Even if that's true it's kind of interesting because the models were asked to give their most controversial opinion. A controversial opinion is supposed to be something that most people don't agree with. Taken literally you cannot have an opinion that's both the average of what most people think and also controversial.
So either that's not what's happening here, or what we're getting is "what do most people think most people don't think about humanity".
If the latter, then it casts new light on the debates over whether LLMs can replace all white collar work. Because some of the most valuable human work is breaking false consensus. Steve Jobs, early Google, for that matter early OpenAI all succeeded because they bet hard on ideas others were rejecting. Is that a durable source of human advantage?
So either that's not what's happening here, or what we're getting is "what do most people think most people don't think about humanity".
If the latter, then it casts new light on the debates over whether LLMs can replace all white collar work. Because some of the most valuable human work is breaking false consensus. Steve Jobs, early Google, for that matter early OpenAI all succeeded because they bet hard on ideas others were rejecting. Is that a durable source of human advantage?