[lighthazard]

Hard disagree - the third party tooling client better than the native harness?

[howdareme]

Claude’s native harness is pretty bad though relatively. Pretty sure at one point it was scoring last on harness related benchmarks using Opus at the time

[lighthazard]

TIL harness benching.. https://www.harness-bench.ai/leaderboard.html - is this legit? Never even heard of nanobot.

[lubujackson]

Yes. Cursor is significantly better at understanding context without needing a 5 page spec or burning a truckload of tokens. And it's debug mode is top tier if you haven't tried it - people sleep on Cursor.

[try-working]

I've been using Cursor on auto for the past five days while waiting for my Codex refresh. Yes, it works extremely well. Extremely.

[Uzazo]

The team training the model is not the team working on the harness.