We didn't make the models smarter. We built the thing that catches them confidently wrong — and it caught us too.

DEV Community·3 points·bryanw·10 days ago·dev.to

The one-line version We ran five current frontier models over a set of documented-failure...

Read the full article at dev.to →

#didnt #wrong #smarter #confidently #catches #caught #models smarter #didnt make

Related stories

My prompt-injection fix caught 0 of 20 attacks. The part I almost didn't build caught all of them.dev.to
DEV Community · 3 points · 4 days ago
Show HN: Lathoa, a math app for kids where the AI is wrong on purposelathoa.ai
Hacker News · 40 points · about 14 hours ago
Show HN: A Shapefile converter that fixes the winding you didn't know was wrong9revolution9.com
Hacker News · 1 points · 8 days ago
Where Jev worked for us, and where it didn'tgitloom.ai
Hacker News · 2 points · 9 days ago
Dyson Won't Say What's Wrong with Its CameraJet Toothbrushwired.com
Hacker News · 2 points · 8 days ago
Belay – See what keeps going wrong in your Claude Code and Codex sessionsgetbelay.vercel.app
Hacker News · 1 points · 9 days ago

Related topics

ai2044 write73 intentional4 just144 become35 writing69 didnt write3 likely10 hash12 fix37