Hacker News Reader

Yo, just because you can't tell when Claude is wrong, doesn't mean it's right.

I do agree that the Q1 2026 models in general have passed a threshold, but goodness almighty Opus 4.6 still screws up a lot.

It’s wrong. It made large mistakes on my code literally yesterday.

Wrong context

No. Aside from just making an algorithm that didn’t even run, it refused to use an MCP that it had registered in the same context session.

Ah, the eternal handwave for anything the AI doesn't do well - it must be user error.

Can I get your magical version of Opus that works? 4.6 has been a side grade at best, and worse than prior models most days.