@sama All LLMs seem to suffer difficulty with uncertainty. I am studying for a radiology exam and constantly get very confident wrong answers from chat gpt. "I am not sure" needs to be a much more frequent response.
@EvanRobertsWFAN Jeff was clearly not the issue with last year's team and he would be more productive than Robert Jr., the corpse of Semien, and Polonco combined.
@LeadingReport In the fullness of time this will likely be true but this is absolutely not the case at all right now. There are a bunch of commercial models that preform well for a relatively small number of indications but AI can't replace a single radiologists today