The agent bug I thought was the model turned out to be the harness (www.reddit.com)
model roundup
GPT 5
-
Spent 3 days debugging an agent that kept looping on the same web search tool call. First things that came to mind was the model couldn't handle the schema.
-
During this discovery utilizing only free AI services I have managed to undeniably prove both conjectures. This would absolutely not have been possible without using GPT5+ as the critic for my work.
-
Why do GPT 5 models not allow temperature change on API? (www.reddit.com)
I think this is so boring of them. Messing with temperature is one of the easiest ways to get more creative outputs and OpenAI doesn't let people edit their newer models' temperature at all.
-
Professor’s bold prediction: AI could help cure all diseases within a decade (excitech.media via reddit)
In the article, the professor Derya Unutmaz specifically mentions an experience with an OpenAI model (GPT-5) where it explained a mechanism from an experiment that he and his colleagues couldn't figure out. What would have taken human rese…
-
Show HN: Which public repos are friendliest to an AI coding agent? (www.agentfriendlycode.com via hn)
Public leaderboard ranking GitHub, GitLab, and Bitbucket repos by how agent-friendly they are for Claude Code, Cursor, Devin, GPT-5 Codex, Gemini CLI, Aider, OpenHands, and Pi — per model, with AGENTS.md / CLAUDE.md, CI, tests, and dev-env…
-
The downfall of OpenAI and who will follow (msukhareva.substack.com via hn)
The Downfall of OpenAI And Who Will Follow Sora dead. GPT-5 flopped.
-
Show HN: Single bash command to find the best matching HN jobs (news.ycombinator.com)
Today I learned that I can find the most interesting jobs for myself in the "Who's Hiring" thread with a single command: curl https://news.ycombinator.com/item?id=47975571 | \ uvx html2text | \ llm --model gpt-5-nano "These are Hacker News…