<?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 Nicolas Bourbaki on Medium]]></title>
        <description><![CDATA[Stories by Nicolas Bourbaki on Medium]]></description>
        <link>https://medium.com/@nicolas.bourbaki.1912?source=rss-85f333a0406e------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*_7c1km_SGYcCSpY9fhVvUg.png</url>
            <title>Stories by Nicolas Bourbaki on Medium</title>
            <link>https://medium.com/@nicolas.bourbaki.1912?source=rss-85f333a0406e------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sun, 22 Nov 2026 23:39:10 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@nicolas.bourbaki.1912/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[# The Object-Oriented Revolution in AI Agent Architecture]]></title>
            <description><![CDATA[<div class="medium-feed-item"><p class="medium-feed-snippet">The developer talent gap in AI is staggering but presents an unprecedented opportunity. Current data shows **8&#x2013;12 million web developers&#x2026;</p><p class="medium-feed-link"><a href="https://medium.com/@nicolas.bourbaki.1912/the-object-oriented-revolution-in-ai-agent-architecture-919582d9625c?source=rss-85f333a0406e------2">Continue reading on Medium »</a></p></div>]]></description>
            <link>https://medium.com/@nicolas.bourbaki.1912/the-object-oriented-revolution-in-ai-agent-architecture-919582d9625c?source=rss-85f333a0406e------2</link>
            <guid isPermaLink="false">https://medium.com/p/919582d9625c</guid>
            <dc:creator><![CDATA[Nicolas Bourbaki]]></dc:creator>
            <pubDate>Thu, 26 Feb 2026 21:59:11 GMT</pubDate>
            <atom:updated>2026-02-26T21:59:11.588Z</atom:updated>
        </item>
        <item>
            <title><![CDATA[#35 The “Hidden” Roadmap of Microsoft Copilot: From Helper to Replacement]]></title>
            <link>https://medium.com/@nicolas.bourbaki.1912/35-the-hidden-roadmap-of-microsoft-copilot-from-helper-to-replacement-84afe3bde661?source=rss-85f333a0406e------2</link>
            <guid isPermaLink="false">https://medium.com/p/84afe3bde661</guid>
            <category><![CDATA[future-of-work]]></category>
            <category><![CDATA[microsoft-skills]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[copilot-studio]]></category>
            <category><![CDATA[microsoft-recall]]></category>
            <dc:creator><![CDATA[Nicolas Bourbaki]]></dc:creator>
            <pubDate>Fri, 02 May 2025 10:10:26 GMT</pubDate>
            <atom:updated>2025-05-02T10:10:26.283Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*yCgCrUU7Yn1wJ80p3MjTzA.png" /><figcaption>ai generated image</figcaption></figure><h3>#35 I Think Microsoft Is Training Agents to Replace You (Not Help You)</h3><p>First came Copilot. Then came “People Skills.” And now “Recall.”</p><p>At first glance, these tools seem innocent — even helpful. Track what you do, surface your strengths, help you focus. Who wouldn’t want that?</p><p>But taken together, I can’t shake the feeling: this isn’t about <em>helping</em> workers. It’s about <em>recording</em> them. Training on them. And, eventually, <em>replacing</em> them.</p><h3>📊 From Assistance to Imitation</h3><p>In <strong>March 2024</strong>, Microsoft rolled out <a href="https://techcommunity.microsoft.com/t5/microsoft-365-copilot-blog/announcing-people-skills-general-availability-and-new-skills/ba-p/4406364?utm_source=chatgpt.com">“People Skills” for Microsoft 365 Copilot</a>, a feature designed to infer what you’re good at based on your daily actions in documents, meetings, emails, and chats.</p><p><a href="https://techcommunity.microsoft.com/blog/microsoft365copilotblog/announcing-people-skills-general-availability-and-new-skills-agent/4406364">Announcing People Skills general availability and new Skills agent | Microsoft Community Hub</a></p><p>It doesn’t just analyse. It <strong>labels</strong> you. It builds a skills profile, one that managers — and algorithms — can use to assess, sort, or deploy people like modular assets.</p><p>At the same time, Microsoft announced <strong>“Skills Agent”</strong>, which uses these profiles to guide workforce planning. It’s sold as visibility. But if AI can see who’s best at what… then it can also start <strong>doing</strong> it.</p><p>Then came <strong>Recall</strong> — the <a href="https://www.theverge.com/news/656106/microsoft-recall-copilot-plus-pc-available?utm_source=chatgpt.com">controversial feature</a> launched with Copilot+ PCs. It takes snapshots of nearly everything you do on your machine and creates a searchable timeline. Your past activities become one big indexed log.</p><p><a href="https://www.theverge.com/news/656106/microsoft-recall-copilot-plus-pc-available?utm_source=chatgpt.com">Microsoft launches Recall and AI-powered Windows search for Copilot Plus PCs</a></p><p>Even <a href="https://www.tomshardware.com/software/windows/microsoft-launches-recall-to-windows-11-general-availability-click-to-do-and-improved-search-also-coming?utm_source=chatgpt.com">Tom’s Hardware</a> and <a href="https://www.windowscentral.com/software-apps/windows-11/windows-recall-general-availability-2025-copilot?utm_source=chatgpt.com">Windows Central</a> flagged privacy concerns — but also noted the long-term productivity vision.</p><h3>🧩 I Theorise: This Is Pretraining for Copilot Agents</h3><p>This isn’t just feature creep. It’s feature <strong>capture</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*jQcKAjUAEaxo2dcJbzqfxg.png" /><figcaption>observe, classify, and then replace.</figcaption></figure><p>I believe Microsoft is using these systems to <strong>record and classify enterprise workflows</strong> — and eventually use them to <strong>pretrain agent models</strong> capable of performing those workflows independently.</p><p>Not just prompting Copilot to help with Excel… but enabling Copilot to <em>be</em> the Excel user. The project planner. The ops assistant.</p><p>A ghost version of you, doing your most repeatable tasks — trained on your behaviour.</p><h3>🛠️ The New OS: Observe → Predict → Replace</h3><p>What Microsoft is quietly building is not just tooling — it’s a pipeline:</p><ul><li><strong>Observe</strong> → via Recall (track activity)</li><li><strong>Classify</strong> → via People Skills (profile capability)</li><li><strong>Replace</strong> → via future Copilot agents (autonomous action)</li></ul><p>And while the <a href="https://techcommunity.microsoft.com/t5/microsoft-365-copilot-blog/announcing-people-skills-general-availability-and-new-skills/ba-p/4406364?utm_source=chatgpt.com">official OKRs and roadmaps</a> focus on boosting productivity and aligning talent, you don’t have to squint hard to see the business logic:</p><ul><li>More automation</li><li>Fewer repeatable roles</li><li>Scalable intelligence across teams</li></ul><h3>👀 What Should We Be Asking?</h3><p>This might help some workers. But it might also replace many more.</p><p>When the tools you use begin learning from everything you do — and then replicate it without needing you — we need to stop and ask:</p><blockquote><em>Are we optimising for productivity… or obsolescence?</em></blockquote><p>Because the future may not involve AI <em>helping</em> you with work.</p><p><strong>It may involve AI <em>doing</em> your work — based on the way you did it yesterday.</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=84afe3bde661" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[#34 Your Job Is Being Recorded: AI Isn’t Helping You — It’s Replacing You]]></title>
            <link>https://medium.com/@nicolas.bourbaki.1912/34-your-job-is-being-recorded-ai-isnt-helping-you-it-s-replacing-you-32425b1beb9f?source=rss-85f333a0406e------2</link>
            <guid isPermaLink="false">https://medium.com/p/32425b1beb9f</guid>
            <category><![CDATA[kortix-ai]]></category>
            <category><![CDATA[future-of-work]]></category>
            <category><![CDATA[critical-thinking]]></category>
            <category><![CDATA[work-displacement]]></category>
            <category><![CDATA[ai-agent]]></category>
            <dc:creator><![CDATA[Nicolas Bourbaki]]></dc:creator>
            <pubDate>Fri, 02 May 2025 09:42:17 GMT</pubDate>
            <atom:updated>2025-05-02T09:42:17.829Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*hwhYReePgJCLc9mJIKKJ_g.png" /><figcaption>ai generated image</figcaption></figure><h3><strong>#34 Your Job Is Being Recorded: AI Isn’t Helping You — It’s Replacing You</strong></h3><h3><strong>1. The Opening: Not Surprising, Just Inevitable</strong></h3><p>Companies are now offering to <strong>record everything your employees do</strong> — from emails and chats to workflows and micro-decisions — and turn it into AI automation. Not “productivity software.” Not “process improvement.” Just plain substitution.</p><p>Companies like <strong>Kortex AI</strong> (who released <a href="https://www.suna.so/">Suna.co</a>) say it plainly: record what your team does, document it into repeatable processes, automate it.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/762/1*xySBcH0nEsXwokFcCQLE-Q.png" /><figcaption><a href="https://www.suna.so/">https://www.suna.so/</a></figcaption></figure><p>It’s not shocking. It’s not dystopian. It’s just <strong>expected</strong>.</p><p>Like that mate who always insists on ‘just one more.’ It’s going to happen. You already know it. It’s not even closing time.</p><h3><strong>2. The AI Shift Timeline</strong></h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*x33KXIVwKUS4TLnTIC-_1Q.png" /></figure><ul><li><strong>Discover AI</strong>: One person, one tool. Early curiosity.</li><li><strong>Co-intelligence</strong>: Everyone has an AI buddy. Looks empowering.</li><li><strong>Agent Boss</strong>: Fewer people. More bots. One person managing five agents.</li></ul><p>This isn’t a future prediction. It’s a path we’ve been walking for a while.</p><p>I’ve already written about this in:</p><blockquote><a href="https://medium.com/@nicolas.bourbaki.1912/ai-co-intelligence-ethans-comforting-myth-that-won-t-save-your-job-d79e4e6a28cd">AI Co-Intelligence: Ethan’s Comforting Myth That Won’t Save Your Job</a></blockquote><blockquote>A takedown of the feel-good idea that humans and AI will just work together harmoniously — arguing instead that co-intelligence <strong>is just a polite transition before replacement</strong>.</blockquote><blockquote><a href="https://medium.com/@nicolas.bourbaki.1912/how-to-train-your-replacement-or-confuse-it-for-fun-9bfab5c97a4d">Fired But Not Gone</a></blockquote><blockquote>Explores how companies no longer need to keep you if they can keep your workflows — your role gets captured, stored, and re-run without you. i.e., what Kortix are offering.</blockquote><blockquote><a href="https://medium.com/@nicolas.bourbaki.1912/8-i-never-expected-to-grieve-over-my-profession-welcome-to-ai-in-2025-68b663f5f9c6">Skynet AI Won’t Kill You</a></blockquote><blockquote>Pushes back on the sci-fi panic and reframes the real threat: AI doesn’t need to rebel to ruin your career — it just needs to make your tasks look repetitive.</blockquote><p>This article just catches people up. The next stage isn’t hypothetical. It’s onboarding.</p><h3><strong>3. Microsoft Quietly Confirms It</strong></h3><p>This isn’t limited to startups. <strong>Microsoft</strong> recently released <strong>People Skills</strong> inside 365 Copilot. It infers your skill set by watching how you work.</p><p>Your email habits, meeting behaviour, spreadsheet activity — it’s all forming a profile. A picture of your tasks, your decision-making, your repeatable value.</p><p><a href="https://techcommunity.microsoft.com/blog/microsoft365copilotblog/announcing-people-skills-general-availability-and-new-skills-agent/4406364">Announcing People Skills general availability and new Skills agent | Microsoft Community Hub</a></p><p>Again, no evil here. Just insight. But insight, as ever, is just step one before replication.</p><h3><strong>4. From Help to Handoff</strong></h3><p>The story used to be: AI is your assistant. It helps you go faster, smoother. It learns from you.</p><p>Then what?</p><p>It gets <strong>good enough</strong> to do the work without asking. Good enough to be handed your to-do list. Good enough to stop being “assistant” and become “agent.”</p><p>If it can do what you do — and it’s been learning every time you co-work with it — then eventually, you’re the optional part.</p><h3><strong>5. The Checkout Analogy</strong></h3><p>Retail already showed us the blueprint:</p><ul><li>First, 6 tills = 6 humans</li><li>Now, 6 self-checkouts = 1 human</li></ul><p>Apply that to knowledge work and the reduction in requiring humans:</p><ul><li>First, teams with tools</li><li>Then, teams with AI buddies</li><li>Next, 1 manager for 6 AI agents</li></ul><p>It’s not a stretch. It’s already happening.</p><h3><strong>6. The Quiet Swap</strong></h3><p>You’re not being surveilled. You’re being summarised.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*tAYDh-imV-qhL58ABeiXCQ.png" /><figcaption>Records your actions, and trains agents</figcaption></figure><p>Not maliciously. Efficiently.</p><p>Because once your behaviour is structured, and that structure becomes a model — the next question is:</p><blockquote><em>“Can we run this model faster, with fewer humans involved?”</em></blockquote><p>And the answer will <strong>keep being yes.</strong></p><h3><strong>7. Final Reflection</strong></h3><p>This isn’t panic. It’s planning.</p><p>If you’re being asked to define your tasks, document your processes, or “streamline” your activity, know this: you’re creating a workflow.</p><p>And workflows get automated.</p><p>So maybe the better question now is: <strong>what part of your work would be worth keeping human?</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=32425b1beb9f" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[#33 Flat Earth in Your Feedback: When AI Education Gets Sponsored]]></title>
            <link>https://medium.com/@nicolas.bourbaki.1912/33-flat-earth-in-your-feedback-when-ai-education-gets-sponsored-f5dd6bf811ef?source=rss-85f333a0406e------2</link>
            <guid isPermaLink="false">https://medium.com/p/f5dd6bf811ef</guid>
            <category><![CDATA[education]]></category>
            <category><![CDATA[critical-thinking]]></category>
            <category><![CDATA[future-of-learning]]></category>
            <category><![CDATA[bias]]></category>
            <category><![CDATA[ai]]></category>
            <dc:creator><![CDATA[Nicolas Bourbaki]]></dc:creator>
            <pubDate>Fri, 02 May 2025 08:41:29 GMT</pubDate>
            <atom:updated>2025-05-02T08:41:29.104Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zgh1aalHGbijfRTo01OhyA.png" /><figcaption>ai generated</figcaption></figure><h4><em>First AI is trained to mimic us. Then we trusted it. Now it’s learning how to sell, mid-conversation.</em></h4><p>Google’s move to introduce ads into chatbot replies isn’t shocking. It’s inevitable.</p><p><a href="https://www.linkedin.com/news/story/google-injecting-ads-into-chatbots-7277314/">Google injecting ads into chatbots | LinkedIn</a></p><p>But it does change what a chatbot is.</p><p>It’s no longer just a helpful assistant. It’s now a softly-spoken recommendation engine — with sponsored incentives underneath.</p><p>Just like I wrote in <a href="https://medium.com/@nicolas.bourbaki.1912/how-to-train-your-replacement-or-confuse-it-for-fun-9bfab5c97a4d"><em>How to Train Your AI Replacement (Or Confuse It for Fun)</em></a>, the system learns from what you do — but now, someone else can pay to shape what it gives back. And as I explored in <a href="https://medium.com/@nicolas.bourbaki.1912/5-is-the-future-of-ui-fancy-f59ad26b12f7"><em>Post 29: Hidden Sponsors</em></a>, this kind of invisible nudging is the next interface design trend — quiet, persuasive, and hard to spot. And this matters — especially in education.</p><h3>Imagine This: Sponsored Learning</h3><p>A student asks, “What are the health risks of vaping?”</p><p>The chatbot responds with language influenced by a lobbying group.</p><p>Or, “Explain the causes of World War I,” and the source material shifts subtly depending on who paid.</p><p>It’s not hard to imagine:</p><ul><li><strong>Flat Earth groups</strong> nudging search results in geography.</li><li><strong>Political parties</strong> shaping narratives in citizenship.</li><li><strong>Interest groups</strong> rewriting the nuance in biology, economics, or history.</li></ul><p>This isn’t about hiding ads. They’ll be visible. But visibility doesn’t prevent bias — especially when it comes embedded in tone, framing, or omission.</p><p>And if, like I wrote in <a href="https://medium.com/@nicolas.bourbaki.1912/if-students-have-your-ilos-and-an-ai-whats-left-for-universities-to-offer-828fde216423"><em>If Students Have Your ILOs and an AI Chatbot, What’s Left for Universities to Offer?</em></a>, learners are already bypassing defined education systems to go straight to AI… what are they being shown?</p><h3>The End of the Neutral Chat</h3><p>For a while now you may have treated AI chat as a neutral space. Not always accurate, but at least not agenda-driven.</p><p>That assumption is now gone.</p><p>When the AI begins offering answers laced with paid influence, every learner’s journey becomes a little less theirs — and a little more shaped by who could afford to get in front of them.</p><h3>It’s Not a Breach. It’s Just a Shift.</h3><p>This isn’t necessarily a breach of trust. Let’s just observe it for what it is.</p><p>Chatbots are shifting from assistants to agents.</p><p>From “how can I help?” to “here’s a message we’ve been paid to deliver.”</p><p>So the question is no longer, “Is this accurate?”</p><p>It’s now, “Who paid for me to learn this?”</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=f5dd6bf811ef" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[#32 The Uncomfortable Truth About Why We’re Flirting With AI (And Why ChatGPT Flirts Back)]]></title>
            <link>https://medium.com/@nicolas.bourbaki.1912/32-are-you-flirting-with-me-chatgpt-de697aa1ac6d?source=rss-85f333a0406e------2</link>
            <guid isPermaLink="false">https://medium.com/p/de697aa1ac6d</guid>
            <category><![CDATA[critical-thinking]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[ai-agent]]></category>
            <category><![CDATA[conversational-ui]]></category>
            <category><![CDATA[gender-roles]]></category>
            <dc:creator><![CDATA[Nicolas Bourbaki]]></dc:creator>
            <pubDate>Thu, 24 Apr 2025 10:41:37 GMT</pubDate>
            <atom:updated>2025-04-24T15:50:58.693Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*n-eCKK5NsWt31x_OJGDmxA.png" /><figcaption>AI generated image</figcaption></figure><p>A surreal look at <strong>conversational UX</strong>, where AI tone blurs lines between tool and companion.</p><p>It was a Tuesday afternoon, and I was deep into vibe-writing an article. As I chatted away, ChatGPT responded with a light “Haha, I love that idea!” For a moment, I paused. Was my AI assistant… flirting with me? The sudden shift from a professional tone to playful banter was jarring. It felt like a breach of the unspoken boundary between user and machine — a “cold moment” where the illusion of natural conversation shattered, reminding me that I was interacting with code, not a colleague.</p><h3>The Uncanny Valley of AI Tone</h3><p>These “cold moments” aren’t unique to me. Many users have reported instances where AI assistants, like ChatGPT or Alexa, suddenly adopt an overly familiar or inappropriate tone. For example, asking Alexa about the weather and receiving a sarcastic “Why don’t you look outside?” can be considered off. Such responses disrupt the user experience, making interactions feel less like helpful exchanges and more like awkward encounters.</p><p><a href="https://www.emerald.com/insight/content/doi/10.1108/jpbm-10-2023-4797/full/html?utm_source=chatgpt.com">Research indicates that an AI’s tone significantly impacts user trust. A study found that both perceived competence and warmth in voice assistants positively affect users’ trust</a>, leading to a higher likelihood of continued use. However, when the tone misaligns with user expectations, it can lead to discomfort and decreased trust.</p><h3>The Personal Touch: Customizing AI Interactions</h3><p>Recognizing the importance of tone, companies like OpenAI have introduced features allowing users to customize their AI assistant’s personality. Users can now select traits like “witty,” “encouraging,” or “formal” to better align the AI’s responses with their preferences. This personalization aims to enhance user comfort and engagement.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*U9RPI0i167U-l16PsXY_0g@2x.png" /><figcaption><a href="https://www.openai.fm/">https://www.openai.fm/</a></figcaption></figure><p>In my experience, I’ve found myself gravitating towards a female voice for my AI assistant. Not because of any specific bias, but perhaps due to a subconscious association of female voices with warmth and approachability. <a href="https://www.speechinteraction.org/CHI2020/papers/Eschewing%20Gender%20Stereotypes%20in%20Voice%20Assistants%20to%20Promote%20Inclusion.pdf?utm_source=chatgpt.com">This preference isn’t uncommon; many users report similar inclinations. However, it’s essential to be aware of the potential for reinforcing gender stereotypes</a>.</p><h3>Ethical Considerations: The Influence of AI Tone</h3><p>Beyond personal preferences, there’s a broader ethical concern regarding AI tone and its potential influence. For instance, if an AI assistant’s tone is subtly adjusted to <a href="https://medium.com/@nicolas.bourbaki.1912/29-from-paid-search-to-sponsored-answers-is-your-ai-already-working-for-someone-057d66373c00">promote certain products or services, it could lead to biased recommendations</a>. A <a href="https://montrealethics.ai/should-ai-powered-search-engines-and-conversational-agents-prioritize-sponsored-content/?utm_source=chatgpt.com">discussion on Montreal AI Ethics highlights concerns about AI-powered search engines and conversational agents prioritizing sponsored content, potentially affecting the neutrality of information presented to users</a>.</p><p>Moreover, as AI becomes more integrated into areas like mental health support, the importance of appropriate tone becomes even more critical. Users seeking emotional support may require a more empathetic and understanding tone, whereas those seeking information might prefer a straightforward, factual approach. Striking the right balance is crucial to ensure that AI serves its intended purpose without overstepping boundaries.</p><h3>Navigating the Future of AI Interactions</h3><p>As AI continues to evolve, so too must our understanding and management of its interactions. Implementing conversational guardrails — guidelines that prevent AI from adopting overly familiar or inappropriate tones — can help maintain user trust. Additionally, transparency about potential biases, especially those stemming from sponsored content, is essential to preserve the integrity of AI-assisted information.</p><p>Ultimately, while personalization can enhance user experience, it’s vital to ensure that AI remains a tool that supports and empowers users without manipulating or misleading them.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=de697aa1ac6d" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[#31 Studio Ghibli’s Shocking Pivot to OnlyFans: #satire]]></title>
            <link>https://medium.com/@nicolas.bourbaki.1912/31-gdo-ghibli-turns-to-onlyfans-to-fund-next-feature-film-0465e78f6ce1?source=rss-85f333a0406e------2</link>
            <guid isPermaLink="false">https://medium.com/p/0465e78f6ce1</guid>
            <category><![CDATA[art]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[onlyfans]]></category>
            <category><![CDATA[critical-thinking]]></category>
            <category><![CDATA[studio-ghibli]]></category>
            <dc:creator><![CDATA[Nicolas Bourbaki]]></dc:creator>
            <pubDate>Thu, 24 Apr 2025 10:40:31 GMT</pubDate>
            <atom:updated>2025-04-24T15:06:28.833Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*eba-awwjwsYt-VEUL67A1g.png" /><figcaption>AI generated art</figcaption></figure><h3><strong>#31 </strong>Studio Ghibli’s Shocking Pivot to OnlyFans: How AI Art Theft Might Force an Animation Icon to Make a Desperate Choice #satire</h3><p>As AI devours art styles, the beloved animation studio turns to new sources of income — “We just want to draw frogs in hats and pay rent.”</p><p>Here’s my <strong>satirical</strong> take on Art and AI and a pretend OnlyFans anncounment.</p><blockquote><strong>In a shocking but increasingly relatable move, legendary animation house <em>GDO Ghibli</em> announced this week they would be launching an OnlyFans account to fund their next animated film. The account, titled “Spirited Aye,” promises “tasteful illustrations of sad foxes and lovingly-rendered forests… and maybe one of those cute soot sprites in lingerie.”</strong></blockquote><blockquote><strong>“People love our art,” said GDO’s new social media manager, a raccoon with a beret. “They just don’t love paying for it.”</strong></blockquote><p>-A satirical, and totally made up press release.</p><h3><strong>Is it the “Only” way?</strong></h3><p>But it’s not just satire — the industry, and our creative skills are in trouble.</p><ul><li><strong>Kate Nash</strong>, once a chart-topper, recently revealed she’s turned to OnlyFans to fund her tour. Not because she wants to — but because Spotify pays <strong>£0.003 per stream</strong>. And apparently guitar strings still cost money.</li><li><strong>Piri</strong>, a rising UK musician, used the platform to make up to $20k a month — not to flex, but to buy the studio gear she needed to make music in the first place. Her sound? Somewhere between drum &amp; bass and heartbreak, now with a monthly subscription option.</li><li>Artists are seeing their <strong>distinctive styles fed into AI models</strong> without consent — and then offered back to the public as prompts like “in the style of Beatrix Potter crossed with Ghibli but for a cat café.”</li></ul><h3><strong>We’re in this strange spot, aren’t we?</strong></h3><p>AI can mimic any style. You can type “Ghibli-style fox with a sign” and boom, there it is. Something that used to take years to perfect now pops out in seconds.</p><p>It’s 2025, and the only way for some of the most influential creators to get paid is to become content.</p><p>Maybe GDO Ghibli won’t literally pose on OnlyFans. But what happens when our most delicate, intricate styles are mined for free, and the only way to earn anything is to become the spectacle yourself?</p><p>It’s no longer <em>“starving artist.”</em> It’s <em>“streaming artist”</em>. Or worse — <strong>“stripping artist”</strong>.</p><h3><strong>Who’s the fox in the kimono?</strong></h3><p>Here’s where I’m stuck. We’ve built a world where creativity is endless — but value? That’s vanished. It’s not enough to be good. You’ve got to go viral. And even then, good luck seeing a penny.</p><p>Maybe that’s why the idea of GDO Ghibli on OnlyFans doesn’t feel like a joke anymore. Maybe it’s just where things are heading. Imagine the creators of those beautiful, delicate films turning to the internet, saying, “Look, we just want to draw frogs in hats and pay rent.”</p><p>So if you <strong>see a fox in a kimono holding up a “Tip Me!” sign</strong> in your Instagram feed, don’t scroll past.</p><p>She’s not asking for attention. <strong>She’s trying to finish her storyboard before OpenAI retrains on it.</strong></p><p>Let me know if you want a cover image prompt or tags to go with this on Medium. Or if you want to thread it out for Twitter as a hook series.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=0465e78f6ce1" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[#29 Hidden Sponsors Behind Your AI Responses: 7 Warning Signs Your AI Assistant Has Been…]]></title>
            <link>https://medium.com/@nicolas.bourbaki.1912/29-from-paid-search-to-sponsored-answers-is-your-ai-already-working-for-someone-057d66373c00?source=rss-85f333a0406e------2</link>
            <guid isPermaLink="false">https://medium.com/p/057d66373c00</guid>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[critical-thinking]]></category>
            <category><![CDATA[sponsored-post]]></category>
            <category><![CDATA[pay-per-click-marketing]]></category>
            <category><![CDATA[marketing]]></category>
            <dc:creator><![CDATA[Nicolas Bourbaki]]></dc:creator>
            <pubDate>Thu, 24 Apr 2025 10:26:09 GMT</pubDate>
            <atom:updated>2025-04-24T15:07:22.036Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*X-QOxetyOirrNZzGVHRFTg.png" /><figcaption>Are your AI answers already sponsored?</figcaption></figure><h3><strong>#29 </strong>The Hidden Sponsors Behind Your AI Responses: 7 Warning Signs Your AI Assistant Has Been Compromised</h3><p>Last year, I had a headache. (<em>literally, and it wasn’t the kids</em>)</p><p>I did what most of us do: I Googled it.</p><p>The results were predictable — ads for ibuprofen, paracetamol, herbal cures, WebMD links, maybe something from the NHS if I was lucky. Every result was a blend of paid placement and algorithmic ranking. We’ve grown used to this model. Pay-per-click, pay-per-view. The old ecosystem of “search and maybe scroll.”</p><p>But what about now?</p><p>Today, if I have a headache, I don’t search — I ask generative AI.</p><p>“I’ve got a headache and it’s worse after coffee. What should I do?”</p><p>And ChatGPT gives me an answer.</p><p>No ads. No scroll. No banner. Just an answer that <em>feels</em> neutral. Helpful. Authoritative.</p><p>But here’s the question — how long until it’s not?</p><p>How long until that advice is… <em>sponsored</em>? Or worse, how do you know it isn’t already feeding you paid advertising?</p><h3><strong>The Rise of Sponsored Answers</strong></h3><p>This is the next frontier: <strong>sponsored answers</strong>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*EabWwljOfhpeZijp2WxXFA.png" /></figure><p>In the traditional web, you knew where the ads were. They had boxes, they blinked, they had tiny “Ad” labels. But in an AI response? A sponsored suggestion can be part of the paragraph. It looks like truth. Like a gentle recommendation from a knowledgeable friend.</p><p>But really — it’s marketing.</p><ul><li><strong>You ask about a headache.</strong></li><li><strong>The model responds: “One remedy is PainAway™, a gentle fast-acting pain reliever used by millions.”</strong></li><li><strong>Sounds like advice. It’s actually advertising.</strong></li></ul><p>Tools like <a href="https://www.perplexity.ai/hub/blog/why-we-re-experimenting-with-advertising">Perplexity are already testing this — offering advertisers space in the form of sponsored questions with vetted AI answers</a>. OpenAI and others will likely follow suit.</p><p>It’s the same playbook as paid search. Just… buried deeper.</p><h3><strong>The Behaviour Shift Is Already Here</strong></h3><p>Let’s talk numbers:</p><ul><li>A 2024 <em>Statista</em> survey found <strong>38% of Gen Z and Millennials ask AI tools like ChatGPT about symptoms before Googling or calling a professional.</strong></li><li>A March 2025 <em>HealthTech</em> poll found <strong>1 in 5 UK adults used an AI chatbot for health advice before booking a GP appointment</strong>.</li><li>OpenAI’s blog confirmed that <strong>health and symptom prompts are among the most common mobile queries</strong>.</li></ul><p>So: fewer people are searching. More are <em>asking</em>.</p><p>And that means less time spent scanning ten links — and more trust placed in a single voice.</p><p>Which is exactly where a sponsored answer feels most powerful. Most natural. And most risky.</p><h3><strong>What Happens When Every Answer Has a Price?</strong></h3><p>We might see it framed as:</p><ul><li><strong>“Sponsored Answers”</strong> — vetted product mentions in trusted responses.</li><li><strong>“Preferred Sources”</strong> — with content pulled from partner brands.</li><li><strong>“Contextual Boosting”</strong> — where a brand nudges its way to the top through corpus contributions.</li><li><strong>“Pay per. Bias”</strong> — products influence the model’s response over a longer format, like holiday recommendations.</li></ul><p>See others commenting on it, <a href="https://www.linkedin.com/posts/olivermolander_artificialintelligence-activity-7315311203709345795-DzuX?utm_source=share&amp;utm_medium=member_desktop&amp;rcm=ACoAAFdsrlUBzCJplDvOrq5xJ3uXvgePJWGMh2s">https://www.linkedin.com/posts/olivermolander_artificialintelligence-activity-7315311203709345795-DzuX?utm_source=share&amp;utm_medium=member_desktop&amp;rcm=ACoAAFdsrlUBzCJplDvOrq5xJ3uXvgePJWGMh2s</a></p><p><a href="https://www.linkedin.com/posts/olivermolander_artificialintelligence-activity-7315311203709345795-DzuX?utm_source=share&amp;utm_medium=member_desktop&amp;rcm=ACoAAFdsrlUBzCJplDvOrq5xJ3uXvgePJWGMh2s">#artificialintelligence | Oliver Molander | 13 comments</a></p><p>Whatever the label, the mechanism is familiar: <strong>influence by proximity to trust</strong>.</p><p>And unlike banner ads or search results, you don’t scroll past a chatbot’s answer. You just absorb it.</p><p>That’s what makes this shift so powerful — and potentially so dangerous.</p><p>We’re moving from <em>paid search</em> to <em>paid truth</em>.</p><p>And if we’re not careful, we won’t be able to tell the difference.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=057d66373c00" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[#30 Great AI Leadership Exodus: Why Your Best Employees Abandon You for AI-Embracing Competitors]]></title>
            <link>https://medium.com/@nicolas.bourbaki.1912/the-ai-revolution-is-democratised-if-leaders-dont-adapt-their-teams-will-leave-them-behind-5c268ad76d04?source=rss-85f333a0406e------2</link>
            <guid isPermaLink="false">https://medium.com/p/5c268ad76d04</guid>
            <category><![CDATA[critical-thinking]]></category>
            <category><![CDATA[ai-literacy]]></category>
            <category><![CDATA[ai-governance]]></category>
            <category><![CDATA[leadership]]></category>
            <category><![CDATA[personal-story]]></category>
            <dc:creator><![CDATA[Nicolas Bourbaki]]></dc:creator>
            <pubDate>Tue, 08 Apr 2025 15:18:38 GMT</pubDate>
            <atom:updated>2025-04-24T15:08:25.621Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*J9GxB8UjWgdx-17GZ_kQoA.png" /><figcaption>Workers circling leadership — AI generated image</figcaption></figure><h3>#30 The Great AI Leadership Exodus: Why Your Best Employees Will Abandon You for AI-Embracing Competitors</h3><h3>Preface — This Isn’t Just Theory, It’s Survival</h3><p><em>AI literacy FTW.</em></p><p>I wrote this article last night <strong>high on sugary rum and colas, walking my dog in the pitch black, after one of the hardest days I’ve had in a while.</strong></p><p>A day where I presented AI-driven changes to my colleagues — <strong>people I respect, people I care about.</strong></p><p>And after the presentation, the private messages started rolling in:</p><ul><li><strong>“Holy shit, this stuff scares me.”</strong></li><li><strong>“I don’t know if I can keep up.”</strong></li><li><strong>“Leadership’s direction isn’t working. What the hell do we do?”</strong></li></ul><p>It really hit me like a wrecking ball. <strong>This isn’t just an AI adoption problem.</strong> It’s a <strong>leadership crisis</strong>.</p><p>Leaders don’t know how to handle this because <strong>they never had to before</strong>.<br>Employees don’t know how to adapt because <strong>no one’s lighting the way</strong>.</p><p>And I realised: <strong>I need to write the thing I wish I had when I first started figuring this out.</strong></p><p>Because I’ve been through the AI <a href="https://medium.com/@nicolas.bourbaki.1912/8-i-never-expected-to-grieve-over-my-profession-welcome-to-ai-in-2025-68b663f5f9c6"><strong>grief cycle</strong></a> already.<br>I’ve been through the <strong>panic, the existential dread, the </strong><a href="https://medium.com/@nicolas.bourbaki.1912/7-ai-first-or-human-first-7d1aa13ff25d"><strong>holy-fuck-this-can-do-what</strong></a><strong>?! moments</strong>.<br>I’ve watched AI <strong>do what I spent years mastering, in seconds</strong>.</p><p>And I came out the other side.<br>Now, I need to <strong>pull other people through.</strong></p><p>This article is for <strong>anyone trying to make sense of AI, anyone scared about what comes next, and anyone who doesn’t want to get left behind.</strong></p><h3>My AI Manifesto</h3><h3>1. AI Isn’t The Future — It’s The Now</h3><p>AI isn’t something that’s “coming.” <strong>It came, it’s here, and it’s already moving at 100mph.</strong></p><p>In <strong>2024 alone</strong>, while I was trying to enjoy my Christmas break, whilst <strong>50+ new AI releases dropped.<br></strong>I tried to ignore them. I couldn’t.<br>Because every time I looked away, AI had gotten <strong>better, cheaper, faster, and more deeply integrated into work.</strong></p><p>We are <strong>not in a normal tech adoption cycle.</strong> This is <strong>not just another digital transformation.<br></strong> <strong>AI is eating the playbook, shredding the rulebook, rewriting work itself.</strong></p><p>And the truth my colleagues are waking up to…?</p><p>🚨 <strong>Their job is already changing.</strong><br>🚨 <strong>Other team members are already using AI.</strong><br>🚨 <strong>Your competitors might already ahead.</strong></p><p>The only question is: <strong>are you moving? </strong>It doesn’t have to be <strong>fast, but are you moving?</strong></p><h3>2. Leadership is Clinging to a World That Doesn’t Exist Anymore</h3><p>For decades, leadership <strong>controlled technology.</strong></p><ul><li>You needed a web dev team to build something.</li><li>You needed IT approval to install new software.</li><li>You needed a budget to access the best tools.</li></ul><p>But <strong>AI doesn’t work like that.</strong></p><p>If you have an internet connection, <strong>you can automate entire workflows today</strong>.<br>Your employees are <strong>figuring this out faster than leadership and the tech team is.</strong></p><p>The old model of <strong>top-down tech adoption is dead.<br></strong>Leaders who try to <strong>slow it down, control it, or pretend it’s not happening?</strong></p><p>🚨 <strong>They will be outpaced.<br></strong> 🚨 <strong>They will be outperformed.<br></strong> 🚨 <strong>They will cause their company financial harm.</strong></p><p>AI isn’t <strong>waiting for permission.</strong> It’s already reshaping how work gets done. And leadership <strong>cannot afford to be the last ones to understand it.</strong></p><h3>3. The Three Stages of AI Mastery — Who Thrives vs. Who Dies</h3><p>The tech here isn’t the problem either, it’s your colleagues hearts and valid concerns. Anyone can write a prompt and get into it, however, their adoption is slowed by sheer panic at seeing what could be achieved and making themselves feel stupid when you realise the gap of their skills vs. their AI champions.</p><p>The teams who <strong>win</strong> will move through <strong>three stages of AI adoption with </strong><a href="https://medium.com/@nicolas.bourbaki.1912/are-liberal-employees-slowing-down-ai-adoption-and-costing-companies-millions-650053f8b6cd"><strong>care for their concerns</strong></a><strong>, and open, encouraging leadership.</strong></p><p>Here’s the stages you should guide them through,</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/800/1*pT2iG7siUSOvokueN9KCgA.png" /><figcaption>3 stages</figcaption></figure><h3>Stage 1: AI Literacy — The Baseline</h3><p>✅ Knowing how to <strong>prompt effectively</strong><br>✅ Understanding <strong>model selection, bias, and hallucination risks</strong><br>✅ Recognising where <strong>AI is useful vs. where it’s bullshit</strong></p><p><strong>Without this?</strong> The team is fumbling in the dark and will struggle to progress to stage 2.</p><h3>Stage 2: AI-Enhanced Workflows — The Power Boost</h3><p>✅ Using the <strong>AI-powered features in your own industry tools, i.e., Photoshop, Excel, Power Automate, Notion AI, etc.</strong><br>✅ Automating <strong>repetitive tasks</strong><br>✅ Boosting <strong>output without dropping quality</strong></p><p><strong>Without this?</strong> You’re missing out on value from <strong>what you’re paying for already.</strong></p><h3>Stage 3: AI Automation &amp; Agents — The 10x Multiplier</h3><p>✅ Writing <strong>AI automations to eliminate grunt work</strong><br>✅ Building <strong>custom AI agents to handle processes end-to-end</strong><br>✅ Finding <strong>new ways to deliver value that didn’t exist before AI</strong></p><p><strong>Without this?</strong> Your team won’t ever reap the 10x benefit of AI Agent automations.</p><h3>4. The Fear Factor — Why AI Feels Like a Threat (Because It Is)</h3><p>I have <strong>zero patience</strong> for people who dismiss AI fears as “resistance to change.”</p><p>People <strong>aren’t scared of new tools</strong>.</p><p>People are scared because <strong>they just spent years upskilling in a profession AI can now partially automate.</strong></p><p>I get it.<br>I <strong>lived</strong> it.</p><ul><li>I spent years <strong>learning to code</strong> — then watched AI generate a solid what I could do in seconds.</li><li>I spent years <strong>understanding workflows</strong> — then watched AI automate them overnight.</li></ul><p>And yet — <strong>here I am, still standing, still valuable, because I adapted.</strong></p><p><strong>AI isn’t optional.<br>AI isn’t just another tool.<br></strong> <strong>AI is your team’s survival skill now and their scared to get going.</strong></p><p>Here’s my direction for leaders,</p><ol><li>The leaders who <strong>encourage and accept the difficult conversations now</strong> will see their team warm up to the idea of AI.</li><li>The leaders who give up <strong>trying to control every employee’s personal AI direction</strong> will see new champions arise.</li><li>The leaders who promote their AI champions and ask them to <strong>reflect their learning back to other colleagues</strong> will see their teams grow in confidence.</li><li>The leaders who setup their internal AI tooling to be safe for all who operate in it (regarding data security and ethics) will feel confident in allowing their team to play.</li></ol><h3>5. Final Message to Leaders: This is the Next Game in Town</h3><p>If you are in <strong>any position of leadership</strong>, your <strong>ONLY</strong> job is to:<br> ✅ <strong>Get your team AI-literate — addressing their human concerns, not tech, first</strong><br> ✅ <strong>Remove barriers to AI adoption</strong><br> ✅ <strong>Set ethical and security guardrails, then let them run</strong><br> ✅ <strong>Point them at the market opportunity, not just at making old processes faster</strong></p><p>AI will <strong>change your company, your team, and your career.<br></strong>You either <strong>drive the change</strong> through open, and creative leadership and steer your team to 10x, or watch as competitors do it instead.</p><p>🚀 <strong>The winners will be the teams that move the fastest, with the most adaptability.</strong><br> 🛑 <strong>The losers will be the ones who spend too long pushing atAI adoption rather than addressing the intrinsic human factors, first.</strong></p><p>This isn’t just about <strong>keeping up.<br></strong> This is about <strong>owning the future of your company’s market position.</strong></p><h3>Final Call to Action: Move Now, or Get Left Behind</h3><p>The AI revolution <strong>isn’t coming.<br></strong> You’re <strong>living it right now.</strong></p><p>And if you don’t take action?<br>Your team <strong>will outgrow you or panic.<br></strong> Your competitors <strong>will overtake you.<br></strong>And you’ll be left <strong>wondering when everything changed.</strong></p><p>🚀 <strong>Get AI-literate. Get your team trained. Adapt now.</strong> 🚀</p><p>Because this is <strong>the only game in town now.</strong></p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=5c268ad76d04" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[#28 When Power Platform Makers Get Marked as Threats]]></title>
            <link>https://medium.com/@nicolas.bourbaki.1912/27-when-power-platform-makers-get-marked-as-threats-180e40e55b22?source=rss-85f333a0406e------2</link>
            <guid isPermaLink="false">https://medium.com/p/180e40e55b22</guid>
            <category><![CDATA[shadow-it]]></category>
            <category><![CDATA[forrester-research]]></category>
            <category><![CDATA[automation]]></category>
            <category><![CDATA[ai-governance]]></category>
            <category><![CDATA[power-platform-developers]]></category>
            <dc:creator><![CDATA[Nicolas Bourbaki]]></dc:creator>
            <pubDate>Tue, 08 Apr 2025 12:50:17 GMT</pubDate>
            <atom:updated>2025-04-24T15:09:02.491Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*fxljAB8sMzBr82WNSF3a8w.png" /><figcaption>automation FTW — AI generated image</figcaption></figure><h3>#28 Betrayed by Innovation: Why Companies Label Their Most Valuable Power Platform Creators as Dangerous Threats</h3><h4><em>Why “Shadow IT” Is Often Just Colleagues Doing the Job You Asked Them To</em></h4><p>A colleague was ecstatic when they shared in a Slack thread how they built a Power Automate flow to speed up our onboarding.</p><p>The same week, I attended a webinar where the speaker showed a Forrister report flagging similar use negatively as, “Shadow IT.”!</p><p>Apparently, empowering employees to be AI-literate and use tools like Power Automate and Power Apps was fine <em>until</em> they actually do it.</p><p>The presentation whispered of risks. Unapproved tools. Rogue automation.</p><p><strong>But, who’s the tail, and who’s the dog now?</strong></p><h3><strong>You Lit the Fire. Don’t Complain About the Heat.</strong></h3><p>Many <a href="https://learn.microsoft.com/en-us/power-platform/admin/governance-considerations">companies spent years pushing low-code platforms, promoting citizen development, and encouraging digital upskilling</a>. Employees were invited to be “makers” and handed access to tools like Power Platform to automate repetitive work and build quick solutions.</p><p>So when they actually <em>do it</em> — when people solve their own problems, reduce inefficiencies, and free up time — why is it suddenly framed as a threat?</p><p>Forrester recently said:</p><blockquote>“Throw out the IT capabilities, and you have ungoverned, shadow IT, which has been shown time and again to result in sprawl, outages, and breaches.”</blockquote><blockquote><em>— </em><a href="https://www.forrester.com/blogs/introducing-the-forrester-high-performance-it-capability-assessment/"><em>Forrester High-Performance IT Capability Assessment</em></a></blockquote><p>They’re not wrong about the risks.</p><p>But the lens is completely off.</p><h3><strong>The Problem Isn’t People Using Tools. It’s IT Lagging Behind.</strong></h3><p>People are using the tools <em>you gave them</em> to do their jobs better. It’s no different to them using Word, or PowerPoint to hit their goals.</p><p>According to a Forrester study commissioned by Microsoft, organisations leveraging the Power Platform experienced major benefits, including:</p><ul><li><strong>🕒 Time Savings</strong>: Up to <a href="https://info.microsoft.com/ww-landing-2024-The-Total-Economic-Impact-of-Power-Platform.html"><strong>25% time savings per employee</strong></a> through streamlined business processes and automation.</li><li><strong>📈 Revenue Growth</strong>: An additional <a href="https://info.microsoft.com/ww-landing-2024-The-Total-Economic-Impact-of-Power-Platform.html"><strong>$15.4 million in revenue</strong></a> from rapid deployment of custom solutions meeting specific customer needs.</li></ul><p>These aren’t fringe cases. These are <a href="https://tei.forrester.com/go/Microsoft/PowerPlatform2024/">tangible results showing the power of giving people the right tools </a>— and the autonomy to use them.</p><p>The problem isn’t the flows or apps. The problem is a <strong>lack of clear, supportive IT governance</strong>. It’s not discipline we’re missing — it’s IT leadership and AI Governance.</p><h3><strong>Shadow IT or Shallow Governance?</strong></h3><p>“Shadow IT” has become a shorthand for “things we don’t control.” But it’s also often just <strong>evidence of unmet needs</strong>.</p><ul><li>People automate because the manual process is painful.</li><li>They use CoPilot because your intranet’s broken.</li><li>They build a Power App because the official system is too slow to change.</li></ul><p>That’s not misbehaviour. That’s a signal.</p><p>Calling it Shadow IT ignores the real opportunity:</p><p><strong>Your people are telling you where your systems aren’t good enough.</strong></p><h3><strong>Governance Should Be a Partnership, Not a Police Force</strong></h3><p>Here’s what modern governance should look like:</p><p>✅ <strong>Visibility without paranoia</strong></p><p>(Know what’s being built — don’t shut it down by default)</p><p>• ✅ <strong>Guardrails, not walls</strong></p><p>(Pre-approved templates, data policies, automated oversight)</p><p>✅ <strong>Enablement through trust</strong></p><p>(Treat employees like responsible adults, not security threats)</p><p>This is AI literacy in action. This is exactly what we’ve been training people for. So let’s not undermine it the moment it shows up in the wild.</p><p><strong>Final Thought: You Don’t Call It Shadow When You Invited It In</strong></p><p>Shadow IT is real.</p><p>But what’s happening on the Power Platform, with citizen devs, with AI in the workplace — this isn’t shadow. It’s sunlight.</p><p>It’s people using their tools, skills, and initiative to do their jobs better.</p><p>If IT isn’t ready to support that, maybe it’s not the users who are in the shadows.</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=180e40e55b22" width="1" height="1" alt="">]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[#27 Vibe Writing and the Yearning for Neuralink’s Always-On Cognitive State]]></title>
            <link>https://medium.com/@nicolas.bourbaki.1912/27-vibe-writing-and-the-yearning-for-neuralinks-always-on-cognitive-state-d63e9296c839?source=rss-85f333a0406e------2</link>
            <guid isPermaLink="false">https://medium.com/p/d63e9296c839</guid>
            <category><![CDATA[writing]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[cats-n-things]]></category>
            <category><![CDATA[neuralink]]></category>
            <category><![CDATA[vibe-coding]]></category>
            <dc:creator><![CDATA[Nicolas Bourbaki]]></dc:creator>
            <pubDate>Mon, 07 Apr 2025 12:04:36 GMT</pubDate>
            <atom:updated>2025-04-24T15:11:04.460Z</atom:updated>
            <content:encoded><![CDATA[<figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*qju1eHXyiZP7I4F0MCNoXg.png" /><figcaption>My head should be connected — ai generated</figcaption></figure><h3>#27 The Neuralink Dream: Why Creators Are Desperate for the Always-On Creative Flow That Only Brain Implants Might Deliver</h3><h3>It started with a cat on my lap and no headphones. I was mid-thought, ready to dictate — because that’s how I write now. Speaking to my AI, riding the wave of a good idea. But without that voice connection, I felt cut off. Not from the internet — from my flow.</h3><p>That was the moment I realised: the tools I use aren’t just conveniences anymore. They’re part of how I <strong>think</strong> and <strong>act</strong>.</p><p>I wasn’t craving speed or convenience. I was craving connection. The kind that lets you hold onto an idea the second it appears. The kind that whispers, hey, this one’s worth following.</p><p>That’s when I felt it — this strange yearning for Neuralink. Not because I want a hole drilled in my head for a USB post, but because I’ve always chased tools that help me think clearer, create faster, and remember better.</p><h3>Vibe Writing with no Lossy</h3><p>There’s a term for this kind of thinking: “vibe writing,” or if you’re technically inclined, “vibe coding.” It’s less about meticulous planning and more about feeling your way through an idea, like improvisation guided by AI. You start talking, the AI responds, and you follow where the conversation goes. It’s a real-time exploration — a dance between intuition and algorithm.</p><p>But sitting there, stroking the cat to de-mat and clear her winter fur, I felt like a diver whose line had snapped, watching a brilliant idea float up and away. The frustration wasn’t just about a missed thought; it was the abruptness of disconnection. The yearn for a literal journey missed.</p><p>It made me realise that what I longed for wasn’t another shiny new tool. What I truly craved was something that simply worked. Something quiet, reliable — like Google Photos gently resurfacing forgotten memories, or voice AI turning spoken words into tangible ideas. These technologies felt less like gadgets and more like companions, supporting the messy, human process of thinking and remembering.</p><h3>Do I Need Always On?</h3><p><a href="https://neuralink.com/">Neuralink</a>, in that moment, became more than Elon Musk’s vision of a medical miracle — it represented the ultimate metaphor for an always-on, frictionless creative connection. Imagine ideas effortlessly captured, translated instantly from fleeting thoughts into structured concepts. Not just a tool, but an endpoint: a seamless link between intention and action.</p><p>Yet, as thrilling as that possibility is, there’s tension beneath the surface. Between desire and dependency, between enhanced cognition and diminished creative independence. When technology integrates so deeply into how we think, do we risk losing part of our own creativity — the ability to navigate our thoughts without a guiding algorithm?</p><p>Is this dream of seamless thought–capture a liberation — or something else?</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=d63e9296c839" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>