Re:Ask HN: Are you tired of reading ChatGPT headlines?
by silversmith on Saturday February 04, 2023 @05:13PM
(#34658955)
The thing is, it's always confident. It gave you the highest confidence answer, out of many high confidence answers. Low confidence would imply the model has not seen the particular words before. A Markov chain does not suffer from imposter syndrome.
[ Parent
| on HN ]
Re:Ask HN: Are you tired of reading ChatGPT headlines?
by mech422 on Saturday February 04, 2023 @05:32PM
(#34659128)
Heh - I say the thing is to get the general public to realize that, and not assume "computers are always right" ? (eg, I don't think the problem is with chatGPT being 'wrong' - I think the problems start when people assume its right...)