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

Vanishing Gradients

Hugo Bowne-Anderson

The Rise of Agentic Search

DEC 19, 202551 MIN
Vanishing Gradients

The Rise of Agentic Search

DEC 19, 202551 MIN

Description

<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>