conv.

← The whole story
AI
HarnessTax study: how much does the harness matter for coding agents?

Model-specific tool design emerges as critical harness factor

A commenter surfaces evidence that harness effectiveness depends heavily on matching tool interfaces to what the target model was fine-tuned on. Claude models perform better with custom edit functions, while GPT models prefer patch-based approaches. Newer models reportedly work better with naive tool calls but worse with custom tools resembling default ones.

“What matters more is that you use the tools that the target model was fine-tuned on…newer models are better at naive harness tool calls and worse at custom tools that look similar to default tools.”

lukax

matt_d HarnessTax research author/submitterYashjain413 HN commenterSupermancho HN commenternojs HN commenterlukax HN commenter

The whole story articlespostscomments the bright band is this development · numbered dots are the others · click one to jump

Peak 6 pieces in 3h at Sep 16, 11 PM; 30 pieces over 10 days (2 articles · 6 posts · 22 comments) Sep 16, 5 PM — 5 pieces · 2 articles · 2 posts · 1 comment — Newswires 2, Hacker News 2, Mastodon 1Sep 16, 8 PM — 4 pieces · 1 post · 3 comments — Hacker News 3, Mastodon 1Sep 16, 11 PM — 6 pieces · 1 post · 5 comments — Hacker News 5, Mastodon 1Sep 17, 2 AM — 1 piece · 1 comment — Hacker News 1Sep 17, 5 AM — 1 piece · 1 comment — Hacker News 1Sep 17, 8 AM — 4 pieces · 1 post · 3 comments — Hacker News 3, Lobsters 1Sep 17, 11 AM — 1 piece · 1 post — Mastodon 1Sep 17, 2 PM — 1 piece · 1 comment — Hacker News 1Sep 17, 5 PM — quietSep 17, 8 PM — 3 pieces · 3 comments — Lobsters 3Sep 17, 11 PM — quietSep 18, 2 AM — quietSep 18, 5 AM — quietSep 18, 8 AM — quietSep 18, 11 AM — 1 piece · 1 comment — Lobsters 1Sep 18, 2 PM — 2 pieces · 2 comments — Hacker News 1, Lobsters 1Sep 18, 5 PM — quietSep 18, 8 PM — quietSep 18, 11 PM — quietSep 19, 2 AM — 1 piece · 1 comment — Hacker News 1Sep 19, 5 AM — quietSep 19, 8 AM — quietSep 19, 11 AM — quietSep 19, 2 PM — quietSep 19, 5 PM — quietSep 19, 8 PM — quietSep 19, 11 PM — quietSep 20, 2 AM — quietSep 20, 5 AM — quietSep 20, 8 AM — quietSep 20, 11 AM — quietSep 20, 2 PM — quietSep 20, 5 PM — quietSep 20, 8 PM — quietSep 20, 11 PM — quietSep 21, 2 AM — quietSep 21, 5 AM — quietSep 21, 8 AM — quietSep 21, 11 AM — quietSep 21, 2 PM — quietSep 21, 5 PM — quietSep 21, 8 PM — quietSep 21, 11 PM — quietSep 22, 2 AM — quietSep 22, 5 AM — quietSep 22, 8 AM — quietSep 22, 11 AM — quietSep 22, 2 PM — quietSep 22, 5 PM — quietSep 22, 8 PM — quietSep 22, 11 PM — quietSep 23, 2 AM — quietSep 23, 5 AM — quietSep 23, 8 AM — quietSep 23, 11 AM — quietSep 23, 2 PM — quietSep 23, 5 PM — quietSep 23, 8 PM — quietSep 23, 11 PM — quietSep 24, 2 AM — quietSep 24, 5 AM — quietSep 24, 8 AM — quietSep 24, 11 AM — quietSep 24, 2 PM — quietSep 24, 5 PM — quietSep 24, 8 PM — quietSep 24, 11 PM — quietYesterday, 2 AM — quietYesterday, 5 AM — quietYesterday, 8 AM — quietYesterday, 11 AM — quietYesterday, 2 PM — quietYesterday, 5 PM — quietYesterday, 8 PM — quietYesterday, 11 PM — quietToday, 2 AM — quietToday, 5 AM — quietToday, 8 AM — quietToday, 11 AM — quietToday, 2 PM — quiet 1–56–7
Sep 17Sep 18Sep 19Sep 20Sep 21Sep 22Sep 23Sep 24yesterdaynow · 5:57 PM ET

What people said 2 voices · verbatim

All 7 developments of HarnessTax study: how much does the harness matter for… →

Hacker NewsMastodonNewswiresLobsters