Practitioner: new models still make serious mistakes on moderately hard applied ML work

ivan_bezdomny · x · 2026-09-07

The author notes that every time he worries a new model might write good code, docs, and plans without user input, digging in reveals it's making really serious critical mistakes — at least for moderately challenging applied ML work. His conclusion: human practitioners are still necessary.

Original post →

More from Models

Models channel →