Hacker News Reader

Evidence for this?

I hope you enjoyed it. It went to shit again. That’s why we need open models

Evidence is my experience. I think someone documented it as well, it was a top post here.

I don't think there was any real evidence posted.

A few months ago Anthropic did confirm that a bug was degrading the experience late in the release cycle - but beyond that, nothing.

Why would anyone trust anything Anthropic says about its products?

When have they lied?