Building Search for AI Agents with Exa CEO Will Bryk (50 min)
ai-driven-innovation-economy
ai-in-workforce-disruption
ai-social-media-dynamics
- Release date: 2026-06-04
- Listen on Spotify: Open episode
- Episode description:
Sarah Wang speaks with Exa cofounder and CEO Will Bryk about building search infrastructure for the AI era. The conversation covers Exa’s origins, why traditional search engines were not designed for AI agents, and how search changes when the user is no longer a human but an autonomous system. They discuss retrieval, agent workflows, coding agents, data access, and why search may become a foundational layer for the emerging agent economy. Along the way, Bryk shares his views on AI-native products, the future of information discovery, and why some of the most important problems in technology can ultimately be framed as search problems. Resources: Find Will on X: https://x.com/WilliamBryk Find Sarah on X: https://x.com/sarahdingwang Check out everything a16z is doing with artificial intelligence here, including articles, projects, and more podcasts. Please note that the content here is for informational purposes only; should NOT be taken as legal, business, tax, or investment advice or be used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund. a16z and its affiliates may maintain investments in the companies discussed. For more details please see a16z.com/disclosures. Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.
Summary
- 🔍 Agent-First Search: Exa was built from the ground up for AI agents that demand exhaustive, complex, and controllable retrieval rather than the human-optimized experience of traditional search engines.
- 🚀 LLM-Era Advantage: Transformers and the bitter lesson enabled a small team to outperform Google on deep queries by focusing on neural retrieval instead of decades of human click data.
- 🌍 Search as Solution: Many societal problems (polarization, loneliness, recruiting) are fundamentally search problems that perfect retrieval could help solve.
- 💰 Token Efficiency: High-quality retrieval lets smaller models perform like larger ones, dramatically cutting token costs and addressing the emerging ‘tokenpocalypse’.
- 📈 Massive Future TAM: Agentic search volume is projected to dwarf human search, potentially exceeding Google Ads scale by the early 2030s as agents perform millions of retrievals daily.
Insights
- How might search engines designed specifically for AI agents fundamentally reshape what counts as ‘comprehensive’ information access compared to human-centric tools like Google?
- Time: 0:30 – 1:14
- Answer: The transcript highlights that agents require exhaustive results (thousands of documents), complex semantic+keyword queries, and extreme customizability rather than the 10 blue links optimized for human clicks. This shift enables deeper knowledge work like competitor analysis or historical research that traditional search fails at.
- What new architectural and data decisions are required when building search infrastructure that must serve millions of agent queries instead of a few human searches per day?
- Time: 9:12 – 12:17
- Answer: Exa was built from scratch around agent needs (complex queries, comprehensiveness, customizability, variable latency) rather than human click data, leading to fundamentally different engineering choices and the prediction that agentic search will exceed Google-scale economics by the early 2030s.
- Why might retrieval quality become a bigger differentiator and business opportunity than raw model intelligence as LLMs continue to improve?
- In what ways could perfect retrieval systems address societal issues like political polarization and loneliness by giving people accurate, controllable, and comprehensive information?