Agents differ from models via the machinery: tools, memory, loops
vishalmisra · x · 2026-09-27
In a follow-up to his MIT talk, Vishal Misra argues that agents are different mostly because of the machinery around the model: tools, memory, permissions, feedback, and repeated calls.
The model itself is still doing next-token prediction; it's the surrounding system that turns those predictions into actions, state, and loops.
Related event: MIT Talk Explains LLMs from First Principles, No Transformer Needed(2 posts)→
More from coding & agent
- Devs Are Just Yes Machines Now: Complaining About Claude Extension's Constant Prompts in VS Code — ThePeterMick · 2026-09-27
- Routing Evals Are Missing, Says Elvis Saravia as OpenRouter Criticism Rages Without Evidence — omarsar0 · 2026-09-27
- Opus 5.5 prompt turns code-rendered frames into an epic 'History of Chinese Civilization' film — dotey · 2026-09-27
- Meta's Muse gives every user their own VM, called the most careful consumer agent sandbox yet — AccBalanced · 2026-09-27
- SafeScript: a Turing-incomplete JS subset lets agent policies replace code review — uriwa · 2026-09-27
- Open source AI-SQL engine Quail signals shift toward cheap task-specific models — sh_reya · 2026-09-27