Should agents internalize the loop, or keep enforcement outside?
DanielKhashabi · x · 2026-07-21
A thread argues that the next step after reasoning models and learned tool use is to internalize the agent loop itself.
The proposed direction is for models to own planning, tool use, error recovery, verification, memory, and stopping decisions—possibly even tool design and skill composition. But the author draws a hard boundary: the model should handle cognition, while the harness still provides sandboxing, permissions, budgets, and auditing as an enforcement layer that does not trust the policy.
The core question is where that boundary should sit, and how to distinguish true internalization from superficial tool-form behavior.
More from coding & agent
- DeepMind paper: LLMs can't sample randomly on their own, only convert given seeds — rohanpaul_ai · 2026-10-27
- DeepMind paper: LLMs can't sample randomly on their own — pass random numbers from code instead — rohanpaul_ai · 2026-10-27
- BELL: Open-Source Stack Combines Browserbase, Eve, iMessage and Stripe Link for Personal Agents — jeff_weinstein · 2026-10-27
- Legora runs 2,000 parallel agent evals on Modal Sandboxes for legal AI — blelbach · 2026-10-27
- Two engineers rebuilt a FedEx supplier's billing core in 3.5 months vs an estimated 7-8 — alex_verem · 2026-10-27
- GPT-6.1 Terra scores 48 on AA Coding Agent Index at one-third the cost of GPT-5.6 Terra — ArtificialAnlys · 2026-10-27