Vanishing Gradients

<p><em>We’re really moving from a world where humans are </em><strong><em>authoring search queries</em></strong><em> and humans are </em><strong><em>executing those queries</em></strong><em> and humans are </em><strong><em>digesting the results</em></strong><em> to </em><strong><em>a world where AI is doing that for us</em></strong>.</p><p><strong>Jeff Huber</strong>, <strong>CEO</strong> and <strong>co-founder</strong> of <strong>Chroma</strong>, joins Hugo to talk about how <strong>agentic search</strong> and <strong>retrieval are</strong> changing the very nature of search and software for builders and users alike.</p><p><strong>We Discuss:</strong></p><p>* <strong>“Context engineering”</strong>, the strategic design and engineering of what context gets fed to the LLM (data, tools, memory, and more), which is now essential for building reliable, agentic AI systems;</p><p>* Why simply stuffing large context windows is no longer feasible due to <strong>“context rot”</strong> as AI applications become more goal-oriented and capable of multi-step tasks</p><p>* A framework for precisely curating and providing only the most <strong>relevant</strong>, h<strong>igh-precision information</strong> to ensure <strong>accurate</strong> and <strong>dependable AI systems;</strong></p><p>* The <strong>“agent harness”</strong>, the collection of tools and capabilities an agent can access, and how to construct these advanced systems;</p><p>* <strong>Emerging best practices</strong> for builders, including hybrid search as a robust default, creating “golden datasets” for evaluation, and leveraging sub-agents to break down complex tasks</p><p>* The major unsolved challenge of <strong>agent evaluation</strong>, emphasizing a shift towards iterative, data-centric approaches.</p><p>You can also find the full episode on <a target="_blank" href="https://open.spotify.com/show/3yuz89gqAhcMcdy3SZPe4X?si=AKl2jvIARD2Liw1bBH2Nng&#38;nd=1&#38;dlsi=e91c7411f20b48fb">Spotify</a>, <a target="_blank" href="https://podcasts.apple.com/us/podcast/vanishing-gradients/id1610318868">Apple Podcasts</a>, and <a target="_blank" href="https://youtu.be/l9V1IN8mQIg">YouTube</a>.</p><p><a target="_blank" href="https://notebooklm.google.com/notebook/32983a67-1602-4e3f-be20-2ca2863978f7">You can also interact directly with the transcript here in NotebookLM</a>: If you do so, let us know anything you find in the comments!</p><p>👉 <strong><em>Want to learn more about Building AI-Powered Software? Check out our </em></strong><a target="_blank" href="https://maven.com/hugo-stefan/building-ai-apps-ds-and-swe-from-first-principles?promoCode=ss-rav"><strong><em>Building AI Applications course</em></strong></a>. It’s a live cohort with hands on exercises and office hours. <strong>Our final cohort is in Q1, 2206</strong>. Here is a <a target="_blank" href="https://maven.com/hugo-stefan/building-ai-apps-ds-and-swe-from-first-principles?promoCode=vgch">35% discount code</a> for readers. 👈</p><p><em>Oh!</em> One more thing: we’ve just announced a Vanishing Gradients livestream for January 21 that you may dig:</p><p>* <a target="_blank" href="https://luma.com/6os4cy5g"><strong><em>A Builder’s Guide to Agentic Search & Retrieval with Doug Turnbull and John Berryman</em></strong></a> (register to join live or get the recording afterwards.</p><p>Show notes</p><p>* <a target="_blank" href="https://x.com/jeffreyhuber">Jeff Huber on Twitter</a></p><p>* <a target="_blank" href="https://www.linkedin.com/in/jeffchuber/">Jeff Huber on LinkedIn</a></p><p>* <a target="_blank" href="https://www.trychroma.com/">Try Chroma</a>!</p><p>* <a target="_blank" href="https://research.trychroma.com/context-rot">Context Rot: How Increasing Input Tokens Impacts LLM Performance</a> by The Chroma Team</p><p>* <a target="_blank" href="https://hugobowne.substack.com/p/ai-agent-harness-3-principles-for">AI Agent Harness, 3 Principles for Context Engineering, and the Bitter Lesson Revisited</a></p><p>* <a target="_blank" href="https://high-signal.delphina.ai/episode/context-engineering-to-ai-agent-harnesses-the-new-software-discipline">From Context Engineering to AI Agent Harnesses: The New Software Discipline</a></p><p>* <a target="_blank" href="https://research.trychroma.com/generative-benchmarking">Generative Benchmarking</a> by The Chroma Team</p><p>* <a target="_blank" href="https://www.anthropic.com/engineering/effective-context-engineering-for-ai-agents">Effective context engineering for AI agents</a> by The Anthropic Team</p><p>* <a target="_blank" href="https://hugobowne.substack.com/p/making-sense-of-millions-of-conversations">Making Sense of Millions of Conversations for AI Agents</a> by Ivan Leo (Manus) and Hugo</p><p>* <a target="_blank" href="https://www.anthropic.com/engineering/multi-agent-research-system">How we built our multi-agent research system</a> by The Anthropic Team</p><p>* <a target="_blank" href="https://lu.ma/calendar/cal-8ImWFDQ3IEIxNWk">Upcoming Events on Luma</a></p><p>* <a target="_blank" href="https://youtu.be/l9V1IN8mQIg">Watch the podcast video on YouTube</a></p><p></p><p>👉 <strong><em>Want to learn more about Building AI-Powered Software? Check out our </em></strong><a target="_blank" href="https://maven.com/hugo-stefan/building-ai-apps-ds-and-swe-from-first-principles?promoCode=ss-rav"><strong><em>Building AI Applications course</em></strong></a>. It’s a live cohort with hands on exercises and office hours. <strong>Our final cohort is in Q1, 2206</strong>. Here is a <a target="_blank" href="https://maven.com/hugo-stefan/building-ai-apps-ds-and-swe-from-first-principles?promoCode=vgch">35% discount code</a> for readers. 👈</p><p><a target="_blank" href="https://maven.com/hugo-stefan/building-ai-apps-ds-and-swe-from-first-principles?promoCode=vgch">https://maven.com/hugo-stefan/building-ai-apps-ds-and-swe-from-first-principles?promoCode=vgch</a></p><p></p><p></p><p></p><p></p><p></p><p></p><p></p><p></p> <br/><br/>This is a public episode. If you would like to discuss this with other subscribers or get access to bonus episodes, visit <a href="https://hugobowne.substack.com?utm_medium=podcast&#38;utm_campaign=CTA_1">hugobowne.substack.com</a>

We have been sold a story of complexity. Michael Kennedy (Talk Python) argues we can escape this by relentlessly focusing on the problem at hand, reducing costs by orders of magnitude in software, data, and AI. In this episode, Michael joins Hugo to dig into the practical side of running Python systems at scale. They connect these ideas to the data science workflow, exploring which software engineering practices allow AI teams to ship faster and with more confidence. They also detail how to deploy systems without unnecessary complexity and how Agentic AI is fundamentally reshaping development workflows.

Gemini 3 is a few days old and the massive leap in performance and model reasoning has big implications for builders: as models begin to self-heal, builders are literally tearing out the functionality their built just months ago: ripping out the defensive coding and reshipping their agent harnesses entirely. Ravin Kumar (Google DeepMind) joins Hugo to breaks down exactly why the rapid evolution of models like Gemini 3 is changing how we build software. They detail the shift from simple tool calling to building reliable "Agent Harnesses", explore the architectural tradeoffs between deterministic workflows and high-agency systems, the nuance of preventing context rot in massive windows, and why proper evaluation infrastructure is the only way to manage the chaos of autonomous loops.

Many leaders are trapped between chasing ambitious, ill-defined AI projects and the paralysis of not knowing where to start. Dr. Randall Olson argues that the real opportunity isn't in moonshots, but in the "trillions of dollars of business value" available right now. As co-founder of Wyrd Studios, he bridges the gap between data science, AI engineering, and executive strategy to deliver a practical framework for execution.

Most AI teams find their multi-agent systems devolving into chaos, but ML Engineer Alex Strick van Linschoten argues they are ignoring the production reality. In this episode, he draws on insights from the LLM Ops Database (750+ real-world deployments then; now nearly 1,000!) to systematically measure and engineer constraint, turning unreliable prototypes into robust, enterprise-ready AI.