Dhruv Batra says web agents and robot navigation are the same problem in cheaper form
ziv_ravid · x · 2026-07-21
In a new episode of The Information Bottleneck, Dhruv Batra—co-founder of Yutori and former head of Embodied AI at Meta FAIR—argues that training robots to navigate 3D scans and building web agents are fundamentally similar problems.
Key themes:
- the browser is a cheaper place for agents to fail than the physical world;
- reinforcement learning on live websites is becoming a practical topic;
- as agents browse the web, the unresolved question becomes who should pay for that usage.
The episode is available on the show's website, YouTube, and Spotify.
Related event: Ex-Meta Embodied AI Lead: Training Web Agents is Like Training Robots(3 posts)→
More from coding & agent
- Kolab Benchmark Uses Pro Screen Recordings to Show AI Agents Overestimate Collaboration Quality — minnesotanlp · 2026-10-27
- Dev argues Claude burns far more tokens than GPT models for the same coding work — haider1 · 2026-10-27
- Stop building agents that can't change models: stable contracts, external state, real evals — rseroter · 2026-10-27
- Cheap models can cost more in agent harnesses than frontier ones, researcher warns — vasuman · 2026-10-27
- Kanzo: A 5k-Line Rust Kanban Binary With MCP for Claude Code and Codex — allmudi · 2026-10-27
- 'A human approves everything' is a trap: approval fatigue and how to actually design safe agents — Efistoffeles · 2026-10-27