Web Discovery Summit

Or Lenchner, CEO of Bright Data, analyzes what the DOJ vs. Google case reveals about building competitive search: the infrastructure decisions, the real-time data problem, and operating at Google scale.

Paddy Srinivasan, CEO of DigitalOcean, explores why AI trained on yesterday's internet is already obsolete: how data freshness impacts accuracy, the cost of stale information and the infrastructure needed for real-time knowledge.

In-depth discussion with Parag Agrawal (Parallel), Dhruv Batra (Yutori), and Sarah Sachs (Notion) on how agents consume the web differently than humans: attention spans, feedback signals, economic incentives and the transition from discovery to action.

Stuart Reid, CEO of Nosible, demonstrates how AI monitors financial markets 24/7 to detect early warning signs years before collapse: the graph-based architecture that makes continuous monitoring tractable and reduces search costs by 50-100x.

Will Bryk, CEO of Exa, on why search hasn't evolved since 1998: what perfect search means, treating the web as a filterable database, serving searches from 30ms to 10 minutes, and how bad algorithms create misinformation just like bad content.

Philippe Mizrahi, CEO of Linkup, on why agents can't use Google: they need information chunks, not web pages. By indexing extracted content and eliminating duplication, Linkup reduces index size 5-10x while delivering precise, context-ready data agents can use directly.

Panel with Mike Marino (DuckDuckGo), Rob Chisholm (Qatalyst Partners), and John Gunja (Amplemarket) on AI search economics: free vs paid models, M&A at record multiples, hallucination risks, and why search costs vary from pennies to $20 per query. Moderated by Angela Hoover (Andi).

Jacob Ervin from Mozilla on browsers evolving from passive viewports to AI decision-makers: determining when to search vs research, inverting economics as headless browsers pay for results, and decoupling search volume from user actions as agents trigger exponential queries per session.

Anthony Goldbloom, CEO of Sumble (former CEO of Kaggle), on using LLMs to structure web data into knowledge graphs: extracting intelligence from job postings, reducing cleaning costs from thousands of people to 17, and building technographic data for go-to-market teams. Interviewed by Brooke Larson (Rime Labs).

Richard Socher, CEO of you.com, reflects on building AI-powered search from research to production: what works at scale, where the remaining challenges are and what the next generation of search technology must solve.