so lawyers can devote more time to the work that does.</p><p>The firms that understand this will use AI to become better at what already makes them effective: responding more quickly so clients feel heard, communicating more consistently so fewer details fall through the cracks, reaching the right clients more effectively, and preparing more thoroughly so the legal work itself becomes stronger.</p><p>Not a replacement for the profession, but an amplifier of its best qualities.</p><p>The firms that get this right will not necessarily be the loudest or the fastest. They will be the ones that bring discipline, humility, and sound judgment to a rapidly changing environment.</p><p>In the end, that is not just what good AI governance looks like.</p><p>It is what good lawyering has always looked like.</p><h3>Chapter 14: Final Thoughts: The Profession Needs You in This Conversation</h3><p>I have been thinking about how to end this guide. Not because I have run out of things to say, but because I want the ending to matter.</p><p>Across these articles, I have tried to do something straightforward: take a technology that feels overwhelming, explain it honestly, and help practicing attorneys think about it more clearly. I have shared what I have learned. I have been transparent about what I do not know. And I have tried, at every step, to respect the profession, its traditions, and the people in it.</p><p>Whether I succeeded is for you to decide. But I want to use this final article to say something I believe matters deeply.</p><p>Right now, much of the conversation about AI in law is being driven by people who are not practicing lawyers. Technology companies. Consultants. Commentators. Academics. Nearly all of them are thoughtful and well-intentioned. But they do not carry the obligations lawyers carry. They do not answer to a bar. They do not face malpractice exposure. And they do not sit across from a client who is trusting them with one of the most difficult moments of life.</p><p>You do.</p><p>That is why your voice in this conversation matters.</p><p>Because the alternative is that the profession’s relationship with AI will be shaped primarily by the people building and selling the tools, rather than the people accountable for using them well.</p><p>Every generation of lawyers faces moments when the world changes faster than the rules can fully catch up. How the profession responds in those moments, whether it leads or follows, whether it engages or retreats, whether it brings its values to the table or leaves others to decide for it, says a great deal about what kind of profession it intends to be.</p><p>This feels like one of those moments.</p><p>This series was intended as a primer, not the final word. There is much more to explore, and I will return to many of these questions in greater depth.</p><p>I hope it has contributed something useful to you. It was written in the spirit of service to a profession I have spent my career in and care about deeply.</p><p>If it helped you think more clearly about even one important decision, it was worth writing.</p><h3>A Note on This Material</h3><p>Nothing in this guide constitutes legal advice, technology advice, or professional consulting, and no attorney-client relationship is created by reading or sharing it. The views expressed are solely my own. AI tools, rules of professional conduct, advertising regulations, and state requirements evolve quickly and vary by jurisdiction, so verify anything here independently before relying on it in your practice. References to any product or vendor are illustrative, not endorsements. For questions about your specific circumstances, consult a qualified attorney or technology professional directly.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=585800666995" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Antimicrobial Resistance: A Five-Year Investigation]]></title>
            <link>https://medium.com/@blakevanderhyde/antimicrobial-resistance-five-year-investigation-efc558e21754?source=rss-53e34460e4c0------2</link>
            <guid isPermaLink="false">https://medium.com/p/efc558e21754</guid>
            <category><![CDATA[antimicrobial-resistance]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[medical-devices]]></category>
            <category><![CDATA[healthcare-innovations]]></category>
            <category><![CDATA[public-health]]></category>
            <dc:creator><![CDATA[Blake Vanderhyde]]></dc:creator>
            <pubDate>Fri, 29 May 2026 20:18:42 GMT</pubDate>
            <atom:updated>2026-06-02T15:39:54.002Z</atom:updated>
            <content:encoded><![CDATA[<h3>A chance meeting with a U.S. Senator in a park drew a lawyer into a five-year campaign against antimicrobial resistance, one of humanity’s greatest emerging threats. This is the full account: the people, the science, the technology, and the hard truths about what could still go wrong. Twenty-seven parts, from beginning to end.</h3><figure><img alt="Four men standing together and smiling: from left, Michael Afremov, Senator Rudy Boschwitz, Blake Vanderhyde, and Dr. Eliaz Babaev, principals in the antimicrobial resistance work described in this series." src="https://cdn-images-1.medium.com/max/1024/1*LqNw1BXeeA8-dYbwGEo1Yw.jpeg" /><figcaption>From left: Michael Afremov, Senator Rudy Boschwitz, Blake Vanderhyde, and Dr. Eliaz Babaev.</figcaption></figure><p>For Scientific Discourse &amp; Hypothesis Only</p><p>This series is provided solely to stimulate open scientific discussion and paradigm thinking on antimicrobial resistance (AMR), a top-10 global health threat according to the World Health Organization. The personal stories and technical exploration are meant to humanize the urgency of AMR and spark rigorous scientific inquiry.</p><p>This series is not medical advice, device promotion, or an endorsement of any product, person or use. All references to specific technologies are illustrative and hypothetical only. Readers (including clinicians) must not use any device for uncleared indications.</p><p>IMPORTANT CLINICAL NOTE: The series discusses bactericidal resonance as purely a physics-based hypothesis for research purposes. Clinicians and facilities must not utilize the Qoustic System, or any similar low-frequency ultrasound device, for the purpose of destroying bacteria or treating infections. Such use is not FDA-cleared, and its safety for these specific indications has not been established in human clinical trials.</p><p>FULL DISCLOSURE: This series is provided for informational and educational purposes only. It does not constitute an offer to sell securities, a solicitation of interest in any securities offering, medical advice, or investment advice. The author is not a medical doctor or a registered investment advisor. Certain statements in this series may be forward-looking statements within the meaning of the Private Securities Litigation Reform Act of 1995. Forward-looking statements involve known and unknown risks and uncertainties that could cause actual results to differ materially, including but not limited to: failure of in vivo scientific validation, adverse FDA classification or pathway decisions, inability to secure development financing, unforeseen competitive developments, regulatory changes, and loss of key inventor knowledge. The technology discussed is currently FDA-cleared only for wound cleansing and debridement; the antimicrobial applications discussed here are purely investigational and are not FDA-cleared for antimicrobial indications. Readers are cautioned not to use any medical device for indications not cleared by the FDA. As of the date of publication, no legal entity has been formed or identified for purposes of any securities offering related to this technology. The author is not currently seeking, and will not accept, any inquiries regarding investment or financial participation. Any future financing, if any, would be conducted only through formal offering materials and in compliance with applicable federal and state securities laws. Do not rely on this content for investment, legal, or medical decisions.</p><h3>Table of Contents</h3><ul><li>Part 1 — About a Boy</li><li>Part 2 — The Statesman</li><li>Part 3 — The Builder</li><li>Part 4 — The Partnership</li><li>Part 5 — The Inventor</li><li>Part 6 — Our First Hurdle</li><li>Part 7 — The Memo</li><li>Part 8 — The Chief Financial Officer</li><li>Part 9 — MedTech</li><li>Part 10 — The Pitch</li><li>Part 11 — Lessons Learned</li><li>Part 12 — Press On</li><li>Part 13 — The Journey</li><li>Part 14 — The Living Legend</li><li>Part 15 — The Chief Technical Officer</li><li>Part 16 — The Consultants</li><li>Part 17 — Tachlis Time</li><li>Part 18 — Know Thyself</li><li>Part 19 — The Pivot</li><li>Part 20 — About a Dad</li><li>Part 21 — A Race Against Time</li><li>Part 22 — The Way</li><li>Part 23 — Physics v. Chemistry</li><li>Part 24 — Clinical Evidence</li><li>Part 25 — Navigating the FDA</li><li>Part 26 — What Could go Wrong</li><li>Part 27 — Final Thoughts</li></ul><p><strong>Why I’m Obsessed With AMR.</strong></p><h3>Part 1 — About a Boy</h3><figure><img alt="Senator Rudy Boschwitz and Blake Vanderhyde’s young son Hudson at a park overlooking Lake Minnetonka in 2020, where the chance meetings that began this story took place." src="https://cdn-images-1.medium.com/max/657/1*2yFjmprSK9QGZ2xnHKY14Q.jpeg" /><figcaption>Senator Rudy Boschwitz and my son Hudson at the park overlooking Lake Minnetonka, 2020.</figcaption></figure><p>In 2020, I was a proud new parent. My son Hudson arrived the previous year, and I was running on equal parts joy, fear, and exhaustion. Somewhere in the background, news was trickling in about a strange new virus. A ‘coronavirus’. I remembered SARS and thought it would fade just as quickly.</p><p>Then January 20th came, along with the first U.S. case in Washington State. The updates began to accelerate. CDC activated its emergency operations center. WHO declared a public health emergency. Even the name changed. “Covid-19.”</p><p>By March, Minnesota reported its first case. The governor declared a peacetime emergency. Schools, bars, and restaurants shuttered overnight. A stay-at-home order fell over the state like a heavy curtain. Hospitals braced for surge capacity as the economy jolted.</p><p>Then came Memorial Day, and the killing of George Floyd. Minneapolis ignited. Protests and riots swept the streets, a police station burned to the ground, and the National Guard deployed. It felt as though the world around me was unraveling.</p><p>By July, COVID was everywhere. Death tolls climbed, and collective mental health frayed. I found myself asking a question I’d never imagined: <em>What kind of world would my son inherit?</em></p><p>To cope, I took Hudson to a little park overlooking Lake Minnetonka. Every day it became our refuge, a place where the chaos of the world felt distant. Almost without fail, we would see the same older gentleman there with silver hair and a steady stride, walking the stairs for exercise.</p><p>After his exercise, he’d join us on a bench. We’d talk about life, family, and politics. His presence was grounding, making it feel as if the noise of the outside world couldn’t quite reach us there.</p><p>I didn’t know it then, but the man on that park bench would soon open doors for me and lead me into the fight against one of the greatest threats of our time. That gentleman was Senator Rudy Boschwitz.</p><h3>Part 2 — The Statesman</h3><figure><img alt="Image of former U.S. Senator Rudy Boschwitz, white-haired and wearing glasses, smiling warmly in a tan blazer. The statesman whose chance friendship drew Blake Vanderhyde into the antimicrobial resistance work" src="https://cdn-images-1.medium.com/max/410/1*ukyBABhnTJLl5bAioYHMjw.jpeg" /><figcaption>Former U.S. Senator Rudy Boschwitz</figcaption></figure><p>Walking into Senator Rudy Boschwitz’s office for the first time was humbling. The walls were lined with photos and memorabilia, each one a piece of history.</p><p>Above his computer hung a black-and-white photo of a ship, the SS Majestic. He told me how his family fled Berlin in 1933 after Hitler came to power. The Majestic carried them to the United States. As the ship entered New York Harbor, his mother pointed up at the Statue of Liberty and said, “We are safe now.” Of the family who remained in Europe, only one survived.</p><p>On the wall beside me: countless photos with historic figures. One of Rudy and his sons with President Ronald Reagan in the Oval Office. Another stepping off the Soviet Union’s presidential jet with Mikhail Gorbachev. Others with President George H.W. Bush, John McCain, and President George W. Bush at his Texas ranch. It felt surreal, like walking through a living museum.</p><figure><img alt="A wall in Senator Rudy Boschwitz’s office, densely covered with framed photographs and memorabilia from his decades in public life. Among the images are the Senator with Presidents Ronald Reagan, George H.W. Bush, and George W. Bush, with Senator John McCain, and with his wife Ellen, alongside other historic photographs" src="https://cdn-images-1.medium.com/max/468/1*KJoAFjL50bleEl1DyigyzQ.jpeg" /><figcaption>A wall in Senator Boschwitz’s office, lined with photographs from a life in public service, including images with Presidents Reagan, George H.W. Bush, and George W. Bush, and Senator John McCain</figcaption></figure><p>Mixed in were personal treasures. Photos of his beloved wife, Ellen. He told me marrying her nearly 64 years ago was the best day of his life. A framed portrait of his son Gerry, who passed away in 2018 after a courageous fight with cancer. Rudy’s eyes softened as he told me that was the hardest day of his life. I thought of my own son, Hudson, and shuddered at the weight of such loss.</p><p>Then one photo stood out from the rest: children crowded inside a stripped-down cargo plane. Rudy told me about Operation Solomon, the covert airlift he helped facilitate during Ethiopia’s civil war. In just 36 hours, over 14,000 Ethiopian Jews were rescued and flown to Israel. Babies were even born mid-flight. It remains one of the greatest humanitarian operations in history.</p><p>Rudy has led a big life: entrepreneur, Senator, ambassador, humanitarian, family man.</p><p>As we sat at his desk, I told him about an upcoming meeting with a Russian-born entrepreneur who had co-founded a billion-dollar MedTech company. He was now working with a scientist on a technology that might have the potential to fight COVID-19 and antimicrobial resistance.</p><p>Rudy leaned forward, intrigued.</p><figure><img alt="Senator Rudy Boschwitz, seated in his office, reads a clipped stack of papers, the memo Blake Vanderhyde prepared on Michael Afremov. A framed portrait of Abraham Lincoln hangs on the wall behind him." src="https://cdn-images-1.medium.com/max/981/1*rgWfzv-KskIKoli7p-nviQ.jpeg" /><figcaption>Senator Boschwitz reading the memo I prepared on Michael Afremov</figcaption></figure><h3>Part 3 — The Builder</h3><figure><img alt="Portrait of Michael Afremov, a dark-haired man with a mustache wearing an olive jacket and patterned tie. The entrepreneur and co-founder of AGA Medical Corporation, and Arobella Medical, a central figure in the antimicrobial resistance work." src="https://cdn-images-1.medium.com/max/512/1*xYE53t-O_z_ii5DIepL9Ng.jpeg" /><figcaption>Michael Afremov, co-founder of AGA Medical Corporation, and Arobella Medical</figcaption></figure><p>I first met Michael Afremov at his home on the shores of Lake Minnetonka. He walked me down a long hallway to his office, and we both took a seat. Behind his desk hung a portrait of his mother, Bella. She had survived the 900-day Siege of Leningrad in World War II, a military blockade that claimed 800,000 lives. Later, she and her husband brought Michael to America. The story sounded eerily familiar.</p><p>Michael was soft-spoken, with the quiet confidence of someone who had seen great success. I had done my homework. He had co-founded AGA Medical Corporation, which pioneered minimally invasive devices to repair heart defects, sparing countless patients from open-heart surgery. His company created billions in value and was eventually acquired by St. Jude Medical.</p><figure><img alt="A Wall Street Journal article headlined “St. Jude in Deal for AGA Medical,” dated October 19, 2010. It reports St. Jude Medical’s roughly $1.1 billion cash-and-stock acquisition of AGA Medical Holdings, the heart-device company Michael Afremov co-founded." src="https://cdn-images-1.medium.com/max/468/1*B2eD16VdAzBfBuDw8tEmEQ.jpeg" /><figcaption>A Wall Street Journal report on St. Jude Medical’s $1.1 billion acquisition of AGA Medical in 2010, the company Michael Afremov co-founded</figcaption></figure><p>He was also a philanthropist, giving the largest donation in history to Children’s Hospitals and Clinics of Minnesota. An impressive track record, no doubt.</p><figure><img alt="Three people standing together in a hospital lobby. From left: Lorie Afremov, Dr. Alan L. Goldbloom (center), president and CEO of Children’s Minnesota, and Michael Afremov. From a Children’s Hospital publication on the Afremovs’ $1.35 million donation in 2005 to build a pediatric cardiac surgery suite." src="https://cdn-images-1.medium.com/max/1020/1*YBsOwhfO7EFFyJjrqAbbtQ.jpeg" /><figcaption>Michael and Lorie Afremov with Dr. Alan L. Goldbloom, president and CEO of Children’s Minnesota (center), in the hospital lobby. The Afremovs donated $1.35 million in 2005 to build a pediatric cardiac surgery suite</figcaption></figure><p>We talked about the pandemic, and the growing uncertainty. Michael told me about a scientist he had co-founded another startup with years ago. That scientist had invented a device that used ultrasound waves to treat wounds. In published lab tests at the Mayo Clinic and Winston-Salem State University, this device had shown the ability to damage antibiotic-resistant bacteria (in electron microscopy and in vitro tests). Maybe, just maybe, this could be a tool to help fight the pandemic.</p><p>It sounded almost like science fiction. Using sound waves to fight bacteria, how could that possibly work? Could this technology be applied to a virus?</p><p>I didn’t know the answers. But Michael had built a billion-dollar company before. And if he believed there was something here, it was worth listening. I knew my next move. I needed to introduce him to the Senator.</p><h3>Part 4 — The Partnership</h3><figure><img alt="Three men standing in front of a garage with an American flag hanging behind them. From left: Michael Afremov, former U.S. Senator Rudy Boschwitz, and Blake Vanderhyde. The face mask around Vanderhyde’s neck and the blue medical glove on his hand mark the August 2020 pandemic setting at the early phase of the antimicrobial resistance work." src="https://cdn-images-1.medium.com/max/810/1*99MCcBe_YEMqdLor7APHUA.jpeg" /><figcaption>Michael Afremov (left), former U.S. Senator Rudy Boschwitz (center), and Blake Vanderhyde (right), August 2020</figcaption></figure><p>Michael and I arrived at the Senator’s home. When we first walked in, there was an impressive library filled with books on the wall, many of them biographies of great world leaders.</p><p>Ellen, the Senator’s wife, welcomed us warmly and led us to a four-season porch in the back, overlooking a blue lake framed by trees. I wore gloves and an N-95 mask, a reminder of the time, and the risks.</p><p>Michael opened his laptop and began a PowerPoint. He told us about his Second Act: a new start-up he had co-founded with another scientist, built around a technology that used low-frequency ultrasound.</p><p>The presentation was long and technical. Michael spoke of stable cavitation, wavelength parameters, bacterial shells, and virus capsids. These were terms that washed over me.</p><figure><img alt="A clinical wound-progression image showing three photographs of the same foot in sequence. Day 0 (left): the untreated wound, with toes already amputated due to necrotic tissue and the foot scheduled for amputation. Day 1 (center): the wound after debridement. Day 5 (right): continued healing. Documentation of the wound treatment described in the series." src="https://cdn-images-1.medium.com/max/1024/1*-uX1mqIaLmLVejePKhiSSA.jpeg" /><figcaption>A foot wound scheduled for amputation, toes already removed due to necrotic tissue, shown at three stages: Day 0 untreated, Day 1 after debridement, and Day 5.</figcaption></figure><figure><img alt="A clinical wound-progression image showing three photographs of the same foot at Days 37, 96, and 183. Day 37 (left): the wound is substantially smaller with scar formation. Day 96 (center): further healed, with only a small open wound remaining. Day 183 (right): the foot is largely healed with visible scar tissue, the limb salvaged. Documentation of the wound treatment described in the series." src="https://cdn-images-1.medium.com/max/1024/1*nrui8-p45DYYwpLMnPyW1Q.jpeg" /><figcaption><em>The same foot at Days 37, 96, and 183. The patient was able to salvage the limb.</em></figcaption></figure><p>Then came a slide that made the Senator and me lean forward: photographs of a foot ravaged by infection, several toes already gone. The prognosis had been amputation. According to the presentation, this technology was later used and the foot was ultimately saved.</p><p>Part of me was skeptical. It sounded too good to be true. Destroying bacteria with sound waves? Saving a man’s foot from amputation? But then Michael showed us studies including one from the Mayo Clinic and another from Winston-Salem State University. The studies suggested potential antimicrobial effects in laboratory settings. Maybe it was possible.</p><figure><img alt="A grid of four electron micrographs of methicillin-resistant Staphylococcus aureus (MRSA), at 40,000x magnification. Top row, control period: a scanning electron micrograph (left) of intact, smooth round cells in clusters, and a transmission electron micrograph (right) of intact dark cross-sections (scale bar 0.5 micrometers). Bottom row, experimental period: a scanning electron micrograph (left) of cells appearing disrupted, and a transmission electron micrograph (right) of cross-sections with" src="https://cdn-images-1.medium.com/max/1024/1*w2GW3ys-sQRkpCPwjI7g3A.png" /><figcaption>Scanning and transmission electron micrographs of methicillin-resistant Staphylococcus aureaus during the control and experimental periods (x 40,000). Journal of the American Podiatric Medical Association — Vol 97, №2, March/April 2007</figcaption></figure><p>Michael spoke of his co-founder fondly, another immigrant from Russia. “People throw this word around too much today, but he’s brilliant when it comes to ultrasound technology. He’s dedicated his life to it.”</p><p>The Senator looked over at me. “I think we ought to meet him.”</p><h3>Part 5 — The Inventor</h3><figure><img alt="Portrait of Dr. Eliaz Babaev, a gray-haired man in a dark suit and patterned tie. The Soviet-trained physicist and inventor of the Qoustic Wound Therapy System, the technology at the center of the antimicrobial resistance work in this series." src="https://cdn-images-1.medium.com/max/497/1*iJ_qxle_ETeniZ7G3wElEg.jpeg" /><figcaption>Dr. Eliaz Babaev</figcaption></figure><p>The Senator and I pulled up to a modest building tucked into an office park. Inside, we were greeted by Dr. Eliaz Babaev.</p><p>Born in Baku on the Caspian Sea, Eliaz was identified early as a prodigy. He rose quickly through Soviet academia, earning three advanced doctorates in record time, credentials that placed him among the scientific elite. One day during his research in Russia, Eliaz observed a phenomenon in physics that he did not fully understand. It happened in a split second, and defied conventional wisdom. Yet, it was repeatable.</p><p>Rather than see his discovery misused, Eliaz chose a different path. Penniless and barely speaking English, he brought his wife and small children to the United States. Once in the U.S., he rebuilt his career step by step: first as a design engineer, then a research scientist, later a chief technical officer, and ultimately a co-founder.</p><p>Despite being a prolific inventor and holding 100 international patents, Eliaz’s office was understated. And so was he. With a thick Russian accent and a stern manner, he reminded me of ‘Q’ from the old James Bond films. Like Q, his gadgets were deceptively ordinary-looking. Simple on the outside but hinting at something more beneath the surface. His English was functional, sometimes literal, and it made me lean in closely to catch every word.</p><p>He rolled out one of his inventions: the Qoustic Wound Therapy System, FDA-cleared in 2007. At first glance, it didn’t look like a device that could change medicine. There was a small control box, a portable IV stand with wheels, and a handpiece with a dome-shaped tip. When he powered it on, a fine mist sprayed into the air.</p><figure><img alt="In an office workspace, Dr. Eliaz Babaev (right) demonstrates the Qoustic Wound Therapy System to former U.S. Senator Rudy Boschwitz (left). The device’s control unit and a saline bag are visible on the counter in the foreground. Both men wear face masks, placing the meeting in the pandemic era during the early antimicrobial resistance work." src="https://cdn-images-1.medium.com/max/499/1*i4jSrIAB74aU4G4AuVc_VQ.png" /><figcaption>Dr. Babaev gives Senator Boschwitz a demonstration</figcaption></figure><p>I had looked into the device before the meeting. It was designed for wound care: cleaning chronic ulcers, burns, and infected tissue by removing debris, biofilms, and bacteria as part of standard wound management. It was especially important for diabetic patients, where infections often led to amputations. This was more than simple cleaning. It offered hope, a chance to preserve limbs and improve outcomes.</p><p>As Eliaz tried to explain the technology, how it worked, I tried to follow along, but much of this was foreign to me. I wasn’t a scientist, and Eliaz’s accent, combined with the complexity of the science, made it hard to grasp every detail. Still, one fact remained: this was the very device that had shown the ability to damage several types of antibiotic-resistant bacteria in preliminary laboratory studies. That left me with one important question: if it could affect bacteria, could it also have applications worth exploring against a virus?</p><h3>Part 6 — Our First Hurdle</h3><figure><img alt="The monument-style signage outside the U.S. Food and Drug Administration headquarters, displaying the seals of the U.S. Department of Health and Human Services, the U.S. Public Health Service (1798), and the FDA logo. The regulatory agency whose pathway is examined in the later parts of this series." src="https://cdn-images-1.medium.com/max/645/1*Eo0ubgSmXkdE8AploC8EQQ.jpeg" /><figcaption><em>The U.S. Food and Drug Administration.</em></figcaption></figure><p>Autumn was descending on Minnesota. Leaves were falling, and urgency was in the air. Hospitals were overwhelmed. Headlines were grim. An election loomed, and the country was on edge.</p><p>Amid this chaos, the Senator and I met Michael and Eliaz. They had developed a remarkable technology: sound waves that, in preliminary laboratory tests, had shown the ability to damage bacteria. Could this possibly have applications worth exploring in the fight against COVID-19?</p><p>The question quickly became: what now?</p><p>We devised a game plan. Cast a wide net and tell our story, again and again. The Senator tapped his deep network: fellow senators, White House advisors, NIH leaders, hospital CEOs. Doors opened. Meetings happened. Hope grew.</p><figure><img alt="Three figures stand in front of a red fire truck marked “Wayzata Fire Dept.” From left: former U.S. Senator Rudy Boschwitz, smiling in a plaid shirt and brown jacket, and Blake Vanderhyde in a dark jacket and cap, with his toddler son Hudson seated on his shoulders. An evening sky and the American flag are visible behind them. Wayzata Fire Station, October 2020, during the early phase of the antimicrobial resistance work." src="https://cdn-images-1.medium.com/max/810/1*lpWkHyx86ySg_vrF5MAqXQ.jpeg" /><figcaption>A hopeful Senator Rudy Boschwitz, Hudson, and Blake Vanderhyde at the Wayzata Fire Station, October 2020</figcaption></figure><p>Not everyone agreed on the path forward, but one thing was clear: we needed speed.</p><p>After much debate, we decided to apply for Emergency Use Authorization (EUA). If approved, Eliaz’s device could potentially be authorized for emergency use against COVID-19. We knew it was a long shot, but it was a shot worth taking.</p><p>We submitted our application. A microbiologist was assigned as our reviewer. He pressed us for more details, and we quickly delivered. For a moment, it felt like maybe, just maybe, we had a shot.</p><p>As the holidays approached, I took some time to be with my family. Then on December 20, my phone rang with news that nearly broke me. One of my best friends from college, Steve, had taken his life the night before.</p><p>I was in shock.</p><p>A month earlier, I had lost a cousin to suicide. Now I had lost Steve too. The pandemic was taking a heavy toll. On all of us.</p><p>The next day was the final blow. We received our response from the FDA. <strong>Rejected.</strong></p><p>The weight of grief, of loss, and now this rejection. It was too much.</p><p>That night I tucked my son Hudson into bed. I sat in his room quietly, watching him sleep. I was reminded of how fragile and precious life truly was.</p><figure><img alt="Three friends stand together in front of Ruby Hill Winery in Pleasanton, California. From left: Blake Vanderhyde in a cap and gray sweater, Steve in a blue shirt and plaid shorts, and Theresa in a pink cardigan and sunglasses. A sunny afternoon in 2015." src="https://cdn-images-1.medium.com/max/810/1*lkLCRtOssyhMWXXdOL8A9Q.jpeg" /><figcaption>Blake, Steve, and Theresa at Ruby Hill Winery, Pleasanton, California, 2015</figcaption></figure><p>Rest in peace Steve.</p><h3>Part 7 — The Memo</h3><figure><img alt="A wintry view of frozen Lake Minnetonka under soft golden light. The ice-covered surface stretches toward a distant treeline on the horizon, with patches of dark water visible through the thinning ice. A rocky shoreline with bare trees and a weathered dock sits at right. The Minnesota lake that recurs throughout this story." src="https://cdn-images-1.medium.com/max/1024/1*GdCa8HF4cmMD80JfB54VHg.jpeg" /><figcaption>Frozen Lake Minnetonka</figcaption></figure><p>It was January 2021. A new year, but I wasn’t celebrating. I sat by the window, watching snow drift silently to the ground. Outside, it was twenty below zero. Quiet, still, and unforgiving. The dead of winter in Minnesota is not for the faint of heart.</p><p>The sting of the EUA rejection was still fresh. Worse yet, I was grieving the loss of my cousin and my friend, Steve. In that silence, I realized I couldn’t stay in this place, not mentally, not emotionally. I needed a way forward.</p><p>That’s when I began drafting a long memo to the Senator. I called it <em>Hopes and Hurdles.</em></p><p>Our hopes were clear: to see a potentially life-saving medical technology properly vetted and, if successful, delivered into the right hands where it could be further evaluated. If successful, it had the potential to reshape aspects of medicine.</p><p>But the hurdles were just as real. Neither of us had walked this road before. He was a retired statesman. I was a mass tort lawyer. We had no knowledge or understanding of the medical device industry. But we <em>did</em> have persistence, networks, and the will to push forward.</p><p>Yes, Michael and Eliaz had invented something remarkable, but invention wasn’t enough. We needed allies, a global company with the resources and expertise to properly validate the technology through further research and, if supported by the data, to scale it. The question was: who?</p><p>Months passed. Spring eventually came, and with it, my son Hudson’s second birthday. He chose a Batman theme, his favorite superhero. Because of COVID, we kept the party small and outdoors in our backyard.</p><p>In the middle of the celebration, a surprise guest arrived: the Senator, carrying gifts.</p><figure><img alt="A backyard birthday party scene. Hudson, a young child in a blue shirt, sits in a white high chair with a toy and a wrapped gift on the tray while his Mother strokes his hair. Blake Vanderhyde, in sunglasses and a navy polo, kneels beside the high chair smiling. Former U.S. Senator Rudy Boschwitz stands at right in a plaid shirt and gray pants, smiling. Yellow, green, and blue balloons, wrapped presents, and a brown house with a patio umbrella in the background." src="https://cdn-images-1.medium.com/max/810/1*T-J7LW-5C07xUvwxi7VkPA.jpeg" /><figcaption>Hudson’s second birthday, in our backyard during the pandemic, with Senator Rudy Boschwitz</figcaption></figure><p>Later, he pulled me aside. There had been a breakthrough. He had spoken with a friend, the CFO of one of the world’s largest Medtech companies. And she had agreed to meet with us.</p><h3>Part 8 — The Chief Financial Officer</h3><figure><img alt="A bronze bust of Winston Churchill, depicting his head and shoulders with his characteristic firm-set jaw and stern expression. The sculpture sat in the home office of the CFO featured in this part of the series." src="https://cdn-images-1.medium.com/max/227/1*WRZKIoH3QMdXf9q9cSMo7w.jpeg" /><figcaption><em>A bronze bust of Winston Churchill that sat in the CFO’s home office.</em></figcaption></figure><p>I drove down a long oval-shaped driveway to a large stately home. At the door, the Senator and I were greeted by the Household Manager and ushered into the study.</p><p>A bust of Winston Churchill watched us carefully from the corner. As I sat down, a Goldendoodle bounded into the room, tail wagging, before being gently escorted out.</p><figure><img alt="A cream-colored, curly-coated dog with floppy ears, panting happily with its tongue out, photographed against a green outdoor background. A representative photograph of the same breed as the dog that joined the CFO meeting described in this part of the series." src="https://cdn-images-1.medium.com/max/624/1*n22usU2tPXhR0SUkQtXzZg.jpeg" /><figcaption>An Unexpected Guess</figcaption></figure><p>This wasn’t just another meeting, it was our shot at changing the trajectory of the project.</p><p>When the CFO entered, her presence filled the room. Sharp. Poised. Composed. Everything you would expect from a chief officer of a Fortune 500 company. I had done my homework, she had earned every bit of her title.</p><p>We laid our case out plainly: sound waves that, in preliminary lab tests, had shown the ability to damage antibiotic-resistant bacteria.</p><p>We wondered whether this might have any role to play against COVID-19. We mentioned the EUA rejection. Then we went deeper: a technology platform comprising more than 20 inventions and 40 patents focused specifically on low-frequency ultrasound.</p><p>She listened carefully, without interruption.</p><p>I couldn’t read her. Was she intrigued, skeptical, or just being polite?</p><p>Then she picked up the phone and called a Vice President of Business Development. He agreed to meet us at the corporate headquarters, in one week.</p><p>A new door had cracked open, and now we had one week to be ready.</p><h3>Part 9 — MedTech</h3><figure><img alt="A wide view of a modern corporate headquarters building with limestone construction, large windows, and a distinctive sloping roofline. Bare trees in the foreground, manicured lawn, and clear blue sky overhead. The MedTech company headquarters where the meeting described in this part of the series took place" src="https://cdn-images-1.medium.com/max/1024/1*VYZXTSKSR4YxIVQ8BTgmmw.jpeg" /><figcaption>MedTech Headquarters</figcaption></figure><p>I’m sitting at the headquarters of one of the world’s largest MedTech companies, a stunning campus with a sprawling green lawn, flanked by a serene lake on one side and a quiet pond on the other. We pass through security. The building was deserted. The pandemic had emptied offices. It’s just the Senator, Michael, Eliaz, the Vice President of Business Development, and me.</p><p>The VP was professional and personable. He and the Senator connected quickly over a mutual acquaintance. Then it was time for Michael and Eliaz to present the technology.</p><p>And that’s where we hit a speed bump. The science was real, but when Michael and Eliaz began to tell the story, it wasn’t landing. The roadmap was muddled. The business case lacked clarity. I watched as the momentum began to stall.</p><p>Still, one detail caught the VP’s attention: a prototype that could fit into his company’s existing strategy. He agreed to send a team to evaluate it further.</p><p>We had cracked the door open wider, but I could feel the gap. If we were going to be successful, we needed to find someone who could better explain the science and tell the story in a way that resonated.</p><h3>Part 10 — The Pitch</h3><figure><img alt="A marble sculpture titled Dying Achilles, depicting the Greek hero reclining in his final moments, his head turned upward in anguish as he reaches toward his fatally wounded heel. The statue stands in the gardens of the Achilleion Palace on the Greek island of Corfu, with the palace’s white neoclassical facade and cypress trees visible behind. An image Included in the doctor’s slide-deck presentation." src="https://cdn-images-1.medium.com/max/518/1*sGLDCAMK2AX_0HsANrSMUQ.jpeg" /><figcaption>Dying Achilles</figcaption></figure><p>It was the summer of 2021, and the Senator and I were filled with excitement. One of the world’s largest MedTech companies was flying in a team from across the country to meet with us in September. The stakes couldn’t have been higher.</p><p>Michael and Eliaz were brilliant, but when they presented their technology to the VP at the last meeting, something didn’t connect. The science was sound, but the story wasn’t landing. We needed a new approach.</p><p>The Senator tapped his deep network, and we caught another break. He found a doctor who knew Eliaz’s work firsthand. Not just any doctor, but someone who had taught at the Mayo Clinic and had delivered over a thousand scientific presentations. He was well respected at the crossroads of patient care, research, and education. For the first time in months, it felt like the stars were aligning.</p><p>Then suddenly, a blur of heartbreak.</p><p>July 12, 2021. Jed, a friend from high school and my college roommate, had died unexpectedly, leaving behind his wife and three kids.</p><p>Twelve days later, another gut punch: my law school roommate, also named Steve, passed away. He left behind his wife and two children.</p><p>Then August came, and with it a third and final blow. Megan, my wife’s best friend since childhood, was killed by a drunk driver at just 40 years old. She left behind her husband and two kids.</p><p>It felt like the ground beneath me was crumbling. This was too much.</p><p>Then the September meeting came. I was still grieving, but there was no time left to pause. The conference room was packed with various experts from the MedTech company. I introduced our team and sat back to watch.</p><p>The doctor delivered an excellent scientific presentation. He had command of the subject matter. He walked through the clinical studies, the foundational science, the twenty-plus inventions, the forty patents, a full ultrasound platform built over decades. A career’s worth of work.</p><p>But I read the room.</p><p>The visitors were not there for a platform. They were trained professionals, laser-focused on one thing, acquiring a very specific device, ready for the market.</p><p>And that device?</p><p>We didn’t have it.</p><figure><img alt="Three young men stand together on a ship’s deck, arms around each other, smiling. From left: Tony, Blake Vanderhyde (center), and Jed. They lean against a railing with the English Channel visible behind them under a gray sky. A photograph from 1995." src="https://cdn-images-1.medium.com/max/1024/1*n8n6ujAfBHlIojcYVfkVwA.jpeg" /><figcaption>Tony (left), Blake (center) and Jed (right). Crossing the English Channel, 1995.</figcaption></figure><figure><img alt="A group photograph of law students standing behind a large banner reading “Delta Theta Phi Law Fraternity, Ramsey-Amdahl Senate.” Steve appears in the back row, far left, and Blake Vanderhyde stands in the front row, center, in a suit and tie. From the Delta Theta Phi Law Fraternity initiation in St. Paul, Minnesota, 1996." src="https://cdn-images-1.medium.com/max/865/1*HhoPDWLGpYfsv8mA1kjKxw.jpeg" /><figcaption>Steve, (back row far left) and Blake (front row, center). Legal Fraternity Initiation, St. Paul, MN 1996.</figcaption></figure><figure><img alt="Three friends pose together for a group photograph in a dimly lit social setting. From left: Megan, Blake Vanderhyde (center), and Jessica, all smiling. A class reunion in Benson, Minnesota, 2010." src="https://cdn-images-1.medium.com/max/1024/1*KabGWcirZR3salyIDSU4zQ.jpeg" /><figcaption>Megan (left), Blake (center), and Jessica (right). Class reunion in Benson, Minnesota, 2010</figcaption></figure><p>For Jed. For Steve. For Megan.</p><h3>Part 11 — Lessons Learned</h3><figure><img alt="Hudson, a young boy in a blue police-officer costume hat, a Captain America t-shirt, and Paw Patrol sandals, sits on a concrete park step eating an ice cream cone. Beside him sits former U.S. Senator Rudy Boschwitz, smiling and watching him, in a colorful plaid shirt and khaki pants. The steps lead down from the park overlooking Lake Minnetonka where the friendship that began this story first formed." src="https://cdn-images-1.medium.com/max/1024/1*uZQEhyhVnLVexkrSQiXz3g.jpeg" /><figcaption>Ice Cream in the Park</figcaption></figure><p>We were learning our lessons the hard way, and fast.</p><p>Lesson number one. Large MedTech firms are not built to take early-stage bets anymore. They steer clear of risky, expensive R&amp;D. Costs are high, timelines are long, regulatory hurdles keep growing, and shareholder pressure demands predictable quarterly returns.</p><p>Instead of building technology in-house, the big players now acquire smaller startups after the hard work is done: once the science is proven, the regulatory path is cleared, and the product has real traction in the market. In other words, they swoop in when most of the risk is gone.</p><p>We had a promising science story, intriguing and validated in vitro, but it wasn’t what they were structured to buy. They wanted a finished device, not a vision. And despite the brilliance of the underlying science, we simply couldn’t meet them where they were.</p><p><strong>Lesson number two</strong>. We couldn’t give up. The stakes were too high. If developed and properly validated, this technology could push medicine forward and potentially help save lives on a massive scale. It was worth the fight, worth the sacrifice, so that Hudson’s generation might have a chance.</p><p><strong>Lesson number three</strong>. Along the way, I had lost people who meant the world to me. Tomorrow isn’t guaranteed.</p><h3>Part 12 — Press On</h3><figure><img alt="A black-and-white photograph of approximately twenty-five men in heavy winter expedition gear, posed in rows against the wooden hull of a ship trapped in ice. One man at left holds a small dog; a sled rests at bottom left. Ernest Shackleton and the crew of the Endurance during the Imperial Trans-Antarctic Expedition, 1915, after their ship became trapped in the Antarctic ice pack." src="https://cdn-images-1.medium.com/max/1024/1*QiZNZugknwN6MmUI60Bp9w.avif" /><figcaption><em>Ernest Shackleton and the Endurance crew, 1915.</em></figcaption></figure><p>We had been rejected twice. First by the government on our EUA request, and then by the major MedTech firm. If this technology was ever going to reach the world, it would have to be through venture capital. And we knew almost nothing about that world. The outlook felt bleak.</p><p>I sat in my office, staring out the window, trying to process our next move. On the wall beside me was a black-and-white photograph of Ernest Shackleton and the Endurance crew. Years earlier, after the hardest day of my life, I had hung some things up for inspiration. That picture was one of them. Next to it I had taped a quote, they were words from Calvin Coolidge: <strong>Press On.</strong></p><p>“Nothing in this world can take the place of persistence. Talent will not; nothing is more common than unsuccessful men with talent. Genius will not; unrewarded genius is almost a proverb. Education will not; the world is full of educated derelicts. Persistence and determination alone are omnipotent. The slogan ‘Press on’ has solved and always will solve the problems of the human race.”</p><p>I had read that quote hundreds of times. What I had never known was the backstory.</p><p>Calvin Coolidge’s son died at sixteen from blood poisoning after a blister on his toe became infected. It happened in July 1924, long before antibiotics existed. What if we returned to that world, where kids died from infected blisters?</p><figure><img alt="A 1924 newspaper page from Bethlehem, Pennsylvania, headlined “DEATH OF THE YOUNGEST SON OF THE PRESIDENT.” A portrait photograph of sixteen-year-old Calvin Coolidge Jr., son of President Calvin Coolidge, accompanies an article reporting his death from blood poisoning. The pre-antibiotic-era death referenced in this series on antimicrobial resistance." src="https://cdn-images-1.medium.com/max/640/1*o6TghQcT9RvKgoohnX-NbA.jpeg" /><figcaption><em>Calvin Jr. dies of blood poisoning, 1924.</em></figcaption></figure><p>The irony and the heartbreak were impossible to ignore.</p><p>I emailed the quote to the Senator. We could not give up. The stakes were too high. If fully developed, this technology could push medicine forward and save lives on a massive scale.</p><p>If it was ever going to reach doctors, nurses, and patients, we would have to learn the venture capital world from scratch.</p><p>Later, the Senator and I sat down and made a list of people to meet. It was a long one. Founders. Chief Scientific Officers. Investors. Anyone who might help us understand what venture capitalists were looking for and who might be willing to lead the charge.</p><h3>Part 13 — The Journey</h3><figure><img alt="A traditional Chinese painting of Lao Tzu, the legendary founder of Taoism, with a long white beard and flowing blue robes, riding a water buffalo through a misty mountain landscape. The classical image of Lao Tzu’s westward departure, when tradition holds he composed the Tao Te Ching, the foundational text of “the Way.”" src="https://cdn-images-1.medium.com/max/1024/1*ippzNTxIn7QzCodRQxlWTg.jpeg" /><figcaption>Lao Tzu</figcaption></figure><p>Learning the venture capital world felt like stepping into a foreign country without a map. Where do you begin? What do investors expect? I didn’t know. Not even a little.</p><p>Lao Tzu once said, “A journey of a thousand miles begins with a single step.” So I took mine.</p><p>I ordered two books. The Secrets of Sand Hill Road by Scott Kupor and Angel: How to Invest in Technology Startups by Jason Calacanis. That became the start of my crash course.</p><p>Then came the long nights. Podcasts, note-taking, and lists that kept growing and growing. Slowly, a framework emerged: clinical need, execution capability, intellectual property, regulatory strategy, market size, commercialization path. Looking back, I finally understood how much we had been missing.</p><p>Next came the deep dives with Michael and Eliaz. Eliaz’s patent portfolio alone was staggering. More than 1,000 pages of applications, diagrams, physics, and engineering. I had to build mind maps and entire websites just to stay organized.</p><p>Low-frequency ultrasound. Stable cavitation. Bacterial shells. Biofilms.</p><p>I repeated these terms until they no longer felt intimidating. I didn’t have all the answers yet, but I could feel myself moving forward.</p><p>Then the Senator and I started working through a list of people who might help us. His network was vast. We met founders, CTOs, board members, investors. People who had built companies, scaled them, sold them, even taken them public. We weren’t pitching. We were learning. Asking questions, gathering feedback and wisdom. Still searching for the right person to lead. It was a humbling experience, but slowly the fog was beginning to lift.</p><h3>Part 14 — The Living Legend</h3><figure><img alt="The entrance to Manny’s Steakhouse in downtown Minneapolis. A green neon “Manny’s” sign sits above a black canopy reading “Manny’s Steakhouse,” with glass doors set into a stone-faced building. The restaurant named for its founder, referenced in this part of the series." src="https://cdn-images-1.medium.com/max/556/1*dN99FayyIg7U00kSDY3HAA.jpeg" /><figcaption>Manny’s Steakhouse, downtown Minneapolis</figcaption></figure><p>The Senator had arranged a meeting with a serial medical-device entrepreneur. A man who had founded multiple major companies and reshaped modern medicine. A true living legend of medicine. Born in the South Bronx to immigrant parents, he had become world-renowned for pioneering technologies that saved countless lives.</p><p>We met him in a quiet restaurant. The host led us to his table, where he greeted us warmly. Over lunch, his presence carried both ease and authority. He had the unmistakable confidence of someone who had built entire industries from ideas. After the meal, he ordered a single scoop of vanilla ice cream for dessert.</p><figure><img alt="A scoop of vanilla bean ice cream in a clear glass footed dish, with visible vanilla bean specks throughout. The dessert ordered by the famous founder during the lunch described in this part of the series." src="https://cdn-images-1.medium.com/max/1024/1*Ou1qkiwI8xeTlvo2BVF8YQ.jpeg" /><figcaption>Vanilla Ice Cream</figcaption></figure><p>I took a breath and explained our situation: the rejections, the science, the urgency. Then I sat back and listened. He talked about the need for deep clinical insight, sharp business acumen and unshakable resilience. His questions were sharp, his insights immediate. He saw angles we hadn’t considered.</p><p>They say fortune favors the bold, so I asked him outright if he might consider getting involved. Maybe even lead the company. His technologies had become industry standards worldwide. If he said yes, our prayers would have been answered.</p><p>He paused, thoughtful. “If I were a younger man, I might,” he said. “But I’m in the middle of a new venture. I think you may have something here, but I’m not your guy.”</p><p>I couldn’t tell whether he was letting us down gently or simply being honest. Either way, all I felt was respect. Respect for what he had built, the lives his innovations had saved, and the ideas he had just shared with us.</p><p>As we stood to leave, a small group approached the table to greet the Senator. He welcomed them with his usual warmth. While he was momentarily distracted, the founder leaned toward me and spoke quietly, “The Senator is a very powerful spotlight for this project,” he said. “Use it wisely.”</p><h3>Part 15 — The Chief Technical Officer</h3><figure><img alt="A marble sculpture of Saint Jude Thaddeus, with curly hair, a beard, and flowing robes, gazing upward while holding a staff. From the Archbasilica of St. John Lateran in Rome. Saint Jude is the patron saint of difficult cases and hopeless situations." src="https://cdn-images-1.medium.com/max/1024/1*K_5PYq1MBC100Tkzwt-WRw.jpeg" /><figcaption><em>St. Jude Thaddeus, Patron of Difficult Situations</em></figcaption></figure><p>We met with the Chief Technical Officer in the Senator’s office. His name was spoken with quiet respect in medical circles. His résumé ran longer than my arm: a storied career in biomedical and electrical engineering, board roles, expert-witness work, and a tenure as CTO of one of the most iconic MedTech companies in the world. He was also a prolific inventor with more than 300 patents to his name.</p><p>I walked him through our journey and Eliaz’s technology. The potential that sound waves might hold for medicine. He didn’t sugarcoat anything.</p><p>“That’s one hell of a story,” he said. “But you’re not there yet.”</p><p>Then he did something invaluable. He explained what a successful roadmap actually looks like.</p><p>It required a functional, market-validated, disciplined approach to commercialization. A sharp regulatory strategy. A clear value proposition. A deep understanding of the market.</p><p>“Make sure you’re solving a real, urgent problem,” he said. “And tailor your message. Clinicians, procurement teams, administrators, and payers all care about different things.”</p><p>His advice was simple but powerful: do fewer things, better. Nail the core value before chasing anything new.</p><p>Finally, he explained the difference between what a technology does and what makes it hard to copy. The trade secrets, the know-how, the invisible layers that separate a breakthrough from a commodity.</p><p>He compared Eliaz’s technology to a Taser. Easy to understand, very difficult to reverse engineer.</p><p>His involvement would have been transformative, so I asked if he might join us, perhaps even lead the effort.</p><p>He smiled gently and shook his head. “I’ve got more money than time left in my life, so I’ll have to decline. But you definitely have something worth exploring.”</p><p>Then he wrote down a name and number. “Call these consultants. They can help you fill the gaps.”</p><p>I emailed the consultants the same day and got a quick reply. They could meet after the holidays. Hope was returning.</p><figure><img alt="A warm domestic scene in front of a stone fireplace decorated for Christmas, with four red stockings on the mantle and garlanded greenery. Hudson, a young boy in an orange hooded sweatshirt, concentrates on a yellow toy dump truck on the floor. Beside him, former U.S. Senator Rudy Boschwitz, in a gray jacket and glasses, leans down and places his hand on the truck, engaged in the moment. The Vanderhydes’ living room, a few days before Christmas" src="https://cdn-images-1.medium.com/max/1024/1*-m0wdM6mbwnLXBa0mfAP9A.jpeg" /><figcaption>wisdom and wonder</figcaption></figure><p>A few days later, as Christmas approached, the Senator stopped by my home unannounced. He brought Hudson a toy truck and a teddy bear. As they sat together on the couch, playing with the truck, it struck me how much time had passed and how long this journey had taken.</p><h3>Part 16 — The Consultants</h3><figure><img alt="A black-and-white group photograph of twenty-nine physicists arranged in three rows in front of a stone building. Albert Einstein sits in the center of the front row, with Marie Curie, the only woman in the photograph, also in the front row. Seventeen of the attendees were or would become Nobel laureates. From the Fifth Solvay Conference on Physics, held in Brussels in 1927, where leading physicists gathered to discuss quantum mechanics." src="https://cdn-images-1.medium.com/max/960/1*SiQG8GILIT09Bz2-6vu1xw.jpeg" /><figcaption>Solvay Conference, 1927</figcaption></figure><p>Their office was in downtown Minneapolis, a shell of its former self. Many storefronts were boarded up and a bullet hole marked the glass window of their office from nights earlier.</p><figure><img alt="A multi-story historic brick building on a corner of Minneapolis’s Warehouse District, featuring arched windows on the upper floors and classical pilasters. The streets are largely empty under an overcast sky. The neighborhood where the consultants described in this part of the series had their office." src="https://cdn-images-1.medium.com/max/725/1*3HaqbRAerQHFVL2ClpYE-w.jpeg" /><figcaption><em>Warehouse District, Minneapolis.</em></figcaption></figure><p>But the consultants stayed. That told me something.</p><p>I had done my homework. They were real operators with experience in both strategy and execution, who had helped major companies launch products around the world.</p><p>If anyone could help us close the gaps, it was them.</p><p>We met in a large conference room that overlooked the Warehouse District. They opened their presentation with a slide of the Solvay Conference, the historic gathering of the leading minds in physics and chemistry who met to confront the fundamental problems of their era.</p><p>They wove a masterful story from that beginning and carried it through the rest of their pitch. I was impressed.</p><p>When they finished, one of the consultants looked around the table and asked, “Who is the quarterback here?”</p><p>I hesitated.</p><p>He walked to the whiteboard and began sketching numbers. Costs, timelines, burn rates. Line after line. It was staggering to see how many millions would be required to commercialize a single device, let alone the twenty inventions in the portfolio.</p><p>Then he said, “You don’t need us. You need Jamie Dimon.”</p><p>He was right. We were biting off more than we could chew. Every expert we had met had hinted at this reality, and now it landed with full force.</p><p>When the meeting ended, we drove to Michael’s house to regroup and refocus.</p><h3>Part 17 — Tachlis Time</h3><figure><img alt="Four men standing together and smiling in a wine cellar setting, with bottles in storage racks visible behind them. From left: Michael Afremov, Senator Rudy Boschwitz, Blake Vanderhyde, and Dr. Eliaz Babaev. The principals of the antimicrobial resistance work, gathered for the meeting described in this part of the series." src="https://cdn-images-1.medium.com/max/1024/1*LqNw1BXeeA8-dYbwGEo1Yw.jpeg" /><figcaption>Michael Afremov (left), Senator Rudy Boschwitz (center left), Blake Vanderhyde (center right), Dr. Eliaz Babaev (right)</figcaption></figure><p>The Senator looked at me and said, “It’s <em>Tachlis Time</em>.” I had never heard the phrase before. He explained that it was Yiddish for getting down to brass tacks. Speak plainly. Face reality.</p><p>And reality was setting in.</p><p>The Senator and I had been bootstrapping this project for years, and it was leaving its mark.</p><p>I had lost several million dollars in time away from my mass tort practice. The Senator’s time was even more valuable. We had poured ourselves into this effort, and the runway was shrinking fast.</p><p>The consultants had only confirmed what others had hinted at all along: pitching an entire technology platform to venture capitalists would be too broad, too unfocused, and far too expensive.</p><p>Commercializing twenty inventions would demand institutional capital, a long horizon, and a team with deep resources. And time was the one thing we didn’t have. The Senator was in his 90s. Eliaz was almost 80.</p><p>If we didn’t narrow the plan now, we would risk watching the technology fade before it ever reached the world.</p><p>We needed to hire the consultants. We needed a new strategy. And we needed to speak the truth, to each other and to ourselves.</p><h3>Part 18 — Know Thyself</h3><figure><img alt="The Temple of Apollo at Delphi, an ancient Greek sanctuary with weathered limestone columns standing on a stone foundation, set high in the mountains of central Greece. The site of the Oracle of Delphi, where the maxim “Know Thyself” was inscribed." src="https://cdn-images-1.medium.com/max/960/1*fw7qE44M5YEZeAlOc3nq-w.jpeg" /><figcaption>Temple of Apollo (Delphi)</figcaption></figure><p>Inscribed on the Temple of Apollo are the words: <em>Know Thyself.</em></p><p>People interpret the phrase in different ways. Know your limits. Stay grounded. Make wise choices.</p><p>For me, the meaning became more personal. I kept hoping that if I stayed committed long enough, the project would finally attract the leadership it needed. That someone would step forward and take the helm.</p><p>But no one did. Too much was at stake to let the effort collapse, so I made the decision to hire the consultants and pay for them out of my own pocket.</p><p>The next day, I returned to the Warehouse District to begin an intensive round of onboarding sessions. I felt like I was under a microscope. I laid everything out on the table: the missteps, the rejections, the losses, and the lessons I had learned the hard way. I admitted that I was not a scientist or an engineer. I was a lawyer stepping into unfamiliar territory.</p><p>The consultants asked why I cared so deeply. I told them about the people I had lost. I told them how personal this mission had become, and how I felt like time was running out.</p><figure><img alt="Former U.S. Senator Rudy Boschwitz seated at a restaurant booth, smiling, in a plaid shirt and glasses. A slice of birthday cake with one lit candle sits on a white plate in front of him, alongside a glass of white wine. Through the window behind him, the partial words “GREEN WE” from a sign across the street are visible. From the Senator’s 93rd birthday dinner at a restaurant in Wayzata, Minnesota." src="https://cdn-images-1.medium.com/max/1024/1*LDw_k7xhGsUuNkufvvzUXw.jpeg" /><figcaption>Senator Rudy Boschwitz’s 93rd Birthday</figcaption></figure><p>They asked about my background, how my law practice worked and what my business model was. I explained it.</p><p>In the late 1990s, a disruptive technology was emerging: people were starting to use search engines instead of phone books to find lawyers. Large firms were slow to adapt, and I saw an opportunity. I taught myself search engine optimization, learned how to rank number one for key terms, and began generating thousands of cases.</p><p>I couldn’t litigate that volume of cases on my own, so I partnered with firms that had the capacity to handle them. I mitigated risk by diversifying case types and timelines and built systems to prevent competitors from reverse-engineering the model. It ultimately became one of the earliest SEO-based marketing networks for lawyers in the country.</p><p>One consultant smiled. “So you introduced a new disruptive technology, protected your trade secrets, mitigated risks, and scaled partnerships to commercialize it? Your weakness is your strength. You are an outsider, and that is your advantage. The skill set is the same, you’re just seeing the problem from a different angle.”</p><p>That is when it clicked. I had done this before, just in a different arena. My skill wasn’t SEO. It was arbitrage, system design and risk mitigation. It was understanding how complex, high-risk environments behave and building structures that turn complexity into opportunity.</p><p>Building a MedTech company is no different. You manage risk across regulatory, clinical, technical, and financial systems. The widget changes, but the architecture stays the same. I realized I had the tools to devise the strategy. I just needed the courage to tell the story publicly.</p><p>Over the next year, we held countless discussions. We continued to refine the strategy, mapped the landscape, and identified what we still needed to learn. There were plenty of laughs, and even a few tears.</p><figure><img alt="A modern preclinical surgical lab interior. A surgical table draped in blue sits in the foreground beside imaging equipment and a large medical monitor displaying vascular imagery. Supply shelving with blue bins lines the wall at right. A representative photograph of the type of preclinical lab toured in this part of the series." src="https://cdn-images-1.medium.com/max/1024/1*Wb--lWcgK6NrvCeD1UbfzA.jpeg" /><figcaption>Preclinical Lab in Minnesota</figcaption></figure><h3>Part 19: The Pivot</h3><figure><img alt="Hudson, a young boy seen from behind in a blue and turquoise life vest, grips the wooden steering wheel of a powerboat. Through the windshield, Lake Minnetonka stretches toward a wooded shoreline in golden evening light. The family with friends on a boat on a hot summer night." src="https://cdn-images-1.medium.com/max/880/1*ZRgm97bfmiPcmZ9AbeaVgQ.jpeg" /><figcaption><em>Hudson at the helm.</em></figcaption></figure><p>The consultants helped sharpen the vision. A broad technology platform was too diffuse. We needed to pivot. One focus, one clear direction. That is when everything finally began to take shape.</p><p>During this journey, another remarkable shift was happening in the background. Artificial intelligence was improving at a pace no one could have predicted. What felt impossible for me to accomplish just six months earlier suddenly became manageable.</p><p>I used AI to pressure-test the advice I had received from founders, CTOs, and consultants. I ran “red team” simulations against our regulatory strategy. I used it to synthesize and transform twenty years of complex physics and engineering data into clear financial models and commercialization paths.</p><p>AI didn’t think for me, it gave me leverage. It allowed me to operate at the speed of a 20-person team. It was a force multiplier.</p><figure><img alt="A stylized digital illustration: a humanoid robot in profile, with a white face and blue mechanical accents, extends its hand toward a glowing holographic medical symbol of a staff and serpent surrounded by geometric network lines, set against a blurred futuristic blue background. An AI-generated image illustrating the theme of artificial intelligence as a force multiplier in this part of the series." src="https://cdn-images-1.medium.com/max/1024/1*kXa5dKTmv4Jg6bJCn2hWTA.jpeg" /><figcaption>AI is a Force Multiplier</figcaption></figure><p>With that leverage came more clarity. I finally understood my role. I didn’t need to be the scientist or the engineer. My job was to connect the dots. To shape the narrative. To prioritize what mattered most. To focus the mission and put it into motion.</p><p>And now, I am going to do something a bit unorthodox, because it feels like we are running out of time. This story began with a little boy in the park. He is six now.</p><p>That is why I’m going to open source part of the strategy and the mission with all of you. But before I do that, there are two final things I need to share.</p><h3>Part 20 — About a Dad</h3><figure><img alt="A black-and-white high school graduation portrait of a young Thomas C. Vanderhyde. He has light hair neatly styled, wears a dark suit jacket and white collared shirt, and looks slightly off-camera with a soft smile. Blake Vanderhyde’s father, photographed at his graduation from Southwest High School in Minneapolis, Class of 1963." src="https://cdn-images-1.medium.com/max/1024/1*UFwVgrLWp2GzlKV7dBJBJQ.jpeg" /><figcaption><em>Thomas C. Vanderhyde. May 22, 1946 — October 12, 2001.</em></figcaption></figure><p>Before I share the strategy, you deserve to know a little more about me.</p><p>I have hinted at the hardest day of my life and how it shaped my perspective on this project. It does not feel good to be this vulnerable, but it feels necessary. So, I will take you back to that day.</p><p>October 12, 2001, was a beautiful, sunny autumn day in Minnesota.</p><p>I was driving home from work when my cell phone rang. It was my mom, weeping.</p><p>Months earlier, she had been diagnosed with breast cancer. She had undergone a double mastectomy, endured chemotherapy, and was fighting for her life.</p><p>Whenever she called, I answered.</p><p>Her words were broken and strained, but I caught them: “Your father shot himself… come quick.” Then the line went dead.</p><p>I sat in rush hour traffic trying to process what I had just heard.</p><p>When I finally arrived, several police cars were already there.</p><p>On the ground was my father’s body.</p><p>I knelt beside him, held his hand, and made him a promise.</p><p>I swore three things: that I would take care of my mom, that I would not let his death destroy me, and that I would find a way to turn this pain into something positive.</p><p>Words fail to fully capture an experience like that. How it feels. How it marks you. How it permanently changes the way you see the world.</p><p>I know what it is like to feel completely helpless.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/300/1*yOWLw9wkWzg1gobd24TREw.jpeg" /><figcaption>Show Me the Way</figcaption></figure><p>There is immense suffering in this world. I know many of you reading this have felt it.</p><p>Every year, antimicrobial resistance kills millions of people, many of them children.</p><p>If there is a chance to prevent some of that suffering, then I think this is a fight worth fighting.</p><figure><img alt="A color photograph from April 1973 of a joyful young family. A young woman with long brown hair in a white shirt with yellow trim, Blake Vanderhyde’s mother, sits laughing alongside a young man in a green collared shirt, Thomas C. Vanderhyde. Their toddler son Blake, wearing a Mickey Mouse t-shirt, sits on his mother’s lap and looks up at her. A birthday cake with a single candle waits on the table at right. Blake’s first birthday in Minneapolis." src="https://cdn-images-1.medium.com/max/842/1*vwtNyjhXGDdWr40Hx30spQ.jpeg" /><figcaption><em>My first birthday, Minneapolis, 1973.</em></figcaption></figure><p>Rest in peace, Dad.</p><h3>Part 21 — A Race Against Time</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SUZBLLNmJ0UHPurqFXvtpQ.jpeg" /><figcaption>Time</figcaption></figure><p>There is one final piece of the puzzle you need to understand before I share the strategy. It is the reason I still struggle to sleep at night.</p><p>We tend to think of technology as permanent. Once something is invented, we assume it belongs to the world forever. But that is not always true. Some breakthroughs are fragile. They do not live in textbooks or patents. They live in the mind of the inventor.</p><p>Dr. Eliaz Babaev is nearly eighty years old. Decades ago, while working in the Soviet Union at the height of the Cold War, he observed what he described as a phenomenon in physics. He learned how to fine-tune low-frequency ultrasound to create a resonance that may, among other things, be capable of damaging bacteria in laboratory settings.</p><p>It was a discovery Western physics largely missed. Not because the science was impossible, but because we lacked the art of the tuning.</p><p>The proprietary physics Eliaz carries remains largely uncodified, an intuition refined over a lifetime, residing almost entirely in his mind.</p><p>This is not a question of patents or intellectual property alone. It is a question of capability that is difficult to replace. Right now, I believe significant capability exists largely in one man.</p><p>We have a narrow and closing window to capture and preserve this knowledge.</p><p>It must be documented in a way that makes it measurable, teachable, and repeatable. From there, it could one day be translated into a scalable platform, should further research support it.</p><p>Achieving this will require some of the brightest scientific minds, working alongside AI and advanced modeling, to extract and formalize Eliaz’s intuition before it disappears forever.</p><p>The theoretical framework I am about to share may offer a path forward, but only through broad collaboration with the real heroes of this story: the scientists and engineers who advance it, the clinicians and regulators who evaluate it, and the healthcare professionals who determine its practical value.</p><p>And you, the reader. You are a part of this story too. You can help bring this vision to life by advancing the scientific conversation and contributing your own expertise to this global challenge.</p><p>But we must move quickly, we are racing against time.</p><p>All hands on deck.</p><h3>Part 22 — The Way</h3><figure><img alt="A Romantic-era oil painting titled “The Ninth Wave” by Ivan Aivazovsky, completed in 1850. A small group of shipwreck survivors cling to debris in the foreground as a massive turquoise wave rises behind them, set against a sky of brilliant orange and gold breaking through dark clouds. Held in the State Russian Museum in Saint Petersburg. An emblematic painting of hope through catastrophe." src="https://cdn-images-1.medium.com/max/468/1*yNpZRAhpkqPhgmcuolFc-Q.png" /><figcaption>Antimicrobial resistance is a gathering storm. The question is whether our world finds a way through.</figcaption></figure><p>This is Part 22 of an ongoing series exploring antimicrobial resistance and the urgent search for solutions.</p><p>If you are joining for the first time, the earlier articles trace the path that brought me here: from personal loss to a chance meeting with a U.S. Senator, to the discovery of a technology that may have the potential to change everything.</p><p>With this article, the series shifts toward purely academic and hypothetical territory, exploring possibilities that require rigorous and independent scientific validation beyond the scope of any single person or project.</p><p>Consequently, the articles become longer and more technical. I know that is not ideal in an age of shortened attention spans, so thank you for staying with me.</p><p>At the end of each article, you will find a disclaimer. I encourage you to read it.</p><p>Throughout this series, I have focused on one technology as an illustrative case study because it is the one I know best. But let me state this clearly: I refer to the Qoustic System as an example, not as an endorsement.</p><p>The scientific questions at issue are not limited to any one device. They extend to any acoustic or physical energy platform that may demonstrate similar mechanisms.</p><p>One final note: I realize I am not the most qualified person to articulate this vision. But if this series sharpens the public conversation around antimicrobial resistance and encourages more rigorous scientific inquiry, it will have served its purpose.</p><p>Let’s begin.</p><p><strong>The Silent Pandemic: Why Antibiotics Are Losing the War</strong></p><p>We need to talk about something most people are not paying attention to.</p><p>A pandemic is unfolding right now, but it does not make the evening news.</p><p>There are no press conferences, no emergency lockdowns, no daily case counts scrolling across the bottom of the screen.</p><p>It is quieter than that, but in some ways, even more dangerous.</p><p>The World Health Organization has identified antimicrobial resistance (AMR) as one of the top ten global public health threats.</p><p>By 2050, credible models predict up to 10 million deaths per year, more than cancer.</p><p>Other models estimate cumulative economic damage in the tens of trillions of dollars in lost global GDP.</p><p>Of course, the exact numbers are impossible to know with certainty, but the direction seems clear.</p><p><strong>What Exactly is AMR?</strong></p><p>Antimicrobial resistance (AMR) occurs when microorganisms such as bacteria, viruses, fungi, and parasites evolve mechanisms that allow them to survive the drugs designed to treat them.</p><p>As a result, our medicines become less effective, leading to more persistent infections and deaths.</p><p>In this series, I focus primarily on bacterial infections, and you will understand why later.</p><p>If this is your first time hearing about AMR, it is important to understand how serious this is.</p><p>The clock is running out, and one of our best weapons is failing.</p><p><strong>The Weapon That Changed Everything</strong></p><p>In 1928, Alexander Fleming discovered penicillin.</p><p>It is hard to overstate what that changed.</p><p>Infections that had once been death sentences suddenly became treatable.</p><p>Antibiotics did not just save individual lives, they made modern medicine possible.</p><p>Without them, there are no organ transplants, no chemotherapy, no joint replacements, and no cesarean sections without extraordinary risk.</p><p>Nearly every advanced medical procedure we rely on today depends on the assumption that we can control infection.</p><p>For decades, that assumption held.</p><p>Each time bacteria evolved resistance to one drug, the pharmaceutical industry developed another.</p><p>It was an arms race, and for a long time, we were winning.</p><p>We are not winning anymore.</p><p><strong>Why Are We Losing? Evolution.</strong></p><p>Here is the fundamental problem.</p><p>Bacteria evolve extremely fast.</p><p>A single bacterium can produce a million descendants in about seven hours.</p><p>Every time we deploy an antibiotic, we create a brutal form of natural selection.</p><p>Most bacteria die, but the ones that survive are left with an open field.</p><p>They multiply.</p><p>Within weeks, and sometimes days, a resistant population emerges.</p><p>This is not a flaw in our antibiotics, it is biology doing what biology does.</p><p>Bacteria have been evolving and adapting for billions of years. Every new chemical becomes another pressure to evolve around.</p><p><strong>Why Pharma Is Walking Away</strong></p><figure><img alt="An AI-generated infographic titled “The Failing Economic Model of Antibiotic Development.” A large funnel labeled “HIGH INVESTMENT ($1B+)” represents the pipeline of basic research, discovery, and Phase I, II, and III clinical trials. A single dollar-shaped drop emerges from the narrow end as the resulting new antibiotic, falling toward cracked ground labeled with the outcomes “SMALL ROI,” “RAPID RESISTANCE,” and “SHORT LIFESPAN.” An illustrative diagram of the economic challenges discussed." src="https://cdn-images-1.medium.com/max/1024/1*5A198pay5ei0g9pUV43Bvw.jpeg" /><figcaption>The broken pipeline</figcaption></figure><p>The pharmaceutical industry understands this dynamic, and increasingly, it is walking away.</p><p>Why?</p><p>Because developing a new antibiotic can cost more than $1 billion and take over a decade.</p><p>If resistance emerges within months of a drug’s release, the economic model collapses.</p><p>You cannot reliably recoup a billion-dollar investment on a therapy that may become obsolete before the patent expires.</p><p>For that reason, many major pharmaceutical companies have shut down or drastically scaled back their antibiotic research programs, at the very moment we need them the most.</p><p>We Have Another Problem: Biofilm</p><figure><img alt="An AI-generated stylized microscopy image depicting a bacterial biofilm. Clusters of golden-yellow spherical bacteria are shown embedded among red and pink membranous structures, set against a dark blue background. An illustrative visual of the biofilm communities discussed in this part of the series on antimicrobial resistance." src="https://cdn-images-1.medium.com/max/624/1*5sQZ3PoqriJR1UYKS7khBg.jpeg" /><figcaption>Bacterial biofilm</figcaption></figure><p>There is another layer of this that makes everything harder.</p><p>Many infections are not just free-floating bacteria.</p><p>In real life, bacteria build communities.</p><p>They form what are called biofilms.</p><p>Think of biofilm as a protective fortress that shields bacteria from the outside world.</p><p>That fortress can block antibiotics from reaching them.</p><p>It can slow bacterial metabolism, making drugs less effective.</p><p>It can even help bacteria share resistance traits.</p><p>This is why chronic wounds do not heal, why urinary tract infections keep coming back, and why implant infections are so devastating.</p><p>The bacteria are literally sitting behind a shield our drugs often cannot penetrate.</p><p>And here is the part that should keep you up at night: biofilms are everywhere.</p><p>They form on medical devices, implants, catheters, even wound beds.</p><p>In the petri dish, antibiotics can look like heroes.</p><p>At the bedside, they are failing us.</p><p>Today, there is no widely adopted clinical solution for reliably disrupting biofilms in human tissue, and that is a big problem.</p><p><strong>Where the War Is Being Fought: Six Battlefields</strong></p><p>In the United States alone, the CDC estimates that approximately 2.8 million people are infected with antibiotic-resistant bacteria each year.</p><p>That is nearly 40 football stadiums filled with people.</p><p>Globally, the picture is far worse.</p><p>Recent models place the figure in the tens of millions, although methodologies and definitions vary.</p><p>Let us assess the battlefield where most of the infections occur.</p><figure><img alt="An AI-generated anatomical illustration showing a human body in semi-transparency. The brain, heart, lungs, liver, intestines, and vascular system are all visible. A diagram of the six body systems most affected by antibiotic-resistant bacterial infections: the blood, lungs, urinary tract, stomach and intestines, brain and spinal cord, and skin and wounds." src="https://cdn-images-1.medium.com/max/864/1*7gKS6EAbzSf74z8eCJkhcA.jpeg" /><figcaption>Where the war is being fought</figcaption></figure><p>Antibiotic-resistant bacterial infections primarily affect six body systems:</p><p><strong>1) blood,</strong></p><p><strong>2) lungs,</strong></p><p><strong>3) urinary tract,</strong></p><p><strong>4) stomach and intestines,</strong></p><p><strong>5) brain and spinal cord, and</strong></p><p><strong>6) skin and wounds.</strong></p><p>Let’s take them one at a time.</p><p><strong>Blood</strong>. Many of you have heard of sepsis, a life-threatening condition often triggered by infection in the bloodstream. It kills more than 250,000 Americans each year and is one of the leading causes of death in U.S. hospitals.</p><p><strong>Lungs</strong>. Hospital-acquired pneumonia, especially ventilator-associated pneumonia, is increasingly caused by resistant organisms. Some patients survive the initial illness only to succumb to an infection that their immune system, and our drugs, cannot clear.</p><p><strong>Urinary tract</strong>. Resistant urinary tract infections are one of the fastest-growing categories of AMR. In elderly patients and those with catheters, these infections can escalate rapidly to sepsis.</p><p><strong>Stomach and intestines</strong>. Clostridioides difficile, or C. diff, kills nearly 30,000 Americans each year. Ironically, it is often triggered by antibiotic use itself, when the drugs wipe out beneficial gut bacteria and allow C. diff to flourish.</p><p><strong>Brain and spinal cord</strong>. Resistant meningitis and other central nervous system infections are rare, but catastrophic. When they occur, treatment options are severely limited.</p><p><strong>Skin and wounds</strong>. Chronic wounds, diabetic foot ulcers, venous leg ulcers, and pressure injuries. When these wounds become infected with resistant organisms such as MRSA or VRE, standard treatment can fail. The progression is grim: weeks of ineffective care, hospitalization, amputation, or death.</p><p><strong>The Human Reality</strong></p><p>I want to be careful not to let this become too abstract.</p><p>Numbers in the millions are hard to internalize, so let’s break this down to something more granular.</p><p>Imagine you are a 62-year-old diabetic.</p><p>You develop an ulcer on your foot.</p><p>Your doctor prescribes antibiotics, and they do not work.</p><p>The infection spreads.</p><p>Your doctor sits you down and explains the options, none of them good.</p><p>You either lose your foot, or you risk losing your life.</p><p>This happens thousands of times a day in hospitals across the country and around the world.</p><p>And the medical community, despite extraordinary effort and dedication, is running out of answers.</p><p><strong>The Question Not Enough People Are Asking</strong></p><p>The conversation around AMR in academia, in pharmaceutical boardrooms, and in government policy circles is typically framed around one question: How do we build a better drug?</p><p>That question assumes the answer is chemical.</p><p>A new molecule, a new compound, or a new biological agent.</p><p>But what if we are asking the wrong question?</p><p>What if the answer is not chemical at all?</p><p>What if, after billions of years of bacteria evolving to survive chemical assault, the solution is to stop playing on their terms?</p><p>What if we approached this problem through physics instead of chemistry?</p><figure><img alt="A Baroque oil painting depicting Alexander the Great in armor and red cape raising a sword to cut the Gordian Knot, while priests and soldiers look on in a classical temple setting. The moment from antiquity when Alexander refused the puzzle’s terms and solved the intractable problem by force." src="https://cdn-images-1.medium.com/max/900/1*0RIrwlL-TVvbIgaZT0p0QQ.jpeg" /><figcaption>The Gordian Knot. Maybe some problems are not solved by pulling harder on the same thread</figcaption></figure><h3>Part 23: Physics vs. Chemistry</h3><figure><img alt="An AI-generated image of Albert Einstein and Marie Curie arm wrestling at a wooden table labeled “PHYSICS” on Einstein’s side and “CHEMISTRY” on Curie’s side. Einstein sits before chalkboards with equations including E=mc²; Curie sits before laboratory glassware. A stylized visual of the physics versus chemistry framing in this part of the series." src="https://cdn-images-1.medium.com/max/832/1*O0vZmLxAFqJxtrD35pTdhQ.jpeg" /><figcaption>Physics v. Chemistry</figcaption></figure><p>I want to tell you about a scientist and inventor named Dr. Eliaz Babaev.</p><p>If you have been following this series, his name has appeared before. He is widely recognized as a leader in non-imaging therapeutic ultrasound. As I have stated previously, I am not endorsing Dr. Babaev or any specific commercial work. My goal is simply to encourage thoughtful, science-based discussion among qualified experts, of which I am not one.</p><p>To do that, it is helpful to examine a real-world example and carefully extrapolate from available data.</p><p><strong>The Idea: Sound as a Physical Force</strong></p><p>When most people hear “ultrasound,” they think of imaging. High-frequency sound waves bouncing off tissue to create pictures.</p><p>What Dr. Babaev works with is fundamentally different, it’s therapeutic low frequency ultrasound.</p><p>Instead of looking at tissue, it acts on tissue.</p><p>The Qoustic Wound Therapy System, the device he developed, operates at 35 kHz.</p><p>It delivers ultrasonic energy through a sterile saline mist via a disposable applicator tip. The energy passes through the mist and into the wound bed.</p><p>At that frequency, two things happen that may be significant.</p><p><strong>Acoustic Cavitation.</strong></p><p>The ultrasonic waves cause microscopic bubbles to form in the fluid at the wound surface.</p><figure><img alt="A stylized rendering of acoustic cavitation: a transparent dome containing a column of light descending into a bright cyan focal point, with concentric ripples radiating outward across a dark glossy surface. The image represents the physical mechanism." src="https://cdn-images-1.medium.com/max/1024/1*GVNXSasFjWcXt7oKZU4t-w.jpeg" /><figcaption>Acoustic Cavitation</figcaption></figure><p>These bubbles oscillate, they expand and contract in rhythm with the sound waves.</p><p>When they collapse, they generate intense, highly localized shockwaves.</p><p>The energy released at the point of collapse is extraordinary and may be sufficient to physically shear bacterial cell walls in a laboratory setting.</p><p>This is not a chemical reaction, no molecule is binding to a receptor, no enzyme is being inhibited.</p><p>It’s lysis through physics, brute mechanical force at the microscopic level.</p><p><strong>Biofilm Disruption.</strong></p><p>Remember the biofilm problem, the fortress that protects bacteria from antibiotics?</p><p>The mechanical energy generated from cavitation may do more than just kill individual bacteria, it may also disrupt and break apart the biofilm matrix itself.</p><p>If that protective shield is compromised, then the bacteria inside may be exposed.</p><p>This suggests that when ultrasound is used alongside antibiotics, it may accomplish something the drugs cannot do on their own: dismantling the barrier and opening the door for treatment to reach its target.</p><p><strong>The Critical Insight: Selective Cytotoxicity</strong></p><p>Here’s the part that makes this interesting: not all cells are the same, they differ in complexity and structure.</p><p>Bacteria are ‘prokaryotic’ organisms. They are a single-celled, no nucleus, relatively simple cell wall structure.</p><p>Human cells are different, they are eukaryotic. Vastly more complex, with reinforced membranes, internal organelles, and structural proteins that give them significantly more mechanical resilience.</p><p>What Dr. Babaev’s work suggests, and what the preliminary evidence appears consistent with, is the hypothesis that a ‘Therapeutic Window’ may exist, an energy density range where the ultrasound could be lethal to prokaryotic bacteria without significantly damaging eukaryotic human cells.</p><p>Defining this window would be the primary objective of any hypothetical pre-clinical program.</p><p>Think about what that possibly means. The ultrasound may be capable of being tuned to selectively destroy bacteria while leaving the patient’s tissue largely intact.</p><p>It’s like a frequency filter: pass through human cells, destroy bacterial cells.</p><p>And here’s the insight that may distinguish this approach from chemical strategies.</p><p>The scientific rationale suggests that bacteria may face fundamental structural barriers to developing resistance to a physical shearing mechanism.</p><p><strong>This of course is a hypothesis</strong>, it has not been confirmed in controlled resistance development studies.</p><p>A bacterium can do a lot of things: evolve a new protein to neutralize a chemical, modify the receptor of antibiotic targets, even develop efflux pumps to push drugs out of the cell.</p><p>These are biochemical adaptations, and bacteria have been perfecting them for billions of years.</p><p>But a bacterium may need to fundamentally alter its physical architecture to withstand mechanical shearing at 35kHz.</p><p>That structural vulnerability is intrinsic; it may not be the kind of surface-level feature that can be patched with a single mutation. Instead, it may simply be a consequence of being a single-celled organism without the reinforced architecture of a human cell.</p><p>If this hypothesis is validated, it would represent a meaningful shift in approach: away from trying to outsmart bacterial evolution with chemistry alone, and toward a physical mechanism that would try to exploit structural vulnerabilities that may be less easily modified by mutation.</p><p>Whether that potential advantage would prove durable over time would need to be demonstrated and confirmed through controlled studies.</p><p><strong>What We Already Know</strong></p><p>This isn’t entirely theoretical, there’s some data.</p><p>The Qoustic Wound Therapy System is FDA 510(k)-cleared for wound debridement, cleansing, and selective tissue dissection. It also holds CE Mark certification and is used in certain clinical settings for those cleared indications. Let me be 100% clear here: <strong>This device is NOT cleared for any bactericidal or antimicrobial claims.</strong></p><p>But two studies suggest something beyond the approved label may be worthy of further investigation with this or a similar system.</p><p><strong>The Mayo Clinic Study (2007).</strong> <strong>Use of Noncontact Low-Frequency Ultrasound in the Treatment of Chronic Foot and Leg Ulcerations: A 51-Patient Analysis — Steven J. Kavros, DPM; Erik C. Schenck, MPT. Journal of the American Podiatric Medical Association (Vol. 97, №2)</strong></p><p>In 2007, Dr. Steven Kavros and his team conducted a study involving 51 patients with chronic foot and leg ulcers, using Dr. Babaev’s earlier device, MIST Therapy, the predecessor to the Qoustic System.</p><p>The study found improved healing and closure rates for ulcers that had resisted other treatments.</p><p>But here’s the critical observation: the researchers noted antimicrobial effects and disruption of bacterial biofilms during treatment.</p><p>These weren’t the study endpoints, the researchers weren’t looking for antibacterial activity. They observed it.</p><figure><img alt="Four-panel electron micrograph of methicillin-resistant Staphylococcus aureus (MRSA). Top row: untreated control bacteria with intact spherical cells (SEM, left) and intact cell walls in cross section (TEM, right). Bottom row: same bacteria after acoustic treatment, showing surface disruption and debris (left) and ruptured cell walls and membranes (right). 40,000x magnification. 0.5 μm scale bars. Journal of the American Podiatric Medical Association, Vol 97, No 2, 2007." src="https://cdn-images-1.medium.com/max/1024/1*w2GW3ys-sQRkpCPwjI7g3A.png" /><figcaption>Scanning and transmission electron micrographs of methicillin-resistant Staphylococcus aureus during the control and experimental periods (x 40,000). Journal of the American Podiatric Medical Association, Vol 97, No 2, March/April 2007</figcaption></figure><p>And in their published conclusions, they wrote that further clinical and basic science investigations using this technology were warranted.</p><p>The Winston-Salem State University Study (2010). The Effects of Low-Frequency Ultrasound (35 kHz) on Methicillin-Resistant Staphylococcus aureus (MRSA) In Vitro, Teresa Conner-Kerr, PT, PhD, CWS, CLT. Wound Management &amp; Prevention (Vol. 56, Issue 5)</p><p>Three years later, researchers at Winston-Salem specifically tested the question.</p><p>Using the Qoustic Wound Therapy System in vitro, in petri dish studies, they examined the effects of 35 kHz low-frequency ultrasound on MRSA, one of the most dangerous and prevalent antibiotic-resistant bacteria.</p><p>The results within the constraints of the in vitro model were interesting: the ultrasound reduced bacterial numbers by inducing direct damage to bacterial cell walls.</p><figure><img alt="Multi-panel figure from a peer-reviewed study comparing the effects of 35-kHz ultrasound on bacteria and human cells. Figure 4 (SEM): MRSA showing few intact cocci after treatment versus numerous intact cocci in control. Figure 5: petri dishes showing zones of inhibition growing larger with 30, 60, and 180 seconds of exposure. Figure 6 (SEM): cultured human neurons, both control and ultrasound-treated panels showing intact neuronal structures." src="https://cdn-images-1.medium.com/max/1024/1*33RQySoNJxT8p8_bpUsLmA.png" /><figcaption>Wound Management &amp; Prevention (Vol. 56, Issue 5)</figcaption></figure><p>The authors also noted, consistent with standard scientific practice, that in vivo confirmation would be required before clinical conclusions could be drawn.</p><p>Two studies, two different institutions, both pointing in the same general direction: This technology may be doing more than cleaning wounds.</p><p><strong>The Gap</strong></p><p>And then progress stalled.</p><p>Despite these early signals, meaningful follow-up work, including robust preclinical dosimetry studies, animal models, and controlled human trials, was never conducted. The in vivo validation called for by the Winston-Salem researchers did not materialize.</p><p>If you’ve been reading this series, you know this is the part of the story that pulled me in and kept me there for years: the technology itself is not purely theoretical.</p><p>The device is already FDA-cleared for safe use in wound beds for its labeled indications. The central open question remains whether the bactericidal effects observed in vitro can be reliably translated and quantified in real clinical wounds.</p><p>That gap, from early feasibility observations to a rigorously validated therapy, is substantial. Closing it would require careful preclinical mapping of the therapeutic window, engagement with regulators, and well-designed clinical studies.</p><p>Any development program would also face the full range of scientific, operational, and regulatory uncertainties inherent in medical device innovation.</p><h3>Part 24: Clinical Evidence. What It Would Take to Prove This Works.</h3><figure><img alt="An AI-generated graphic introducing the section “Clinical Evidence: What It Would Take to Prove This Works.” A stylized hand at the center emits curved waves labeled “35 kHz.” A panel on the right lists three evaluation criteria: Before → After, 2-log reduction, Therapeutic Window. Dark blue background." src="https://cdn-images-1.medium.com/max/1024/1*sBOSkZHADKfkf-H6HJLxiA.jpeg" /><figcaption>A thought experiment, in clinical terms</figcaption></figure><p>There’s a moment in every serious scientific effort where you have to stop talking about potential and start talking about evidence, and this is that moment.</p><p>Over the last several articles, I have laid out the rationale for why low-frequency ultrasound may represent the possibility for a breakthrough approach for antibiotic-resistant infections. The physics is compelling, early evidence is encouraging, and the need is urgent.</p><p>But none of that matters until someone can show that it works in a human being, under controlled conditions, with quantified data that regulators and clinicians can evaluate.</p><p>So let me lay out what that might look like in a hypothetical illustration, in the kind of specific detail a serious clinician, regulator, or engineer would expect to see. Remember, this is strictly a thought experiment only.</p><figure><img alt="Three-step infographic “From Hypothesis to Evidence” showing the validation journey for low-frequency ultrasound in antibiotic-resistant wound infections. Step 1, Compelling Hypothesis: physics rationale, early preclinical data, Winston-Salem study. Step 2, Rigorous Validation Gauntlet: multi-year industry sequence, preclinical safety, IDE Early Feasibility Study (~30 patients). Step 3, Initial Human Proof-of-Concept: bacterial load reduction, biopsy data, foundation for pivotal trials." src="https://cdn-images-1.medium.com/max/1024/1*-d8EIlazd7VFJ6dm8PVNLA.jpeg" /><figcaption>A thought experiment in three stages</figcaption></figure><p><strong>What Someone Would Need to Show</strong></p><p>The clinical question is straightforward:</p><p><strong>Does 35 kHz low-frequency ultrasound reduce the bacterial load in chronic wounds infected with antibiotic-resistant organisms?</strong></p><p>Not “does it clean wounds.” That is already covered by the existing 510(k) clearance.</p><p>Not “does it disrupt biofilms in a petri dish.” Winston-Salem explored that.</p><p>The question is whether this device (or one like it) applied to a real wound in a real patient with a confirmed resistant infection, produces a measurable and statistically significant reduction in bacterial burden.</p><p>If the answer is yes, and if that answer comes from a well-designed, properly controlled study, it could materially change how this technology is evaluated.</p><p>If such data were generated, it could potentially support a regulatory submission seeking a new intended use, possibly making this one of the first FDA-cleared physical mechanisms for addressing antibiotic-resistant bacterial infections in wounds.</p><p>That would be a scientifically and clinically meaningful development, subject, of course, to FDA agreement on the pathway and the data required.</p><p><strong>A Generic Industry Model for Clinical Validation</strong></p><p><strong>The following timeline is a hypothetical/generic illustration only. It is intended to show the logical sequence of a validation program. It does not represent a committed development schedule, a funding plan, or any guarantee of outcome.</strong></p><p><strong>A Pre-Clinical Foundation</strong></p><p>Before testing in humans, the safety parameters would need to be established in animal models.</p><p>This isn’t bureaucratic box-checking, this is where you answer the most important safety question: <strong>Where exactly is the Therapeutic Window?</strong></p><p>As discussed in the last article, the concept of selective cytotoxicity suggests that ultrasound may be able to damage bacteria without materially harming human tissue, but only within a specific range of energy densities.</p><p>Too low, and there may be no meaningful bactericidal effect.</p><p>Too high, and healthy tissue could be injured.</p><p>The preclinical program would aim to map that window as precisely as possible.</p><p><strong>Porcine models</strong></p><p>Pigs are often considered the gold standard for wound research because their skin more closely resembles human tissue in thickness, healing response, and overall wound behavior.</p><p>Researchers could create standardized wound infections using resistant organisms such as MRSA, VRE, and CRE, then apply ultrasound at varying energy levels.</p><p>At each level, they would measure two things: bacterial reduction and tissue viability.</p><p>The result would ideally be a dosimetry map, a dataset showing which energy densities appear to produce meaningful bacterial reduction without unacceptable collateral tissue damage.</p><p>That map would help inform the operating parameters for any future human study.</p><p><strong>Expanded in vitro testing</strong> would also likely happen during this phase.</p><p>The Winston-Salem study focused on MRSA. Additional data would likely be needed on organisms such as VRE, CRE, and Pseudomonas aeruginosa.</p><p>These are among the most dangerous and prevalent resistant pathogens. Demonstrating activity across multiple species could strengthen the argument that this is a broader physical mechanism rather than a narrow, organism-specific effect.</p><p><strong>The Human Trial: IDE Early Feasibility Study</strong></p><p>This would be the centerpiece.</p><p>An Investigational Device Exemption, or IDE, Early Feasibility Study would provide a potential FDA pathway for limited human testing of a device for a new, uncleared indication.</p><p>Here is one way such a study might be structured:</p><p><strong>Design: </strong>Prospective, randomized, 1:1 allocation, single-blind.</p><p>Single-blind means the patient does not know whether they are receiving active treatment or sham treatment. Ideally, clinicians assessing biopsies and endpoint data would also be blinded to treatment assignments where feasible.</p><p><strong>Sample Size:</strong> 30 patients. Fifteen in the treatment arm and fifteen in the control arm.</p><p>This would be a small study by design. An Early Feasibility Study is not meant to establish definitive efficacy across a broad population. It is meant to generate initial human proof-of-concept data and help shape the design of a larger pivotal trial.</p><p><strong>Patient Population:</strong> Adults with chronic lower-extremity ulcers, such as diabetic foot ulcers or venous leg ulcers, with laboratory-confirmed infection by a resistant organism such as MRSA or VRE.</p><p>These would be patients for whom standard therapy has proven inadequate or incomplete.</p><p><strong>Treatment Arm:</strong> Standard of care, including wound care protocols and antibiotics as clinically indicated, plus 35 kHz ultrasound treatment applied according to the dosimetry protocol established in preclinical testing.</p><p><strong>Control Arm:</strong> Standard of care plus a sham procedure. The device would be positioned identically, and saline mist would be delivered, but the ultrasound energy would not be activated.</p><p>This would help control for placebo effects and for any therapeutic contribution from saline irrigation alone.</p><p><strong>Primary Endpoints:</strong></p><ol><li><strong>Safety.</strong> Adverse event monitoring throughout the study period.</li></ol><p>In an Early Feasibility Study, safety is the first concern, and rightly so.</p><ol><li><strong>Bacterial Load Reduction.</strong> Quantitative wound biopsies obtained before treatment and at defined intervals afterward, measured as log10 colony-forming units per gram of tissue.</li></ol><p>A 2-log reduction, meaning a 100-fold decrease in bacterial count, could represent a clinically meaningful proof-of-concept signal, though the precise threshold would likely need to be discussed with FDA and clinical investigators.</p><p><strong>Secondary Endpoints:</strong></p><p>Wound closure rate.</p><p>Time to infection resolution.</p><p>Antibiotic utilization, including whether patients in the treatment arm required fewer or shorter courses.</p><p>Patient-reported outcomes.</p><p><strong>Why This Study Design Matters</strong></p><p>I want to pause here and explain why these design choices are strategically important, because they reflect lessons learned by others over decades of device development.</p><p><strong>The sham-controlled design.</strong> Many device studies use an active comparator, a different device, or no meaningful control at all.</p><p>A sham-controlled trial is far more rigorous.</p><p>It reduces the possibility that any observed benefit is simply the result of wound care protocols, saline irrigation, or patient expectation.</p><p>When regulators review the data, the causal question becomes much clearer: either the ultrasound contributed to the bacterial reduction, or it did not.</p><p><strong>The quantitative biopsy.</strong> Wound appearance is subjective.</p><p>A physician saying “it looks better” is not enough.</p><p>A quantitative tissue biopsy measuring bacterial colony counts is objective, reproducible, and far more persuasive.</p><p>The bacterial burden goes down, or it does not.</p><p><strong>The “adjunct therapy” protocol.</strong> Notice that the treatment arm receives ultrasound plus standard of care, not ultrasound instead of standard of care.</p><p>That matters for two reasons.</p><p>First, ethically, proven therapies cannot simply be withheld from patients with serious infections.</p><p>Second, strategically, it creates a more realistic and achievable clinical question.</p><p>The device does not need to cure the infection by itself. It needs to demonstrate that it improves outcomes compared with standard care alone.</p><p>That is both clinically relevant and more consistent with how the technology would likely be used in practice.</p><p><strong>The Milestone Timeline</strong></p><p>To move from a physics-based hypothesis to a clinical reality, any technology in this space must survive a rigorous, multi-year validation gauntlet. The following framework outlines the standard industry milestones required to prove safety and efficacy:</p><p><strong>Phase 1: Strategic Regulatory Alignment.</strong> This phase begins with a Breakthrough Device Designation request submitted to the FDA, potentially leveraging the public health significance of antimicrobial resistance.</p><p>If granted, this could support prioritized interaction with the agency, though it would not lower the evidentiary standard.</p><p><strong>Phase 2: Preclinical Validation.</strong> This stage involves rigorous preclinical porcine studies and expanded in vitro testing to define the therapeutic window.</p><p><strong>Phase 3: Formal Regulatory Consultation.</strong> This critical milestone involves a Q-Submission (Pre-Sub) meeting with the FDA to align on study endpoints, sample size, and evidence expectations before IDE submission.</p><p>This step is critical. One of the most common mistakes in device development is running a study that does not answer the questions regulators actually care about.</p><p><strong>Phase 4: Investigational Device Authorization.</strong> This phase centers on the formal IDE submission to the FDA to authorize human clinical trials.</p><p><strong>Phase 5: Early Clinical Feasibility Study.</strong> This is the first-in-human stage, typically involving enrollment and treatment of approximately thirty patients to establish proof-of-concept.</p><p><strong>Phase 6: Data Synthesis and Decision Gate.</strong> At this final industry milestone, the resulting evidence is synthesized to determine if the data supports advancing to larger pivotal trials.</p><p>Six phases from initiation to the first meaningful human data that could help determine whether this technology might one day affect the standard of care for antibiotic-resistant wound infections.</p><p><strong>What Successful Data Could Potentially Enable</strong></p><p>If an Early Feasibility Study of this kind were successful, and if it showed meaningful bacterial load reduction with acceptable safety, several scientific and regulatory paths could potentially begin to open.</p><p>It could also become the scientific foundation for what follows: pivotal trials, expanded indications, and global regulatory strategy.</p><p>This first study would not need to prove that the technology works for every infection in every body system.</p><p>It would need to answer a narrower and more important question: Does the mechanism appear to work in one specific clinical context, in human beings, under controlled conditions?</p><p>If that were established, additional questions across other tissues, indications, and clinical settings could then be explored.</p><p>And perhaps most importantly, successful data would change the conversation.</p><p>Right now, low-frequency ultrasound for AMR remains a compelling hypothesis supported by preliminary evidence.</p><p>After a successful Early Feasibility Study, it would become something more concrete: an evidence-based clinical concept supported by initial human data.</p><h3>Part 25: Navigating the FDA — What Would a Hypothetical Regulatory Roadmap Look Like</h3><figure><img alt="An AI-generated graphic for Part 25 titled “Navigating the FDA: What Would A Hypothetical Regulatory Roadmap Look Like?” A maze illustration with sequential labeled steps: Existing medical wound therapy system, Pre-Sub meeting, Breakthrough Device Designation, De Novo submission, leading to a circular FDA Clearance seal marked “New Intended Use: Antibiotic-Resistant Infections.” “35 kHz” appears twice in the background." src="https://cdn-images-1.medium.com/max/1024/1*VskyrgHOA5FOh8-WXevmQw.jpeg" /><figcaption><em>A thought experiment in regulatory terms</em></figcaption></figure><p>People outside the medical device world tend to think of the FDA as a single, monolithic gatekeeper. You submit your application, you wait, and eventually you get a yes or a no.</p><p>It doesn’t work that way.</p><p>The FDA is not one door. It’s a maze of doors, each leading to a different pathway, each with different requirements, timelines, and costs.</p><p>Choosing the right pathway is one of the most consequential strategic decisions in medical device development. Get it wrong and you can lose years of time and substantial development resources pursuing a path that does not fit.</p><p>For a technology like low-frequency ultrasound applied to antimicrobial resistance, pathway selection would not be a minor administrative detail.</p><p>It would be a major scientific, regulatory, and operational decision. The choice would shape the data required, the clinical program, and the realistic timeline to any potential clearance.</p><p>Let me walk through how that could work in theory.</p><p><strong>The Starting Position</strong></p><p>In this hypothetical, you would need an ultrasonic wound therapy system (like the Qoustic) that has an FDA 510(k) clearance. The 510(k) is the most common pathway for Class II medical devices, for wound debridement and cleansing.</p><p>That matters.</p><p>It means the FDA has already reviewed the device for safety and performance within its current indication. It also means this would not begin as a completely unreviewed technology.</p><p>The question would not be whether the FDA is seeing the hardware for the first time.</p><p>The question would be whether the agency would accept a new intended use for a device it has already cleared.</p><p>The hardware is not the novelty.</p><p>The novelty is the claim that it may reduce or destroy antibiotic-resistant bacteria.</p><p>That distinction matters enormously in regulatory strategy.</p><p><strong>Why Not Just Amend the 510(k)?</strong></p><p>The 510(k) pathway works by showing that a device is substantially equivalent to a predicate device already on the market.</p><p>That pathway is generally faster and less burdensome than others.</p><p>But there is an obvious problem here: there appears to be no predicate device cleared for the destruction of antibiotic-resistant bacteria in wounds using this kind of physical mechanism.</p><p>If no suitable predicate exists, then a traditional 510(k) route for that intended use may not be available.</p><p>And paradoxically, that may not be bad news.</p><p><strong>The De Novo Pathway</strong></p><p>When a device is novel, either because the technology is new or because the intended use does not fit an existing predicate category, the FDA may consider the De Novo classification pathway.</p><p>De Novo is how a new regulatory category can be created. You are not fitting into an existing box. You are asking the FDA to define a new one.</p><p>For a bactericidal ultrasound indication applied to wounds, De Novo may be the most logical pathway if the FDA agrees that the risk profile fits Class I or, more likely, Class II with Special Controls.</p><p>That is an important qualifier: the FDA makes the final classification decision.</p><p><strong>Likely Classification:</strong> De Novo devices are classified as either Class I or Class II, with “Special Controls” — specific requirements the FDA defines for the new category. For a wound-applied antimicrobial device, Class II with Special Controls is the likely classification. The risk profile is moderate. We’re applying energy to the skin surface, not to internal organs. The device is already cleared for contact with wound beds.</p><p><strong>What the FDA Needs to See:</strong> A De Novo submission of this kind would likely require a package that includes: preclinical safety data, including porcine work to define the therapeutic window, expanded in vitro data across multiple resistant organisms, and human clinical data, potentially from an IDE Early Feasibility Study, showing bacterial-load reduction with acceptable safety.</p><p>If the FDA were to grant a De Novo classification, it would also define Special Controls for the new category. Those could include performance testing, labeling limitations, manufacturing requirements, and possibly post-market expectations.</p><p><strong>Why De Novo and Not PMA?</strong> There is a more stringent pathway called PMA, or Premarket Approval, generally associated with Class III devices that present higher risk.</p><p>PMA usually requires more extensive evidence, larger clinical programs, higher cost, and longer timelines.</p><p>For a wound-surface device using externally applied ultrasound, there is a plausible argument that the risk profile may not rise to the level typically associated with Class III.</p><p>But there is no guarantee.</p><p>The FDA could still decide that a bactericidal claim introduces enough risk, novelty, or public-health significance to justify a PMA-level evidentiary standard.</p><p>That is exactly why early interaction with the agency would be so important.</p><p><strong>An Important Nuance</strong></p><p>Recent FDA scrutiny of antimicrobial wound dressings has often focused on chemical agents that may contribute to resistance or create other safety concerns.</p><p>A device based on a physical mechanism may present a different regulatory logic.</p><p>That distinction could matter.</p><p>But it should not be overstated. The FDA could still conclude that the claim itself, rather than the mechanism alone, warrants a more demanding pathway.</p><p><strong>Breakthrough Device Designation: Sprinting Instead of Walking</strong></p><figure><img alt="AI-generated comparison infographic, “Breakthrough Device Designation: Sprinting Instead of Walking.” Left: Traditional Pathway, hourglass icon, “Submit-and-wait model with longer review cycles.” Right: Breakthrough Device Designation, four features (priority review and shorter timelines, interactive FDA dialogue, early alignment on data and endpoints, manufacturing flexibility). Bottom note: expedited interaction does not lower evidentiary standards." src="https://cdn-images-1.medium.com/max/1024/1*cIKeiaOIdotBY3d17Phftg.jpeg" /><figcaption>Faster, not easier</figcaption></figure><p>The FDA offers an expedited pathway called Breakthrough Device Designation (BDD) for devices that provide more effective treatment for life-threatening or irreversibly debilitating conditions.</p><p>AMR aligns closely with this criterion.</p><p>The WHO has classified it as a critical global threat. Antibiotic-resistant infections are, by definition, life-threatening. They kill because existing treatments fail.</p><p>Breakthrough Device Designation provides several concrete advantages:</p><p><strong>Priority review.</strong> The FDA commits to shorter review timelines and more responsive communication.</p><p><strong>Interactive dialogue.</strong> Instead of the traditional submit-and-wait model, BDD enables ongoing communication between the device developer and the FDA review team during the review process. Disagreements can be resolved in real time rather than through cycles of rejection and resubmission.</p><p><strong>Manufacturing flexibility.</strong> The FDA can agree to expedited manufacturing review, which is important for scaling production.</p><p><strong>Data-driven collaboration.</strong> Perhaps most importantly, BDD allows the applicant to work with the FDA to agree on exactly what data will be sufficient for clearance <em>before</em> the full submission. This dramatically reduces the risk of undertaking clinical programs only to discover the FDA wanted different endpoints.</p><p>Filing for BDD would ideally occur early in the development program, before the clinical trial begins. The goal would be to obtain early regulatory feedback and alignment on the development plan before substantial time and research resources are invested.</p><p><strong>The Pre-Submission Meeting:</strong></p><p>One of the most important and underappreciated tools in device regulation is the Q-Submission, often called a Pre-Sub.</p><p>This is a formal process for obtaining FDA feedback on a proposed development plan before a major submission is filed.</p><p>In a Pre-Sub, the sponsor can present a proposed study design, endpoints, statistical framework, and regulatory rationale. The FDA then provides written feedback.</p><p>That feedback is not binding, the agency can change its position later.</p><p>But in practice, early written feedback can significantly reduce the chance of building a development program around the wrong assumptions.</p><p>For a program like this, a Pre-Sub would likely occur after meaningful preclinical work is available, but before the IDE is filed.</p><p>The conversation would likely focus on three key questions.</p><p>First: Does the FDA agree that De Novo is the appropriate pathway, or does it believe a PMA framework is required?</p><p>This is critical. If the answer is PMA, the entire development plan changes.</p><p>Second: Would the FDA consider log10 bacterial-load reduction an acceptable primary endpoint?</p><p>There may be other acceptable measures, but this is the kind of issue that should be clarified before the study begins, not after.</p><p>Third: What Special Controls might the FDA expect for a new device category of this kind?</p><p>Understanding that early would allow engineering, quality, and regulatory planning to be built around those expectations from the beginning.</p><p><strong>Phase 2: The Long Game — Internal Organs</strong></p><figure><img alt="AI-generated infographic, “Phase 2: The Long Game — Internal Organs.” Anatomical body illustration in background. Two phases compared. Phase 1, Chronic Wounds: surface application to skin, least risky clinical context, logical first indication. Phase 2, Internal Organs &amp; Compartments: bloodstream, lungs, urinary tract, central nervous system; significantly higher safety considerations; likely PMA regulatory pathway. Note: expansion only if wound data demonstrates robust safety and efficacy." src="https://cdn-images-1.medium.com/max/1024/1*G95Rgkce3UVZaHbvFSdG7w.jpeg" /><figcaption>A thought experiment in two phases</figcaption></figure><p>Everything described so far applies to wounds, meaning external application to the skin surface.</p><p>That is the logical starting point, but the broader strategic vision could extend further.</p><p>If the mechanism were shown to work safely and effectively in wounds, the scientific rationale might eventually support exploration in other body systems: blood, lungs, urinary tract, or even central nervous system infections.</p><p>That does not mean those paths would be easy, far from it.</p><p>Applying ultrasound to internal organs or internal compartments raises a very different safety profile.</p><p>Cavitation in blood vessels, lung tissue, or the brain is not the same proposition as cavitation on a wound bed.</p><p>Those indications would likely require a much more demanding regulatory pathway, potentially including PMA, larger trials, more extensive preclinical testing, and a much longer timeline.</p><p>But strategically, the wound indication could serve as the proof of concept that makes any later internal-organ work scientifically and regulatorily credible.</p><p>You would not begin with intravascular ultrasound for sepsis.</p><p>You would begin with the least risky, most accessible clinical context in which the mechanism can be tested.</p><p>First chronic wounds.</p><p>Then, only if warranted by the data, broader indications.</p><p><strong>What This Could Mean for the Broader AMR Effort</strong></p><p>I want to step back for a moment and place this in context.</p><p>One reason the antimicrobial resistance crisis has become so difficult is that our innovation frameworks were largely built around the pharmaceutical model: long timelines, massive costs, one molecule, one indication.</p><p>That model has delivered extraordinary medicines.</p><p>But it is also part of the reason the antibiotic pipeline has become so fragile. The economics are difficult when resistance can erode the commercial life of a product.</p><p>A device-based approach may involve a different development logic.</p><p>That does not automatically mean it would be faster. It might be, or it might not, depending on classification, data requirements, FDA feedback, and study design.</p><p>That distinction matters, it is important not to overstate speed or simplicity.</p><p>But if there is a credible regulatory path that could validate a non-chemical solution in a clinically meaningful timeframe.</p><p>That possibility deserves serious attention from policymakers, clinicians, regulators, and anyone concerned about the trajectory of antimicrobial resistance.</p><p>This is not an argument against drug development, the world needs both.</p><p>But in a crisis that is accelerating, and already costing lives, alternative paths deserve to be evaluated with urgency and discipline.</p><h3>Part 26: What Could Go Wrong, An Honest Assessment</h3><figure><img alt="AI-generated section header titled “What Could Go Wrong: An Honest Assessment.” Central image of a balance scale. The left pan holds a glowing red-orange element representing the technology; the right pan holds a glowing question mark representing unknowns. Flanking the scale are charts labeled “35 kHz” on the left and regulatory icons labeled “Regulation” on the right. Dark blue background." src="https://cdn-images-1.medium.com/max/1024/1*T7OrztlEwz0k9pGrqhZr2w.jpeg" /><figcaption>Weighing the unknowns</figcaption></figure><p>Over the last several articles, I have laid out the science, the hypothetical path to clinical validation, and a possible regulatory strategy.</p><p>If you have read the full series, you may reasonably be asking where the biggest uncertainties and failure points lie.</p><p>Those are fair and necessary questions.</p><p>Let me tell you where this could fail, because the best way to judge optimism is by how honestly it confronts the downside.</p><figure><img alt="AI-generated infographic split into two panels. Left: ultrasound device applied to a petri dish in a clean laboratory, labeled “In Vitro — Controlled Lab,” with magnified inset showing a single bacterium. Right: same device applied to a cross-section of a human wound containing visible bacteria, with a large overlay question mark, labeled “In Vivo — Real Human Wound.” Title at top: “Risk 1: The Science Doesn’t Translate." src="https://cdn-images-1.medium.com/max/580/1*LVukOzaakTu9uvmMvrzc9g.jpeg" /><figcaption>The translation problem</figcaption></figure><p>Risk 1: The Science Doesn’t Translate</p><p>The Mayo Clinic observed antimicrobial effects in electron microscopy. Winston-Salem demonstrated bacterial kill in a petri dish. Both studies are credible and published.</p><p>But in vitro is not in vivo, and a petri dish is not a wound bed.</p><p>The human body is a complex and dynamic environment.</p><p>A living wound involves tissue composition, blood flow, immune response, polymicrobial infection, wound depth and geometry. These variables do not exist in the same way in a controlled laboratory setting.</p><p>It is entirely possible that the bacterial-kill rates observed in vitro would diminish significantly when the technology is applied to real wounds in real patients.</p><p>This is the core scientific risk.</p><p>It is also the reason the Early Feasibility Study is the most important milestone in the entire development path.</p><p>Until human data shows a meaningful reduction in bacterial load, everything else remains conditional.</p><p>Mitigating factor: the device already functions mechanically in human wounds. It is FDA-cleared for debridement and cleansing. In other words, the question is not whether the hardware can operate in the wound bed. It can.</p><p>The question is whether the bactericidal effect observed in the laboratory persists in a real clinical environment.</p><p>That is still a meaningful leap, but it is a shorter leap than moving from pure theory to first laboratory evidence.</p><p>Risk 2: Collateral Tissue Damage</p><p>The therapeutic window, the energy range that may kill bacteria while sparing human cells, is central to the entire strategy.</p><p>If that window is narrower than the early evidence suggests, or if it varies significantly across patients, wound types, tissue conditions, or anatomies, the clinical application becomes far more difficult.</p><p>In a moderate downside scenario, the technology might work only in a subset of patients or wound types, limiting its practical value.</p><p>In a more serious downside scenario, the window could prove so narrow that achieving a reliable bactericidal effect without unacceptable tissue damage is not clinically practical.</p><p>Mitigating factor: the selectivity hypothesis is grounded in a real structural distinction between prokaryotic and eukaryotic cells.</p><p>This is not merely a subtle pharmacological difference; it is a basic architectural difference.</p><p>Prokaryotic cells lack many of the structural features that make eukaryotic human cells more mechanically resilient.</p><p>That does not prove the therapeutic window will be broad enough, but it does provide a rational basis for testing the idea.</p><p>And that is precisely what the preclinical porcine work would be designed to assess. If the window is too narrow, the data should reveal that before broad clinical use is ever contemplated.</p><p>Risk 3: The FDA Says No, or Says “More”</p><p>Regulatory risk is inherent in any novel medical device program.</p><p>The two obvious risks here are these:</p><p>Classification creep. The strategy assumes a possible De Novo Class II path for the wound indication. The FDA could instead conclude that a bactericidal claim warrants Class III treatment and a PMA pathway, significantly increasing the evidence burden, cost, and timeline.</p><p>Endpoint disagreement. The FDA could determine that log10 bacterial-load reduction is not enough on its own and require different or additional endpoints, such as clinical cure, infection resolution, or wound closure. That would affect study design, sample size, cost, and timing.</p><p>Mitigating factor: this is exactly why the strategy emphasizes early FDA engagement through a Q-Submission, and potentially Breakthrough Device Designation if appropriate.</p><p>Those mechanisms do not eliminate regulatory risk, but they do reduce the likelihood of spending years and significant resources on a study design the agency ultimately rejects.</p><figure><img alt="AI-generated image titled “Risk 4: It Works, But Nobody Adopts It.” A blurred background shows a group of clinicians in white coats meeting around a conference table in a hospital setting. In the foreground sits a medical device with its inner electronics visible. Other clinical devices appear on shelves. Three labeled icons across the top: Reimbursement Issues (person icon), Capital Approval Delays (rising chart icon), Conservative Adoption Culture (clinician with screen icon)." src="https://cdn-images-1.medium.com/max/832/1*hTGkr-6m5wihHZpq9Cvabw.jpeg" /><figcaption>The adoption problem</figcaption></figure><p>Risk 4: It Works, But Nobody Adopts It</p><p>Clinical adoption risk is real, and technologists often underestimate it.</p><p>Hospitals are conservative institutions. Capital purchases require committee approval, budget allocation, clinical champions, workflow integration, and a believable reimbursement story.</p><p>Even compelling data does not guarantee rapid uptake.</p><p>There is also a more specific challenge here: wound care is not usually treated as a prestige specialty. It does not command the same institutional attention as cardiology, oncology, or orthopedics.</p><p>That means the adoption case may need to be made not through excitement, but through economics and outcomes.</p><p>Mitigating factor: if reimbursement, utilization, and outcomes data ultimately support it, the economic case could be significant.</p><p>The basic premise is straightforward. Resistant infections are expensive. They drive longer hospital stays, readmissions, surgical interventions, and downstream complications.</p><p>A technology that reduces those burdens, while supporting reimbursable clinical use, could become economically attractive to hospitals and health systems.</p><p>But that argument would need to be demonstrated with real-world evidence, not assumed in advance.</p><p>Risk 5: It Works, But it’s Not Accessible</p><p>Even a clinically effective solution would face barriers to broad deployment, especially in low- and middle-income countries where the burden of antimicrobial resistance is often highest.</p><p>Device-based therapies require manufacturing, distribution, training, maintenance, and clinical infrastructure.</p><p>A solution that works in principle but cannot scale in practice may still leave much of the world behind.</p><p>Equitable access is critical.</p><p>The Risk I Think About Most</p><figure><img alt="AI-generated image depicting the urgency of time in addressing antimicrobial resistance. A hospital intensive care room is split visually down the middle. On the left, a nurse in scrubs attends to a patient in bed under cool blue lighting, with medical monitoring equipment in the background. On the right, the same scene shifts to a warm orange glow with stylized microbial particles floating in the air. A large translucent clock face dominates the center of the image." src="https://cdn-images-1.medium.com/max/784/1*IqFG_z1stbysjDzXQC04JQ.jpeg" /><figcaption>The temporal risk</figcaption></figure><p>I will be honest, the risk I think about most is not scientific or regulatory, it is temporal.</p><p>This technology exists today. Much of the foundational work was done years ago.</p><p>Meanwhile, the AMR crisis is not standing still. It is accelerating.</p><p>Every year without validated non-pharmaceutical solutions, resistance evolves, clinical options narrow, and the consequences compound.</p><h3>PART 27: Final Thoughts</h3><figure><img alt="A close-up indoor family photograph. At left, Senator Rudy Boschwitz, an elderly man with glasses and silver hair in a plaid shirt, smiles broadly. Center, Blake Vanderhyde’s wife Jessica, with blonde hair, smiles while holding the family rescue dog Chewy, a small fluffy cream-colored terrier. At right, Blake’s young son Hudson, his face deliberately obscured for privacy. Warm indoor lighting." src="https://cdn-images-1.medium.com/max/1024/1*ZDBkDfLLNGiuf2aeLNSBWA.jpeg" /><figcaption>The Senator with my wife Jessica, Hudson, and Chewy</figcaption></figure><p>I began this series with a simple question: What if there is a better way to fight antibiotic-resistant bacteria?</p><p>I have done my best to explore that question honestly, carefully, and with practical focus.</p><p>This series was never about me. It was never about any single inventor, device, or company. It is about a crisis already claiming millions of lives while much of the world continues to look away.</p><p>I am not a scientist, engineer, or regulator. Over the past five years, I have studied this problem, tried to understand the science, and sought out expert perspectives, not because I claim to have the answers, but because the question itself seemed too important to ignore.</p><p>The people who will ultimately determine whether any of these ideas have merit are the people doing the real work: researchers, clinicians, engineers, and regulators. They are the ones who test ideas, set standards, and determine what is real. Nothing I have written here replaces that.</p><p>What I have tried to do is place a serious and specific question into the public record, one concrete enough to test and honest enough to challenge.</p><p>If any part of this has been useful, then take it and do something better with it.</p><p>Use it. Challenge it. Improve it. Replace it.</p><p>The goal was never to produce a final answer. It was to help start a conversation that might lead to one.</p><p>I hope that moment is now.</p><p>This journey began, in part, with my son, Hudson, and the fear I felt as a new father during the COVID years. Like many people, I experienced loss during that time, and some losses do not fully leave you.</p><p>In those days, I read Hudson bedtime stories. I still do.</p><p>One of his favorites was <em>The Lorax</em>, and one line has stayed with me through all of this:</p><p>“Unless someone like you cares a whole awful lot, nothing is going to get better. It’s not.”</p><p>Thank you for reading.</p><figure><img alt="A laughing toddler in a blue and white seersucker outfit with a monogrammed “H” is held aloft by his father, visible from the chest down in a light blue shirt with arms raised. The setting is outside a brick building, at the park where Blake Vanderhyde and his son Hudson first met Senator Rudy Boschwitz." src="https://cdn-images-1.medium.com/max/1024/1*vVwPAaUwAjH_wdOhx8MLlw.jpeg" /><figcaption>For Hudson</figcaption></figure><p>IMPORTANT CLINICAL NOTE: The preceding discussion regarding bactericidal resonance is purely a physics-based hypothesis for research purposes. Clinicians and facilities must not utilize the Qoustic System, or any similar low-frequency ultrasound device, for the purpose of destroying bacteria or treating infections. Such use is not FDA-cleared, and its safety for these specific indications has not been established in human clinical trials.</p><p>FULL DISCLOSURE: This series is provided for informational and educational purposes only. It does not constitute an offer to sell securities, a solicitation of interest in any securities offering, medical advice, or investment advice. The author is not a medical doctor or a registered investment advisor. Certain statements in this series may be forward-looking statements within the meaning of the Private Securities Litigation Reform Act of 1995. Forward-looking statements involve known and unknown risks and uncertainties that could cause actual results to differ materially, including but not limited to: failure of in vivo scientific validation, adverse FDA classification or pathway decisions, inability to secure development financing, unforeseen competitive developments, regulatory changes, and loss of key inventor knowledge. The technology discussed is currently FDA-cleared only for wound cleansing and debridement; the antimicrobial applications discussed here are purely investigational and are not FDA-cleared for antimicrobial indications. Readers are cautioned not to use any medical device for indications not cleared by the FDA. As of the date of publication, no legal entity has been formed or identified for purposes of any securities offering related to this technology. The author is not currently seeking, and will not accept, any inquiries regarding investment or financial participation. Any future financing, if any, would be conducted only through formal offering materials and in compliance with applicable federal and state securities laws. Do not rely on this content for investment, legal, or medical decisions.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=efc558e21754" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>