Re:Does the success of LLM support Wittgenstein's position t...
by karaterobot on Friday August 02, 2024 @02:22PM
(#41141237)
Explain the distinction and relevance for me please!
[ Parent
| on HN ]
Re:Does the success of LLM support Wittgenstein's position t...
by og_kalu on Friday August 02, 2024 @02:53PM
(#41141520)
It can't return a 10x10 grid but it has nothing to do with whether GPT can count or not. Similarly, it took 200+ tries to return a hotdog as you say, and again it has nothing to do GPT or LLMs.
You could just replace GPT with a Human and nothing about your experiment would change.
>My (trollish) position in that thread became that there is no better authority on common use than ChatGPT, whose entire purpose is to distill it and recapitulate it.
You're not querying ChatGPT here. You're querying Dalle-3.
Re:Does the success of LLM support Wittgenstein's position t...
by karaterobot on Friday August 02, 2024 @05:35PM
(#41142887)
Thanks for the reply!