From RL to IRL
Conference Context
- Date/time: 2026-06-30 · 1:30pm-1:50pm
- Track/room: Computer Use · Track 7
- Speaker(s): Gaurav Mishra
- Session type/status: session · confirmed
- Track: Computer Use
- Room: Track 7
- Session type: session
- Status: confirmed
Session Description
Today's agents have to operate in a messy reality of flaky connections, ephemeral credentials, and irreversible actions. They need to navigate real software the way humans do: recovering from failures, learning from feedback, and making sound judgment calls. This talk is about the fundamental changes in RL required to make agents ready for IRL. We'll walk through what it takes for training environments to reflect the complexity of the real world, the perception primitives that let an agent see what a user sees, the harness pieces that help it survive contact with real applications, and the failure modes you only discover when you stop scoring and start shipping.
Media Evidence
No related AI Engineer channel video found yet.
Evidence Graph
This evidence graph is generated from currently linked source material: official schedule text, related video pages, cached transcripts, visible slide text, dense/reconstructed slide pages, and AI slide-classification audits.
Media Signals
No linked video, transcript, or slide source has been attached yet.
Agent Reading Notes
Use these signals to refine the synopsis, topic links, people/company context, and method notes. If a source is a related external video rather than an exact official recording, keep it framed as supporting evidence.
Transcript Status
No official session recording transcript was found by exact title match on the AI Engineer YouTube channel during this run.
People
Notes
- Pending transcript synthesis when an official recording or confirmed matching video is available.
Synthesis
Synthesized Breakdown
From RL to IRL ## Conference Context - Date/time: 2026-06-30 · 1:30pm-1:50pm - Track/room: Computer Use · Track 7 - Speaker(s): Gaurav Mishra - Session type/status: session · confirmed - Track: Computer Use - Room: Track 7 - Session type: session - Status: confirmed ## Session Description Today's agents have to operate in a messy reality of flaky connections, ephemeral credentials, and irreversible actions. They need to navigate real software the way humans do: recovering from failures, learning from feedback, and making sound judgment calls. This talk is about the fundamental changes in RL required to make agents ready for IRL. We'll walk through what it takes for training environments to reflect the complexity of the real world, the perception primitives that let an agent see what a user sees, the harness pieces that help it survive contact with real applications, and the failure modes you only discover when you stop scoring and start shipping.
Speaker And Company Context
- Gaurav Mishra — role not listed at Amazon AGI Lab.
Topics Covered
Derived Links And Source Material
Novel Concepts / Clever Methods
- No highlighted novel concept has been detected yet.
Evidence Boundary
This synthesis is based on the official schedule and linked source pages. It should be revisited when exact session recordings or transcript-backed secondary sources are available.