"Nerf tracker" that shows when AI models quietly get worse
Nnews.ycombinator.com
Runs the same coding and reasoning tasks against each model every day and publishes when quality drops after the launch hype fades.
Read on news.ycombinator.com ↗Summary written by IdeaBank.
More in Developer Tools
- 1Password for API keys, aimed at vibe codersGREG ISENBERG
- Seven-day iPhone Duo compatibility fix for existing appsideabrowser.com
- GitHub for normies: home for apps built by non-developersGREG ISENBERG
- Turn GitHub repos into polished explainer landing pagesIdea Browser
- Git Trophy: 3D-printed trophy of your GitHub contribution chartIdea Browser