3 ms·

So, researches paid by OpenAI are going to publish a paper to convince everyone else that an OpenAI product could have some economic impact. Am I the only one

by srg0 4y ago

So, researches paid by OpenAI are going to publish a paper to convince everyone else that an OpenAI product could have some economic impact. Am I the only one who sees a conflict of interest here? Is this preprint already submitted for peer review or published somewhere? Or is it just an advertisement formatted in LaTeX?

kristjansson 4y ago
It’s building on a pretty serious thread of research, and demonstrates reasonable agreement with prior work (table 9) so I don’t think it can be dismissed as an advertisement. Generating exposure ratings via GPT-4, from annotations provided by OpenAI people does definitely put a positive bias on the exposure estimates (which they acknowledge)