{"type":"rich","version":"1.0","provider_name":"Transistor","provider_url":"https://transistor.fm","author_name":"focal podcast","title":"Why Horizontal Beats Vertical in AI Agents | The Compounding Error Problem Most Founders Miss | The Case For Research-Heavy Teams Win | How to Build AI That Actually Generalizes with Abhishek Das, Co-Founder & Co-CEO of Yutori","html":"<iframe width=\"100%\" height=\"180\" frameborder=\"no\" scrolling=\"no\" seamless src=\"https://share.transistor.fm/e/afeb1c21\"></iframe>","width":"100%","height":180,"duration":3408,"description":"The Horizontal vs Vertical AI Debate: Why This Ex-Meta AI Researcher Is Betting Big on Horizontal Web Agents\nShould you build narrow (vertical) or go broad (horizontal) in AI? This episode unpacks why one PhD researcher abandoned his working vertical product to chase a much riskier horizontal bet - and why VCs leaning heavily into vertical AI might be missing something.\nAbhishek Das is the co-founder and co-CEO of Yutori, which has raised over $15 million from Radical Ventures, Felicis, and prominent angels including Ali Gil, Sarah Guo, Scott Belsky, and Guillermo Rauch. Previously a research scientist at Meta's FAIR lab, Abhishek holds a PhD from Georgia Tech where he pioneered work on AI agents that can see, talk, and act starting in 2016.\nIn Today's Episode We Discuss:\n00:53 - Why how we interact with the web hasn't changed in three decades and what will break that\n02:27 - The coming shift from manual browsing to AI assistants performing tasks in the background\n05:57 - What \"agents\" actually meant in ML research before the term became overloaded\n06:14 - Why 90% accuracy per step creates catastrophic failure rates over multi-step workflows\n08:46 - The behavior pattern humans nail intuitively that machines struggle with: backtracking from errors\n10:11 - The DoorDash experiment: building an end-to-end food ordering agent that never shipped\n12:58 - Why training on sinle websites leads to memorization instead of generalization\n13:03 - The dopamine problem: some tasks users don't want automated\n15:08 - Why capability-scoped beats website-scoped: the pivot to read-only horizontal agents\n18:05 - Three criteria that drove the horizontal decision: research, user value, and data strategy\n24:18 - Scouts API launch: why different channels have different risk appetites for web agents\n26:30 - Flying close to the sun: how Yutori competes with hyperscalers on horizontal AI\n30:32 - What VCs should actually test for in horizontal AI teams beyond founder horsepower\n32:10 - Why...","thumbnail_url":"https://img.transistorcdn.com/my7KgVcEQnLNwqL4F-8_lpRk6UA6mDb7bNK4EyzKfe4/rs:fill:0:0:1/w:400/h:400/q:60/mb:500000/aHR0cHM6Ly9pbWct/dXBsb2FkLXByb2R1/Y3Rpb24udHJhbnNp/c3Rvci5mbS8zOTYw/YzMxMzc0MjQ2MTM2/MjE4ZDFhMDIzNDUx/NWE2ZS5wbmc.webp","thumbnail_width":300,"thumbnail_height":300}