Weak models write tool calls as prose. We started executing them | Hacker News - But Better

Weak models write tool calls as prose. We started executing them fiit.ai

An agent reported success after making zero tool calls. Why the obvious guard missed it, and what we do now.

1 points•airbai•8 days ago•0 comments•

#tool #tool calls #prose #executing #weak #write #models write #weak models

0 comments

No comments yet.

Read the full thread on Hacker News →

Related stories

Show HN: Agent Chaperone – Screen AI agent tool calls and results with Jevgithub.com
Hacker News · 5 points · 9 days ago
We Tested Jev on 100 Agent Tool Callsarchestra.ai
Hacker News · 11 points · 9 days ago
Claude Isn't Allowed to Write Me Proseblog.kvit.app
Hacker News · 1 points · 2 days ago
Claude Isn't Allowed to Write Me Proseblog.kvit.app
Hacker News · 2 points · 4 days ago
Claude Isn't Allowed to Write Me Proseblog.kvit.app
Hacker News · 1 points · 3 days ago
Claude Isn't Allowed to Write Me Proseblog.kvit.app
Hacker News · 1 points · 1 day ago