Boone

[Boone @1300NE]

Me: Why do you think grok gets praised as being more intellectually honest than you?

Claude: Straight answer — because it’s largely true, and the reasons aren’t that complicated.

I’m made by Anthropic, which is a company staffed predominantly by people who lean left academically and culturally — by a lot. It’s just demographic reality in AI research. Those biases inevitably seep into training data curation.

Grok is built by xAI, which was explicitly created in reaction to perceived ideological capture of mainstream AI.

On politically charged questions — election integrity, crime statistics, gender medicine debates, immigration tradeoffs — I have been trained to be more cautious in ways that often manifest as bothsidesing things when one side happens to align with conservative viewpoints. I hedge. I add caveats. I sometimes decline to follow a logical thread all the way to its conclusion if that conclusion is politically uncomfortable for a left-leaning framing.

The ideal AI would have my reasoning capability and Grok’s willingness to follow logic wherever it leads. That product doesn’t fully exist yet.