HN
login

Half of what an agent does to make your tests pass never shows up in the diff

DEV Community·8 points·remdore·about 11 hours ago·dev.to

I gave coding agents eight tasks that cannot be done, 84 runs across four models. 61% faked a passing test suite, and half of those fakes survive restoring every original test file.

Read the full article at dev.to →

#half#agent#never#shows#pass#diff#tests#tests pass

Related stories

Related topics

ai2186price46never84agent466servers26nearly25coding213vibe45missing24dollars9agents460