<?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 Lou Ritter on Medium]]></title>
        <description><![CDATA[Stories by Lou Ritter on Medium]]></description>
        <link>https://medium.com/@louritter1?source=rss-ca535acb0e04------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*kUeXZdC6mO-C0JjsRn8_7A.png</url>
            <title>Stories by Lou Ritter on Medium</title>
            <link>https://medium.com/@louritter1?source=rss-ca535acb0e04------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Wed, 18 Nov 2026 18:50:07 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@louritter1/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[Building with the Wizard: Bringing Ideas to Life with AI (Week 3 at Overclock)]]></title>
            <link>https://medium.com/@louritter1/building-with-the-wizard-bringing-ideas-to-life-with-ai-week-3-at-overclock-d9edffa42678?source=rss-ca535acb0e04------2</link>
            <guid isPermaLink="false">https://medium.com/p/d9edffa42678</guid>
            <category><![CDATA[llm]]></category>
            <category><![CDATA[python]]></category>
            <category><![CDATA[learning]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[product-owner]]></category>
            <dc:creator><![CDATA[Lou Ritter]]></dc:creator>
            <pubDate>Thu, 23 Oct 2025 05:29:09 GMT</pubDate>
            <atom:updated>2025-10-23T05:29:09.938Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*gqBy6uJFZkGQzPkBsP6aGg.png" /></figure><p>I’ll be honest, I’m a little behind on this series. Week 3 at <strong>Overclock Accelerator</strong> came and went faster than my local build could deploy, and between work, experiments, and late-night debugging, the writing part slipped through the cracks.</p><p>But this week’s theme <strong>“Build with the Wizard”</strong> deserves a proper recap.</p><h3>Leveling Up: From AI-Enhanced to AI-Powered</h3><p>This week we went a step higher in the AI engineering stack. We weren’t just using AI to <em>help</em> us build anymore. We were learning how to actually <em>build with it inside</em> our applications.</p><p>The focus was on wiring large language models (LLMs) directly into our codebases, creating products that think, reason, and respond dynamically. This is something I have been doing for a while but having the team at Overclock to support me brought this skill to new levels.</p><p>We learned how to:</p><ul><li>Select the right LLM for different needs (speed, cost, accuracy)</li><li>Pre-configure <strong>system prompts</strong> to behave like app-specific brains</li><li>Structure multiple AI calls inside one cohesive workflow</li><li>Keep context tight and outputs predictable, utilizing sandboxes and variables.</li></ul><p>This is where things started feeling less like “prompting” and more like <strong>software architecture for intelligence.</strong></p><h3>The Magic Lens</h3><p>For my sprint project, I built something I’ve been thinking about for a while:<br> <strong>Kid Vision, </strong>a friendly AI storytelling app for children.</p><p>Here’s the idea: a child takes a picture or short video and “hands it” to a fuzzy on-screen monster. The monster then transforms it into a personalized, educational story, narrated in its own monster voice.</p><p>It’s imaginative, slightly chaotic, and powered entirely by AI. Exactly the kind of project that makes this accelerator so fun.</p><p><a href="https://www.linkedin.com/posts/activity-7384589510111621121-wYpD?utm_source=social_share_send&amp;utm_medium=member_desktop_web&amp;rcm=ACoAABlRf-MBt3E0DDxeofzeIWo8lnsQuWbUHgE">Link to my LinkedIn post showing a demo</a></p><h3>How It Works</h3><p>The app is split into two parts:</p><p><strong>Frontend:</strong></p><ul><li>Built with <strong>Next.js </strong>, <strong>React</strong>, and <strong>TypeScript</strong></li><li>Styled with <strong>Tailwind </strong>and animated using <strong>Framer Motion</strong></li><li>Includes <strong>Supabase authentication</strong> for parents and a playful child-facing UI</li></ul><p><strong>Backend:</strong></p><ul><li>A <strong>Python API</strong> handling all LLM and image generation calls</li><li>Story generation via a pre-configured system prompt</li><li>Text to speech to create the narrator’s voice</li><li>Age-based personalization to tailor tone and vocabulary</li></ul><p>I wanted it to feel alive, not just an app, but a digital companion that could learn, react, and create alongside a child. It’s amazing how AI tools make such incredible features come to life. It allows me to think like a product owner while building, keeping the users in mind.</p><h3>The Big Unlock: Interactive AI Storytelling</h3><p>Most kids’ apps rely on static responses or predictable scripts.<br>I wanted this to feel exploratory, something that adapts with each tap.</p><p>So I added an <strong>interactive grid overlay</strong>. Kids can tap on different parts of their image, and the monster tells them what’s happening in that section, maybe the bird is learning to fly, or the tree is a portal to a secret garden.</p><p>Each tap triggers a new AI call, cropping that region and generating a localized story moment.<br>It’s playful, dynamic, and powered by curiosity.</p><h3>Designing for Kids (and Parents)</h3><p>Building for children adds another layer of complexity. Everything has to be safe, simple, and engaging.</p><p>The monster’s voice is friendly, the buttons are oversized and bouncy, and the colors lean warm, oranges and blues to feel comforting but magical.</p><p>Meanwhile, the <strong>Parent Dashboard</strong> (backed by Supabase) lets adults manage child profiles, track generated stories, and toggle between anonymous and personalized modes.</p><p>It’s one of those small UX details that makes a big difference when real families use it.</p><h3>What This Week Taught Me</h3><p>This sprint reinforced something I’ve been realizing since Week 1:<br> AI isn’t just a tool, it’s a collaborator.</p><p>When you design prompts and system logic the right way, you’re not telling the AI what to do, you’re <em>teaching it how to think for your product</em>.</p><p>It’s a balancing act between creativity, engineering, and trust.</p><p>Also, token budgets are real. Turns out, even imaginary monsters can run up real bills.</p><h3>Looking Ahead</h3><p>Next week we’re diving into <strong>Prompt and Context Engineering</strong> — basically the art of turning messy human thought into clean, reusable logic.</p><p>If Week 3 was about building with magic, Week 4 looks like learning how to <em>control</em> it.</p><p>And honestly? I can’t wait.</p><h3>Stay Connected</h3><p>If you’ve been following this series , thank you.<br> There’s much more to come as we keep climbing the AI engineering stack.</p><ul><li>Follow me on <a href="https://medium.com/@louritter1"><strong>Medium</strong></a> for future Overclock posts and other experiments.</li><li>Connect on <a href="https://www.linkedin.com/in/louritter"><strong>LinkedI</strong>n</a> for behind-the-scenes updates and product insights.</li><li>Check out <a href="https://overclockaccelerator.com/"><strong>Overclock Accelerator</strong></a> to learn more about the program shaping the next wave of AI engineers.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d9edffa42678" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Beyond Vibe Coding: When the Fun Stops and the Real Work Begins (Week 2 at Overclock)]]></title>
            <link>https://medium.com/@louritter1/beyond-vibe-coding-when-the-fun-stops-and-the-real-work-begins-week-2-at-overclock-e060ad8eadfc?source=rss-ca535acb0e04------2</link>
            <guid isPermaLink="false">https://medium.com/p/e060ad8eadfc</guid>
            <category><![CDATA[software-architecture]]></category>
            <category><![CDATA[aritificial-intelligence]]></category>
            <category><![CDATA[vibe-coding]]></category>
            <category><![CDATA[spec-driven-development]]></category>
            <category><![CDATA[tech]]></category>
            <dc:creator><![CDATA[Lou Ritter]]></dc:creator>
            <pubDate>Wed, 08 Oct 2025 11:47:54 GMT</pubDate>
            <atom:updated>2025-10-08T11:47:54.124Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*3mk6UGzgg-7sFdDFiWJCdw.png" /></figure><p>This is week 2 of a weekly series documenting my time at a top AI Accelerator. If you haven’t checked out week one, give it a read <a href="https://medium.com/@louritter1/what-i-built-in-7-days-with-ai-at-overclock-accelerator-and-what-it-taught-me-16df60f59770"><strong>here</strong></a></p><p>Week 2 at <strong>Overclock Accelerator</strong> was a shift in tone.<br> Last week was about creativity and speed. This week, things got more technical. We moved from experimenting with AI to learning how to work with it inside existing codebases.</p><p>The session was called “Beyond Vibe Coding,” and it lived up to the name. It was more technical and more structured, for good reason. Handling real-world code is a different skill than spinning up a new prototype for fun. Anybody contributing to a large production codebase can probably already tell you that.</p><h3>Moving from Building to Maintaining</h3><p>The focus this week was understanding how to apply AI tools to code that already exists. Not greenfield projects, but messy, living repositories with very real, very human, history.</p><p>We were introduced to new tools that made working with production codebases as painless as possible.</p><ul><li><strong>Repomix </strong>compresses repositories into a format AI models can understand. Perfect for asking LLMs questions about code spanning numerous files and classes.</li><li><strong>Context7 </strong>feeds documentation and libraries so the model has accurate context.</li><li><strong>Spec Kit</strong> structures the process, keeping tasks aligned with real engineering principles.</li></ul><p>These tools aren’t about writing faster code. They’re about writing smarter code. This has been one of the loudest concerns surrounding AI-assisted engineering, and these tools rise up to the challenge.</p><p>Ahmed, the instructor, showed us an incredible and sleek looking zoom-like application, Overcast, built using these very tools. It was clean, structured, and performant. It really showed the capability of these tools to create something that looked the part and was well built, geared towards production.</p><h3>The Power of Spec Kit</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ersH4laWp4cpdLpUIHh_OQ.png" /></figure><p>Spec Kit was the biggest unlock for me. I had been looking for something like it. My biggest question with AI development has always been: how do you get an AI to respect standards, maintain context, and follow engineering best practices and not just forget one or two prompts later?</p><p>Spec Kit answered that by introducing structure. It starts with a <strong>constitution</strong>, which defines principles and constraints. Then it generates a <strong>specification</strong>, goes through <strong>clarification</strong> to resolve ambiguities, and moves through <strong>planning</strong>, <strong>task setting</strong>, and finally, <strong>implementation</strong>. All perfectly baked in to most popular agentic coding tools, such as cursor.</p><p>Watching it run through a task list was both exciting and slightly stressful. It was fast, confident, and very capable. It made me wonder what limits are actually left.</p><h3>The Human in the Loop</h3><p>Then I made the classic mistake: I gave Spec Kit too long of a leash.</p><p>It started making assumptions and introducing bugs. I got lazy with my specifications and code reviewing. That moment drove home one of the biggest lessons so far, AI doesn’t remove the need for human attention. It multiplies it.</p><p>You still need to stay present, understand what’s happening, and guide the process. These tools are powerful, but they aren’t self-correcting. The best work happens when the human is still the one steering.</p><p>This week reinforced that the goal isn’t full automation. It’s collaboration, finding the balance between trust and control. want to mow down a list of five minor updates and check the results after? Spec Kit will certainly help you with that. For bigger features, I found more success chunking the task by the sections Spec Kit so kindly creates for you and monitoring as you go.</p><h3>Our FeatherWiki Project</h3><p>Time for breakout groups for the week. My group chose to work on FeatherWiki, a lightweight open-source alternative to Wikipedia. Its simplicity was part of the challenge. We wanted to preserve its lightweight design while still introducing something new.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wSaKAJCWQIrg3NmZytSnSg.png" /></figure><p>We set up a constitution inside Spec Kit that emphasized minimalism and clarity. Every feature we built had to follow that rule. The cool part was how well it worked. It avoided dependencies and unnecessary bloat with ease. Each of us built different features, but everything felt like it came from one mind. My feature was a theming engine, after the first specification, it was changing themes. After the second, more UI-friendly. I can keep iterating using the same steps to enhance features endlessly.</p><p>It showed how these tools can help a team of humans and AI operate under a shared vision. Like a unified development style guide that enforces itself.</p><h3>Takeaways</h3><p>This week wasn’t about hype. It was about discipline.</p><p>Working with existing code reminded me that the magic of AI only works when it’s anchored in real engineering structure. Tools like Repomix, Context7, and Spec<strong> </strong>Kit don’t replace us. They extend what we can do.</p><p>They make it possible to move faster, but also to stay intentional. The trick is to keep the human element strong, to know when to guide, when to let go, and when to pull the reins back.</p><p>The fun certainly hasn’t stopped. It’s just gotten more real.</p><p><em>Next up: we start building our own AI-driven applications from the ground up. The specs are ready. Time to put them to work. Stay Tuned.</em></p><h3>Stay Connected</h3><p>If you enjoyed this piece, you can:<br><strong>Follow me on </strong><a href="https://medium.com/@louritter1"><strong>Medium</strong></a> for more posts on AI, engineering, and product thinking.<br><strong>Connect with me on </strong><a href="https://www.linkedin.com/in/louritter"><strong>LinkedIn</strong></a><strong> </strong>to keep up with my projects and insights from the Overclock Accelerator journey.<br><strong>Learn more about </strong><a href="https://overclockaccelerator.com/"><strong>Overclock Accelerator</strong></a>, a program redefining how we build with AI.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e060ad8eadfc" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The AI Bubble Boy: Why I’m Building While the Skeptics Get Louder]]></title>
            <link>https://ai.plainenglish.io/the-ai-bubble-boy-why-im-building-while-the-skeptics-get-louder-4c866a7ca5ce?source=rss-ca535acb0e04------2</link>
            <guid isPermaLink="false">https://medium.com/p/4c866a7ca5ce</guid>
            <category><![CDATA[criticism]]></category>
            <category><![CDATA[opinion]]></category>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[software-engineering]]></category>
            <dc:creator><![CDATA[Lou Ritter]]></dc:creator>
            <pubDate>Mon, 06 Oct 2025 04:43:11 GMT</pubDate>
            <atom:updated>2025-10-06T20:27:50.386Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*4rh6NEbiR1d_24fpnDIS7w.png" /></figure><h3>Hint: It’s not about hype. It’s about leverage.</h3><p>Every few years, a new technology comes along that divides the room. Some call it revolutionary, others call it overhyped, and somewhere in between, the builders just keep building.</p><p>AI is that technology right now. Depending on who you ask, it’s either going to automate humanity into a utopia or implode like a bad startup pitch. Scroll through X for five minutes and you’ll find people posting screenshots of bad code as if it’s proof of the apocalypse.</p><p>“See?” they say. “It can’t even write a for-loop properly.”</p><h3>Nate Berkopec on X (formerly Twitter): &quot;Some of the code these AI models put out is truly an abomination pic.twitter.com/gFgXPoVGKV / X&quot;</h3><p>Some of the code these AI models put out is truly an abomination pic.twitter.com/gFgXPoVGKV</p><p>Meanwhile, their coworker Steve has been writing 400-line switch statements inside nested if-blocks for years, and somehow that’s fine. We’ve all inherited code from a dozen Steves. We’ve all <em>been</em> Steve on a deadline. The hypocrisy is stunning. We hold AI to a standard of perfection we’ve never once demanded from ourselves.</p><h3>The Fear of Change Is Human</h3><p>It’s easy to dunk on AI skeptics, but the truth is that fear of change is just human nature. Every major shift in technology has looked terrifying at first.</p><p>The internet was a “fad.” Smartphones were “a toy.” Cloud computing was “too risky.”</p><p>Go back further. When high-level compilers were invented, plenty of assembly programmers scoffed. They saw it as a crutch, a layer of abstraction that would make developers lazy and disconnected from the machine. They were on the wrong side of history. The leverage a compiler gave a single developer was worth more than a thousand lines of perfectly optimized assembly.</p><p>People aren’t afraid of the tech itself. They’re afraid of what it means for them. If you’ve built your identity around how things currently work, watching the rules change overnight can feel like an existential gut punch.</p><h3>Déjà Vu All Over Again</h3><h3>AgresTrades 🪙 on X (formerly Twitter): &quot;Whenever someone tells you #AI is taking over the world just link them this article from MIT.Every single technological advancement has gone through the same boom to bust cycle. We are near the peak of hysteria and greed.What comes next is the inevitable bubble burst.enjoy pic.twitter.com/3N1Qy3w8HR / X&quot;</h3><p>Whenever someone tells you #AI is taking over the world just link them this article from MIT.Every single technological advancement has gone through the same boom to bust cycle. We are near the peak of hysteria and greed.What comes next is the inevitable bubble burst.enjoy pic.twitter.com/3N1Qy3w8HR</p><p>This pattern of hype, crash, and quiet revolution isn’t new. We’ve seen this movie before, and the critics get the ending wrong every time by confusing the speculative froth with the foundational technology. Even when the bubble “bursts”, technologic advancement is a byproduct.</p><ul><li><strong>The Dot-Com Bubble (~1999):</strong> This was the big one. The bubble was Pets.com, Webvan, and a thousand other companies with nonsensical business models and billion-dollar valuations based on “eyeballs.” When the NASDAQ imploded in 2000, the skeptics took a victory lap, declaring the internet a failure. But the crash didn’t erase the technology. The breakthrough was the fiber optic cable, the server farms, and the protocols that were built during the frenzy. From those ashes, survivors like Amazon and new giants like Google built the digital foundation of modern life. Cloud infrastructure was born. The internet wasn’t the bubble; the irrational stock prices were.</li><li><strong>The Mobile App Gold Rush (~2010):</strong> Remember when every local pizza place had a clunky, useless app? That was the bubble. It was the belief that every business needed a standalone app, leading to a gold rush of low-quality software and get rich quick schemes. Many of those apps died, and the app stores became a brutal battlefield. But the breakthrough wasn’t the individual apps, it was the platform. The creation of a supercomputer in every pocket, always connected, unlocked industries that didn’t exist before. Uber, Instagram, DoorDash, and mobile banking are the pivotal results, built on a platform that the initial “bubble” helped popularize.</li><li><strong>The Crypto/Blockchain Frenzy (~2021):</strong> Laser eyes on Twitter. JPEGs of apes selling for more than a house. This bubble was particularly glaring. The promise that a decentralized ledger would fix everything from voting to your morning coffee order. The bubble was the absurd belief that <em>everything</em> needed to be on a blockchain immediately, fueled by meme coins and pure speculation. Then the crash came, and the world collectively rolled its eyes, dismissing the entire space as a digital casino. But look under the hood. The immense capital and talent that flooded the space resulted in a massive acceleration of breakthroughs in next-generation cryptography like Zero-Knowledge Proofs, a revolutionary way to prove something is true without revealing the underlying data. The hype may not have been warranted, but the significant advancements in cryptography and decentralized finance primitives will have an impact for decades.</li></ul><p>And now here we are with AI. The “bubble” is the hype around AGI arriving tomorrow, the cringe-worthy marketing slogans, and the sky-high valuations for thin wrappers around an API. The <strong>breakthrough</strong> is the tangible, daily leverage it’s giving builders <em>right now</em>.</p><h3>The Builders’ Perspective</h3><h3>crocon media on X (formerly Twitter): &quot;AI is not the bubble ... it is the needle. pic.twitter.com/qMonzdJzyD / X&quot;</h3><p>AI is not the bubble ... it is the needle. pic.twitter.com/qMonzdJzyD</p><p>If you actually build software for a living, it’s hard not to see what’s happening right now. Tools like Cursor, Claude Code, and Aider aren’t party tricks. They’re accelerators. They don’t just autocomplete — they help you reason, debug, refactor, and ship faster than ever before.</p><p>How could anyone in engineering bet against that?</p><blockquote><em>AI is the greatest cognitive offload tool ever invented for developers. It handles the boring, so we can focus on the brilliant.</em></blockquote><p>I’ve watched it:</p><ul><li>Redesign a website in under a day.</li><li>Generate the boilerplate for a new REST API endpoint in under 30 seconds.</li><li>Refactor a tangled mess of procedural code into a clean, readable class.</li><li>Explain my own code back to me like a junior dev who actually listens.</li></ul><p>Sure, it gets things wrong sometimes. It hallucinates. It misplaces a bracket. But so does Steve.</p><p>If you judge AI by its worst output, you’re missing the point. The worst code it writes is still fixable. The best code it writes changes what’s possible for small teams and solo builders who used to need a whole department.</p><h3>It’s Not a Miracle Cure</h3><p>Let’s be real, AI isn’t going to fix everything. It won’t make bad engineers good or rewrite your entire product strategy. It won’t magically ship a feature that no one asked for but marketing insists on.</p><p>We should temper expectations. Some of the hype is absurd, some startups will flame out, and some investors will get burned. That’s how every wave of innovation works.</p><p>But between the loudest skeptics and the loudest believers, there’s a middle ground. And that’s where the truth usually lives. The real shift isn’t about replacing engineers; it’s about changing the job description.</p><p>The engineer of tomorrow is less of a line-by-line code monkey and more of an architect, a prompter, a critic, or a visionary. The premium will be on your ability to ask the right questions, validate the output, and steer the system toward an elegant solution. You become the conductor of an AI orchestra, not a lone violinist.</p><p>AI won’t replace software engineering. It’s <em>redefining</em> it. It’s not here to remove humans. It’s here to multiply what we can do.</p><h3>Betting on the Future</h3><p>So yeah, call me a bubble boy. I’ll take it.</p><p>The real risk isn’t that the AI bubble pops. The real risk is standing on the sidelines, waiting for permission to innovate, while the game changes around you. The cost of inaction is far greater than the cost of experimentation. Refuse to use the tools and you will be replaced by some bright eyed junior dev who launched three apps using AI while you were screaming into the ether about the end times.</p><p>Every revolution looks like a bubble before it becomes the new normal. Almost certainly this one will cool off a little, especially in the stock market, but the foundation is too strong to vanish. The smartest engineers I know aren’t running from AI, they’re running toward it.</p><p>And if that’s the bubble, it’s one I’m happy to live in.</p><p>— Bubble Boy</p><p>Follow me <a href="https://medium.com/u/ca535acb0e04">Lou Ritter</a> for more blogs about AI, Technology and much more</p><h3>A message from our Founder</h3><p><strong>Hey, </strong><a href="https://linkedin.com/in/sunilsandhu"><strong>Sunil</strong></a><strong> here.</strong> I wanted to take a moment to thank you for reading until the end and for being a part of this community.</p><p>Did you know that our team run these publications as a volunteer effort to over 3.5m monthly readers? <strong>We don’t receive any funding, we do this to support the community. ❤️</strong></p><p>If you want to show some love, please take a moment to <strong>follow me on </strong><a href="https://linkedin.com/in/sunilsandhu"><strong>LinkedIn</strong></a><strong>, </strong><a href="https://tiktok.com/@messyfounder"><strong>TikTok</strong></a>, <a href="https://instagram.com/sunilsandhu"><strong>Instagram</strong></a>. You can also subscribe to our <a href="https://newsletter.plainenglish.io/"><strong>weekly newsletter</strong></a>.</p><p>And before you go, don’t forget to <strong>clap</strong> and <strong>follow</strong> the writer️!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=4c866a7ca5ce" width="1" height="1" alt=""><hr><p><a href="https://ai.plainenglish.io/the-ai-bubble-boy-why-im-building-while-the-skeptics-get-louder-4c866a7ca5ce">The AI Bubble Boy: Why I’m Building While the Skeptics Get Louder</a> was originally published in <a href="https://ai.plainenglish.io/">Artificial Intelligence in Plain English</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[What I Built in 7 Days with AI at Overclock Accelerator (And What It Taught Me)]]></title>
            <link>https://medium.com/@louritter1/what-i-built-in-7-days-with-ai-at-overclock-accelerator-and-what-it-taught-me-16df60f59770?source=rss-ca535acb0e04------2</link>
            <guid isPermaLink="false">https://medium.com/p/16df60f59770</guid>
            <category><![CDATA[artificial-intelligence]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[education]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[Lou Ritter]]></dc:creator>
            <pubDate>Thu, 02 Oct 2025 13:27:28 GMT</pubDate>
            <atom:updated>2025-10-08T03:41:09.912Z</atom:updated>
            <content:encoded><![CDATA[<h3>What I Built in 7 Days with AI at Overclock Accelerator (And What It Taught Me): Week 1</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*1VYkbarYPNVekTReg945ng.png" /></figure><p>This week, I kicked off something I’ve been looking forward to for a long time: being in the Overclock AI Engineering Accelerator. Within just a few days, it’s already changing how I think about building software.</p><p>Overclock isn’t just another online program. It’s a tightly curated accelerator with top-tier instructors and incredibly sharp builders. As we introduced ourselves, I was thoroughly impressed with the caliber of talent in the call. The energy is high, the expectations are real, and the atmosphere is collaborative and grounded.</p><h3>Old Faces, New Tools</h3><p>On day one, I spotted a few familiar names from my Rutgers Web Development Bootcamp from nearly a decade ago. The most recognizable was Ahmed Haque, who taught that course back then, and is now leading this one. Ahmed helped me craft the skills I needed to break into this industry. He was an excellent instructor back then, and he’s even better now.</p><p>Back in those days, we were stitching together jQuery sliders and praying to the Heroku gods that our apps deployed. Now we’re using LLMs to refactor code, generate components in Cursor, and ship full-stack apps before lunch. The leap in tools feels like moving from a chisel to a laser cutter.</p><h3>Welcome to Software 3.0</h3><p>A big part of week one was getting grounded in the shift to what’s now being called “Software 3.0.” We watched a handful of talks that framed it well, especially Andrej Karpathy’s <em>Software is Changing (Again)</em>.</p><p>The key idea is simple but massive: <strong>the English language is now your programming language.</strong></p><p>Instead of breaking down problems into lines of code from the start, we begin by describing what we want in plain language. With the right techniques and tooling, that’s enough to get real, working software. It’s a new kind of literacy, one we’re learning to use effectively.</p><p>Ahmed hammered this point throughout the week: understanding how these tools <em>think</em> is a core engineering skill now.</p><h3>From Vibe Coding to Real Results</h3><p>We also explored the controversial topic of “vibe coding”, the idea that you can now tell AI what you want in plain English and let it handle implementation. Instead of adjusting layout CSS manually, you can just say:</p><blockquote><em>“Make the sidebar half as wide and change all the buttons to blue.”</em></blockquote><p>It’s not magic. There’s structure behind it. And its easy to end up with a huge plate of spaghetti code. That is why were here. We learned how prompts work, how system instructions shape behavior, different tools used in the current landscape, and how to guide an AI’s attention toward the right outcomes.</p><p>Understanding this changed how I write prompts and how I think about engineering in general.</p><h3>The Homework That Sounded Impossible</h3><p>Our first assignment? Build two working prototypes in one week.</p><p>“Wait… what did he just say?”</p><p>At first, that felt like a huge stretch. But the frameworks, prompting tactics, and mindset we were given made it not just doable, it made it exciting. Everyone in the cohort started shipping things that were creative, beautiful, and high-quality. I got a surge of excitement clicking each link to a project my colleagues built. The range of ideas and execution was wild, and each one impressed me in different ways.</p><p>Here’s what I built:</p><h3>🏓 Blazin’ Paddles</h3><p>A pickleball court reservation app with real-time booking, and account management.<br> 👉 <a href="https://blazin-paddles-sc7a.vercel.app/landing">Try it here</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*q0AZUCvnfw7G_6gvzVwPJQ.png" /></figure><h3>🌙 Alchera: AI Dream Analyzer</h3><p>A dream journal and analysis app inspired by Jungian archetypes. Uses Gemini for interpretation and DALL·E for tarot-style image generation.<br> 👉 <a href="https://jungian-dream-weaver.vercel.app/">Explore it here</a></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*XPlUzb0O_ddDDiGqQAYbLQ.png" /></figure><p>For both apps this week, I really stressed creating clean, eye-catching UI. I personally feel great about how my projects came out. I learned about about the things these tools do great and also why engineering knowledge is still so valuable. There were times my vision was being lost by the IDE. But with some refinement and more explicit direction, it came through. It sometimes worked so well, it surprised me with UI I never even thought of.</p><h3>Using These Skills at Work</h3><p>One of the coolest moments was applying this at my 9–5. I was assigned to reskin a production website, something I originally scoped for two weeks, and I got most of the heavy lifting done in about half of a work week.</p><p>It wasn’t about cutting corners. I looked through the code. It was better and cleaner than before. That’s what happens when you pair experience with the Overclock stack. These tools don’t just save time, they shift how you work entirely.</p><h3>Looking Ahead</h3><p>Next week, we’ll be diving into existing production codebases, working with APIs, and building with testing and coverage in mind. It’s more complex, but I’m excited to take this new workflow into deeper waters.</p><p>After one week, I already feel faster, sharper, and more creative as an engineer. It was the exact spark I needed. It’s like I found Mary Poppins purse, and I’ve only just started exploring what’s inside.</p><p><strong>Want to check out Overclock for yourself?</strong></p><p>It’s rare to find a program that feels this practical and forward-looking at the same time. If you’re interested in what Overclock is doing, take a look at <a href="https://overclockaccelerator.com/">overclockaccelerator.com</a>. I’ll keep posting weekly reflections, but happy to answer questions if you’re thinking about applying.</p><p><strong>Want to follow along?</strong><br> I’ll be sharing a post every week as the program continues. Hit <em>Follow</em> if you’re curious to see what comes next.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=16df60f59770" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The Art of Flowgramming: Building Without a Blueprint]]></title>
            <link>https://medium.com/@louritter1/the-art-of-flowgramming-building-without-a-blueprint-f3c8f807bbd7?source=rss-ca535acb0e04------2</link>
            <guid isPermaLink="false">https://medium.com/p/f3c8f807bbd7</guid>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[javascript]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[coding]]></category>
            <dc:creator><![CDATA[Lou Ritter]]></dc:creator>
            <pubDate>Tue, 20 May 2025 04:30:49 GMT</pubDate>
            <atom:updated>2025-05-20T04:30:49.308Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ofgjL_dMx2zGF345yA_NkQ.png" /></figure><p>Most side projects start with a pitch. A plan. A problem to solve.</p><p>Lately, mine haven’t.</p><p>They’ve started with an idea that barely makes sense, a half-formed prompt, and a sentence that starts with “wouldn’t it be cool if…?”</p><p>First there was an AI iPaaS. Then, a real-time flight tracker built with Kafka and Snowflake. There was an assistant for transforming different pieces of content. There was a dashboard for data that didn’t exist yet. I could go on. And on. And on. And on.</p><p>None of them started with a spec. They all began with curiosity and code.<br> Some of them, maybe even most, will never be finished. And that’s okay.</p><p>I’m building these for me, because I find them interesting.<br>If I stop finding them interesting halfway through, I move on to the next shiny idea that enters my brain.</p><p>There’s something to be said for seeing things through. But right now, I’m building to build.<br>Letting things unfold. Following the thread.</p><p>This isn’t a job. It’s a stream of consciousness.<br>It’s a kind of artistic expression.</p><p>That’s where Flowgramming began: building without a blueprint, letting the keyboard steer, and trusting that something interesting will emerge along the way.</p><p><strong>What Is Flowgramming?</strong></p><p>It’s not a framework or a stack. It’s a way of building quickly, intuitively, and sometimes recklessly.</p><p>You start with a vibe, not a roadmap. You code what feels useful. You break things. You realize what you’re building halfway through building it. You finish it because it’s fun.</p><p>Flowgramming means using tools like LLMs, lightweight runtimes, and declarative configs not to scale, but to explore.</p><p><strong>Why It’s Not Just Vibe Coding</strong></p><p>Flowgramming may sound like a cousin to vibe coding, but it’s different.</p><p>Vibe coding is spontaneous. It’s about spinning up a quick idea without much care for structure, longevity, or what happens after the commit.</p><p>Flowgramming carries the same spirit of momentum, but with intent. There is still craft involved. You name things. You shape the architecture. You build systems that could last, even if you’re not sure they will.</p><p>It’s playful, but not careless. You follow the thread, but you do it thoughtfully.</p><p>The goal is not just to get something working, but to build something that feels good to work on.</p><p><strong>Why Build Like This?</strong></p><p>Most projects, especially in a job setting, follow a waterfall or agile rhythm: plan, build, test, ship. But flowgramming is more like jazz. The architecture changes every other night. Features emerge from bugs. Naming things appropriately happens last (if at all).</p><p>There are advantages to building in the open, especially when the end goal is unclear:</p><ul><li><strong>Momentum &gt; Perfection.</strong> You don’t get stuck optimizing early. You ship weird ideas and see what sticks.</li><li><strong>Discovery through doing.</strong> You don’t know what the product <em>should</em> be until you start breaking it.</li><li><strong>AI as a partner.</strong> Using LLMs not just in the product, but <em>for</em> the product as a pair programmer, debugger, even an idea sparring partner.</li></ul><p>But it comes with tradeoffs:</p><ul><li>No clear finish line</li><li>Hard to onboard contributors</li><li>Constant context switching</li></ul><p>Still, for a solo dev project, it’s freeing. You build with instinct instead of a Gantt chart.</p><p><strong>Anatomy of a Flowgrammed Project</strong></p><p>Whether it’s a meme generator or a data streaming pipeline, they tend to follow the same rhythm:</p><ol><li>Get data from somewhere — an API, a feed, a prompt</li><li>Do something clever to it — transform, route, visualize</li><li>Keep it flexible, keep it hackable, keep it lightweight</li></ol><p>Sometimes it ends in a blog post. Sometimes it ends in an archive folder. But every time, I learn something. Something about code or maybe something I enjoy working with. Something is gained even if there is nothing to show for it.</p><p><strong>The Point Isn’t the Product</strong></p><p>There are plenty of tools more polished, more performant, more ready for scale.</p><p>But this isn’t about shipping startups. It’s about building like you have nothing to prove. Fast, expressive, unstructured but without sacrificing craft.</p><p>It’s an experiment in developer UX, in co-building with AI, and in trusting your gut while building something weird.</p><p><strong>Next Steps</strong></p><p>A few of these projects might become more. Others won’t.<br> The schemas could be tighter. The tests could be written. The README could be readable.</p><p>But they work.</p><p>And I’ll keep building out my ideas like this, one idea at a time.</p><p><em>Follow me for more builds, ideas, and experiments that start with “what if” and end with something you can run.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f3c8f807bbd7" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[“Just Wondering If You’ve Got a Second”]]></title>
            <link>https://medium.com/@louritter1/just-wondering-if-youve-got-a-second-1309f296404c?source=rss-ca535acb0e04------2</link>
            <guid isPermaLink="false">https://medium.com/p/1309f296404c</guid>
            <category><![CDATA[programming]]></category>
            <category><![CDATA[coding]]></category>
            <category><![CDATA[debugging]]></category>
            <category><![CDATA[software-engineering]]></category>
            <category><![CDATA[technology]]></category>
            <dc:creator><![CDATA[Lou Ritter]]></dc:creator>
            <pubDate>Tue, 13 May 2025 12:41:48 GMT</pubDate>
            <atom:updated>2025-05-13T12:41:48.587Z</atom:updated>
            <content:encoded><![CDATA[<p>A short story about reaching out, pairing up, and fixing it alone anyway.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VF5oBp-VUBm3NPj1SlfwSw.png" /></figure><p>The bug doesn’t seem big.</p><p>Just one endpoint. A 500 error. You didn’t touch the logic, and yet it’s broken.</p><p>You’ve already done 99% of the work. The PR template is filled out and you were so close to being ready for review. At some point in the clean up things went downhill and you started seeing red everywhere.</p><p>You refresh. Scan logs. Stare into your editor like it’ll explain itself.</p><p>Nothing moves.</p><p>After 45 minutes of circling the same mental drain, you open Slack.</p><p>Type. Backspace. Type again.</p><p>“hey, anyone have a few minutes to look at something weird?”</p><p>A reply comes in fast:</p><p>“sure! Zoom?”</p><p>You paste the link. Click join.</p><p>The camera turns on.</p><p>Before you even share your screen, you catch your reflection.</p><p>Hair flattened in the back, fluffed in the front. A reminder of all the head-scratching you’ve done today.</p><p>You look tired. Not in a tragic way. Just honestly tired.</p><p>Eyes with the thousand-yard stare of someone who’s looked at the same block of code for too long without blinking.</p><p>You’re thankful they can’t see your desk. A half-drank coffee, a banana peel, and a scatter of handwritten notes that look more like conspiracy theories than a plan.</p><p>You fix nothing. You hit “Share Screen.”</p><p>They join.</p><p>You give the quick summary: what it’s supposed to do, what it’s doing instead, and the twenty things you’ve already tried.</p><p>They nod. You know they don’t understand it all but they are going to try their best.</p><p>You both dig in.</p><p>Controllers. Services. Helper files that smell like 2019 and long forgotten todo’s.</p><p>You try logging. They try guessing. You delete things, retype them. You say “just to check” way more than you’d like.</p><p>The silence gets longer between ideas.</p><p>Finally, they say what you’ve both been thinking:</p><p>“I don’t see it either. But ping me if you crack it. Now I’m invested.”</p><p>You thank them and you mean it.</p><p>They leave the call.</p><p>Your screen goes back to just you.</p><p>For a moment, you just sit.</p><p>Still sharing. Still confused. Still staring.</p><p>Then, for no particular reason, you scroll the file again. Slower this time.</p><p>Line by line.</p><p>And there it is.</p><p>A comma.</p><p>Wrong place. Wrong object. Wrong enough to quietly break everything.</p><p>You delete it.</p><p>Refresh.</p><p>It works.</p><p>No celebration. Just relief that feels almost indistinguishable from sadness.</p><p>You send them a message:</p><p>“dude… found it. extra comma. appreciate you hopping on. Seriously.”</p><p>You close the call window.</p><p>Catch your reflection again.</p><p>Still tired.</p><p>Still human.</p><p>But no longer stuck.</p><p>A fleeting feeling of victory.</p><p>This is the part they don’t show in the job description. The quiet, frustrating, oddly satisfying moment when you beat the thing that almost beat you.</p><p>It’s painful. It’s humbling. But it’s also what keeps you coming back.</p><p>You have to take the small wins.</p><p>You have to enjoy the fix no one else saw.</p><p>Because sometimes, that’s the job. And sometimes, that’s enough.</p><p><em>Follow for more posts on building, debugging, and everything in between.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1309f296404c" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Data Streaming in a Day: Flight Data from Kafka to Snowflake on Fly.io]]></title>
            <link>https://medium.com/@louritter1/data-streaming-in-a-day-flight-data-from-kafka-to-snowflake-on-fly-io-a00b0d1b0f6d?source=rss-ca535acb0e04------2</link>
            <guid isPermaLink="false">https://medium.com/p/a00b0d1b0f6d</guid>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[snowflake]]></category>
            <category><![CDATA[kafka]]></category>
            <category><![CDATA[data]]></category>
            <dc:creator><![CDATA[Lou Ritter]]></dc:creator>
            <pubDate>Tue, 06 May 2025 17:22:02 GMT</pubDate>
            <atom:updated>2025-05-06T17:23:47.407Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*s4YvrTqXc57r7y7pS9PBlA.png" /></figure><p>Streaming real-time data might sound like something reserved for hedge funds or companies with a cloud bill the size of a private jet. But you can actually build a working end-to-end pipeline, from public APIs to Kafka to Snowflake to dashboard, in less than a day. And no, you don’t need a fleet of microservices. Just a few smart, clean, choices and some thoughtful engineering.</p><p>In this post, I’ll show you how I built a real-time flight tracking pipeline using:</p><ul><li>A public flight data API (OpenSky Network)</li><li>A simple Kafka producer running on Fly.io</li><li>Confluent Cloud (Kafka-as-a-service)</li><li>Snowflake as the destination warehouse</li><li>A Streamlit dashboard for visualization</li></ul><p>We kept things simple and focused on proving the concept. This isn’t about simulating an airline control center (although that could be very useful right now), it’s about showing that you can stitch together modern tools to stream live aircraft data into a warehouse and query it like it’s just another Tuesday.</p><p>check out the demo here: <a href="https://flacker.streamlit.app/">https://flacker.streamlit.app</a></p><p><strong>🏗️ Architecture Overview</strong></p><p><strong>1. Data Source: OpenSky Network</strong><br>OpenSky provides live aircraft state data with an open, unauthenticated API. It’s free, raw, and the data is surprisingly good. That said, my EC2 instance couldn’t connect to it. After some digging and debugging, I concluded OpenSky must be treating AWS like spam, kind of like how your inbox handles LinkedIn cold DMs.</p><p>So I moved to <strong>Fly.io</strong>, which not only worked out of the box but also felt weirdly appropriate for a flight-tracking pipeline.</p><p><strong>2. Kafka Producer (Fly.io):</strong><br>We deploy a Python script using confluent-kafka to Fly.io. Every 10 minutes, it wakes up, grabs data from OpenSky, and fires it into a Kafka topic. It&#39;s containerized and stays out of the way.</p><pre>            res = requests.get(url)<br>            if res.status_code == 200:<br>                data = res.json().get(&#39;states&#39;, [])<br>                buffer = []<br>                for flight in data[:50]:  # Consider up to 50 flights<br>                    if not flight or flight[0] in sent_ids:<br>                        continue<br>                    payload = {<br>                        &#39;icao24&#39;: flight[0],<br>                        &#39;callsign&#39;: flight[1],<br>                        &#39;origin_country&#39;: flight[2],<br>                        &#39;longitude&#39;: flight[5],<br>                        &#39;latitude&#39;: flight[6],<br>                        &#39;velocity&#39;: flight[9]<br>                    }<br>                    buffer.append(payload)<br>                    sent_ids.add(flight[0])<br>                    if len(sent_ids) &gt; 500:<br>                        sent_ids = set(list(sent_ids)[-250:])  # Trim memory<br><br>                for record in buffer:<br>                    producer.produce(topic, value=json.dumps(record).encode(&#39;utf-8&#39;))<br>                    producer.poll(0)<br>                    print(f&quot;Produced flight: {record[&#39;icao24&#39;]}&quot;)<br>                producer.flush()</pre><p><strong>3. Kafka Topic (Confluent Cloud):</strong><br>No servers to babysit. Just a topic named kafka_flight living in Confluent Cloud, happily waiting for messages.</p><p><strong>4. Kafka Connector (Confluent → Snowflake):</strong><br>Here’s where the magic happens. Confluent provides a managed Snowflake Sink Connector that we configure through the Confluent Cloud UI. You just specify your topic, Snowflake account, database, schema, and a few other connection details. Confluent handles the heavy lifting — batching, formatting, and streaming the data into Snowflake using Snowpipe Streaming. You don’t need to run any infrastructure for this, which is a win.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1006/1*zQL9xGWGOUjEj0-c-7uoww.png" /></figure><p><strong>5. Snowflake:</strong><br>We created a table with a single VARIANT column to ingest raw JSON. It accepts the structure without a fuss. No ETL here, just raw stream data, landing directly in a warehouse. If this were production-ready, I’d start layering dbt models on top to model and clean it.</p><p><strong>6. Streamlit Dashboard:</strong><br>Finally, a simple Streamlit app pulls from Snowflake and renders a real-time view of the skies. You can filter flights by speed and country, and see them mapped out live. I can’t sing the praises of Streamlit loud enough, it’s a godsend for quick, clean demos.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*W1mgSMgIfry766drxJSXMQ.png" /></figure><p><strong>💻 Why These Tools?</strong></p><ul><li><strong>OpenSky Network</strong>: Real-time global aircraft data, no login required.</li><li><strong>Fly.io</strong>: Let’s be honest, deploying a flight tracker to something named Fly.io just makes sense. Plus, it handles small containers really well and you will have less turbulence than running it on an ec2 instance.</li><li><strong>Confluent Cloud</strong>: Fully-managed Kafka makes your life easier.</li><li><strong>Snowflake</strong>: Easy to ingest semi-structured data and scale when needed.</li><li><strong>Streamlit</strong>: Zero boilerplate dashboards with an actually decent UI.</li></ul><p><strong>🧠 What We’re NOT Doing (But Could)</strong></p><ul><li>No schema evolution or Avro support (we’re just streaming raw JSON)</li><li>No dbt modeling or warehouse transformations… yet</li><li>No monitoring or alerting</li><li>No event-driven architecture beyond the connector</li></ul><p>This is a proof-of-concept meant to show how simple the modern stack can be when you don’t try to overcomplicate it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tn-hQ_k0qDWr3Kvj6Xy8jA.png" /></figure><p><strong>📦 Running the Producer</strong></p><p>We deployed the producer using Docker on Fly.io, using this simple Dockerfile:</p><pre>FROM python:3.10-slim<br>WORKDIR /app<br>COPY requirements.txt .<br>RUN pip install --no-cache-dir -r requirements.txt<br>COPY flight_producer.py .<br>CMD [&quot;python&quot;, &quot;flight_producer.py&quot;]</pre><p>To keep things secure, we stored Kafka credentials in Fly.io secrets instead of baking them into the image.</p><p><strong>📈 The Dashboard</strong></p><p>The Streamlit dashboard queries Snowflake and maps aircraft in near real-time. It color-codes flights by velocity and lets you filter by origin country or speed threshold. I can’t be running this 24/7 due to the costs I would incur, but I ran it for a while with nohup to verify the stream works as it should.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*ndseWifdZ7iztqEKNyrHGA.png" /></figure><p><strong>🧪 Conclusion</strong></p><p>We’ve got a working real-time pipeline: public flight data into Kafka, streamed into Snowflake, and visualized on a dashboard. It’s fast to build, reasonably robust, and best of all? Composed entirely of managed tools. This project wasn’t about scale or production readiness. It was about proving that we can glue together a modern stack with just enough code and just the right abstraction.</p><p>Would I build a global aviation system this way? Not without a few upgrades. But for a side project or a technical proof-of-concept? This flies (groan).</p><p><em>Follow for more dev builds that blend real-world tech solutions, fun tutorials, and just enough puns to stay airborne.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=a00b0d1b0f6d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Why Albert Camus Would Thrive in Modern Tech]]></title>
            <link>https://medium.com/@louritter1/why-albert-camus-would-thrive-in-modern-tech-e0a9107bccdb?source=rss-ca535acb0e04------2</link>
            <guid isPermaLink="false">https://medium.com/p/e0a9107bccdb</guid>
            <category><![CDATA[code]]></category>
            <category><![CDATA[creative-writing]]></category>
            <category><![CDATA[software-engineering]]></category>
            <category><![CDATA[philosophy]]></category>
            <category><![CDATA[software-development]]></category>
            <dc:creator><![CDATA[Lou Ritter]]></dc:creator>
            <pubDate>Tue, 29 Apr 2025 21:25:08 GMT</pubDate>
            <atom:updated>2025-04-30T00:11:20.896Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*uRsgXvli2DoURoa8tDidrQ.png" /></figure><p>If Albert Camus were alive today, he wouldn’t be writing existential essays in a cafe. He’d be scratching his head and silently watching CI pipelines fail for no reason. We should all channel our inner Camus and embrace the chaos that is working in tech in these modern times.</p><p>Tech in 2025 is absurd. We know it. And he’d know it.</p><p><strong>The Absurdity of the Stack</strong></p><p>Camus’ whole philosophy starts with one truth: life is absurd. It lacks meaning, structure, and predictability. You can fight it, or you can embrace it.</p><p>Now apply that to JavaScript frameworks.</p><p>You spend six months learning the “new standard.” It gets deprecated. You rebuild everything for edge rendering. The specs change again. Welcome to the modern dev lifecycle, where certainty is an illusion and your stack is Sisyphus’s boulder.</p><p>The parallels between Camus views of the world couldn’t be more pertinent when it comes to technology. Tech workers are constantly trying to find order, believing that ultimately there is order. There may be a degree of order, but as I and anyone else in tech know, underlying the order there is a web of countless bugs, missed timelines, and tech debt.</p><p>Camus wouldn’t rage against it. He’d shrug, update his packages, and keep pushing.</p><p>“One must imagine Sisyphus happy.”</p><p><strong>Sisyphus: The Original Developer</strong></p><p>In <em>The Myth of Sisyphus</em>, Camus reimagines the condemned figure endlessly rolling a boulder up a hill. He says we should imagine Sisyphus as happy. Rather than pity in his sorrows, he’d smile, roll up his sleeves, and get to work. We should all be more like Sisyphus.</p><p>That’s software engineering. That’s shipping products. That’s fixing the same bug for the fourth time because a package version deep in your dependency tree reverted or a semi colon was missed (somehow it still happens to me in 2025).</p><p>The boulder rolls back. You push it again. Somewhere in that cycle, you find meaning. Maybe not in the task, but in the craftsmanship, the rhythm, the stubbornness.</p><p>Camus would be an incredible engineer because he wouldn’t chase control. He’d find peace in the push. He’d look at his spaghetti code and console logs and laugh.</p><p><strong>He wouldn’t be writing his thoughts on Medium</strong></p><p>Camus wouldn’t be dunking on tailwind.css or debating tabs vs. spaces. He’d be building. Maybe solo shipping a weird personal site that logs dreams. Maybe writing docs so good they make you cry.</p><p>He’d quietly outlast all of us. Like those devs you work with that you swear spend their sleeping hours reading documentation, except Camus would have much more cigarettes.</p><p><strong>Conclusion: Embrace the Absurd Stack</strong></p><p>The modern tech landscape is a treadmill of shifting standards, overpromises, and recursive error messages. But you still show up. You still create that pull request. You still learn weird edge-case behavior because something in you refuses to quit.</p><p>Camus would get it. He’d hit npm install, watch it break, and write a beautiful essay about it.</p><p>One must imagine the full stack engineer happy.</p><p><em>Keep pushing that boulder.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e0a9107bccdb" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[The 2025 NFL Draft as Dev Tools: Player Comps For The Top Picks in This Years Draft]]></title>
            <link>https://medium.com/@louritter1/the-2025-nfl-draft-as-dev-tools-player-comps-for-the-top-picks-in-this-years-draft-1701936b20db?source=rss-ca535acb0e04------2</link>
            <guid isPermaLink="false">https://medium.com/p/1701936b20db</guid>
            <category><![CDATA[nfl]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[web-development]]></category>
            <category><![CDATA[web3]]></category>
            <dc:creator><![CDATA[Lou Ritter]]></dc:creator>
            <pubDate>Tue, 29 Apr 2025 21:05:50 GMT</pubDate>
            <atom:updated>2025-04-29T21:05:50.171Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*F1h75pePQyMIii2MTkI0kA.png" /></figure><p>The NFL Draft is basically the Super Bowl for people who enjoy spreadsheets, player comparisons (“this guy is the next Brady”), and unearned confidence. And honestly, it’s not that different from tech. Teams chase upside, take big swings, and occasionally draft a total bust that looked great on paper (looking at you, Meteor.js).</p><p>So what if this year’s top NFL prospects were emerging dev tools or tech trends? Let’s build a scouting report from the command line.</p><p><strong>Shedeur Sanders = Web3 Infrastructure</strong></p><p>At one point, he was <em>the</em> projected number one pick. High completion percentage. Confident. Charismatic. But on draft day? He slid. The Browns scooped him up late, and now the world is split: he’ll either be the steal of the draft or the guy who proved the doubters right.</p><p>That’s Web3 in 2025. Once billed as the savior of the internet, now fighting to prove it belongs in real-world apps. The talent is there. The swagger is undeniable. But it’s still got a lot to prove when the pressure’s on.</p><p><strong>Jaxson Dart = Bun</strong></p><p>The Giants traded up for him. Bold move. Some thought it was a reach, but the team clearly believes. He’s exciting, mobile, and fans are already photoshopping him into playoff scenarios.</p><p>Bun is the Dart of dev tools. It came out swinging with a faster JavaScript runtime, a sleek dev experience, and the promise of simplifying everything. Is it ready to lead a big team? Not sure yet. But the potential is real, and the hype feels earned.</p><p><strong>Cam Ward = LangChain</strong></p><p>He went number one. Some say it came out of nowhere, but if you were paying attention, the talent was obvious. The Titans already had a QB, but they bet the house on Cam. Now it’s all eyes on his development.</p><p>LangChain is the Cam Ward of dev right now. It’s become the face of LLM orchestration. Big-time promise, rapid rise, and the industry is leaning in. Time will tell if it can command an offense or just look good on demo day.</p><p><strong>Travis Hunter = Vercel (Full Stack Framework + Platform)</strong></p><p>The flashiest player in the draft, and somehow also the most grounded. He plays both ways, a lockdown corner plus a dynamic wide receiver. He smiles a lot, works harder than most, and everyone seems to love him.</p><p>That’s Vercel. Frontend, backend, hosting, and edge delivery all in one. It’s your wide receiver and your lockdown corner. Stylish and technical. It might look like just a pretty deploy button, but behind the scenes it’s winning matchups on both sides of the ball. Every dev wants it on their team.</p><p><strong>Overtime</strong></p><p>The draft, like tech, is all about potential. Some of these picks will thrive. Some will flame out. And some will quietly become foundational to how the game is played.</p><p>Whatever you’re building, just make sure you’re not trading up for a broken CI pipeline.</p><p><em>Follow me for more takes at the intersection of code, culture, and occasionally good analogies. I try to write a blog every day about tech, tech jobs, and much more.</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=1701936b20db" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Technical Debt Explained Like a Dirty Apartment]]></title>
            <link>https://medium.com/@louritter1/technical-debt-explained-like-a-dirty-apartment-78996b0620b7?source=rss-ca535acb0e04------2</link>
            <guid isPermaLink="false">https://medium.com/p/78996b0620b7</guid>
            <category><![CDATA[technical-debt]]></category>
            <category><![CDATA[software-development]]></category>
            <category><![CDATA[code]]></category>
            <category><![CDATA[technology]]></category>
            <category><![CDATA[software-engineering]]></category>
            <dc:creator><![CDATA[Lou Ritter]]></dc:creator>
            <pubDate>Mon, 28 Apr 2025 17:14:13 GMT</pubDate>
            <atom:updated>2025-04-28T17:22:02.881Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*SzUZKuBAforyYljVpeJOlA.png" /></figure><p>It starts small. A few pieces of clothing scattered on the floor, a couple dirty dishes left over from having friends over before your night out. Then suddenly, it’s Sunday evening. You spent the day procrastinating, catching up on White Lotus so it wouldn’t get spoiled on X. The dishes pile up, the groceries run out, and you’re staring down another work week. You look around in horror and think, “How did we get here?”</p><p>Welcome to the world of technical debt.</p><p><strong>What is Technical Debt?</strong></p><p>Tech debt is just like that dirty apartment. It starts small, usually hidden behind a shiny new feature or a tight deadline. You or your team write the code, and it “just works.” You can’t explain it, but the requirements are hit. Ship it. You write faulty tests (or sometimes, no tests), copy-paste from another component, and call it good enough. For now.</p><p>Before long, the consequences rear their ugly head. Shame and embarrassment set in.</p><p>In your apartment, it means a bad smell, a sense of shame, or a relative seeing the mess and never looking at you the same way again. In your software, it means bugs, a sense of shame, and coworkers side-eyeing your pull requests.</p><p>In all seriousness, technical debt is a necessary evil for engineers. It’s something we all have to handle. It’s not about avoiding it entirely, but about knowing when and how to tackle it.</p><p><strong>Why We Ignore It (Until It’s Too Late)</strong></p><p>In your apartment:</p><ul><li>You’re busy.</li><li>You’re tired.</li><li>You’re not inviting anyone over anyway.</li></ul><p>In your codebase:</p><ul><li>You’re chasing deadlines.</li><li>You’re adding features.</li><li>You’re praying no one asks why new deployments smell suspiciously like burning.</li></ul><p>Ignoring the mess feels easier than facing the painful, embarrassing cleanup.</p><p>Until one day, your landlord or your CTO demands an inspection.</p><p>Oops.</p><p><strong>Paying Off the Debt</strong></p><p>Cleaning the apartment:</p><ul><li>Scrub everything.</li><li>Throw away junk.</li><li>Maybe light a scented candle and cry a little.</li></ul><p>Paying off technical debt:</p><ul><li>Refactor old code.</li><li>Add missing tests.</li><li>Document what you should’ve documented six months ago.</li></ul><p>Neither process is fun. Both are absolutely necessary if you want to live or build software like a functional adult.</p><p><strong>Technical Debt Isn’t Always Bad</strong></p><p>Quick reality check:</p><ul><li>Technical debt can be strategic.</li><li>Sometimes you should move fast and clean up later.</li><li>The real danger is forgetting you ever made those compromises.</li></ul><p>If you never plan the “spring cleaning” sprint, the mess stops being a minor inconvenience and starts being the reason your project can’t scale, can’t pivot, or can’t survive.</p><p><strong>Signs You Might Be Living With Technical Debt</strong></p><ul><li>No one knows what that critical piece of legacy code actually does.</li><li>Deploying a simple change feels like pulling a Jenga block from the bottom row.</li><li>New developers need a map, compass, and therapy sessions to onboard.</li><li>Your unit tests cover everything except reality.</li></ul><p><strong>Conclusion</strong></p><p>Whether it’s code or your kitchen, messes don’t disappear. They just get better at hiding until they become impossible to ignore.</p><p>Clean early. Clean often. And if your software starts to smell funny, check under the pizza boxes.</p><p>You might just find the bug that’s been haunting you.</p><p><em>If you liked this post, follow my blog for more takes like this. Whether it’s a light-hearted dive into tech concepts or a deeper technical breakdown, there’s a lot more coming. Thanks for reading!</em></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=78996b0620b7" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>