Cognition Lab Talk: RL and Inference Optimization Are Converging
AAAzzam · x · 2026-07-30
Silas Alberti, Head of Research at Cognition, discussed the intersection of reinforcement learning (RL) and large-scale inference.
Key points from the interview include:
- Convergence of RL and Inference: RL training and underlying inference optimization are gradually merging, moving towards "Auto inference."
- Inference Procurement: Weighing the latency/throughput Pareto frontier when buying compute, and distinguishing between memory-bound and compute-bound scenarios.
- Advanced Optimization: Explored tree-based speculation and online training of the speculator.
- Agent Competition: Agents will increasingly compete on solving optimization problems.
More from coding & agent
- From Tech Lead to Engineering Manager: Shifting Gears with Coding Agents — dotey · 2026-07-30
- Neocarta: Build a Semantic Layer Graph for Text-to-SQL Agents to Fix Table Understanding Failures Upstream — JeremyCMorgan · 2026-07-30
- gNMIBuddy MCP Server: Enables LLMs to Retrieve Structured Network Data via gNMI — modelcontextprotocol · 2026-07-30
- ShippingRates MCP Server: Ocean Container Shipping Intelligence for AI Agents via Apify — modelcontextprotocol · 2026-07-30
- YC-Backed Speko Launches as an OpenRouter for Voice Models — ycombinator · 2026-07-30
- Creating Map Animations with Codex and Remotion: An AI-Assisted Workflow — Vjeux · 2026-07-30