[Zigurd]

That sounds plausible. But it doesn't explain why LLM's make laughably bad errors that even a biased and haphazard human researcher wouldn't make.

[Zigurd]

Gemini seems to have a user interface that, for the way most people encounter Gemini, is more closely linked to search results. This leads me to suspect that Google's approach to training could be uniquely informed by both current and historic web crawling.

[chpatrick]

I think that's been a lot less true over the last year or so. Gemini 2.5 Pro is the first LLM I actually find pretty damn reliable.