@nalinrajput23 2 people can have same password, unless you want to tell them it’s taken by someone, 💀imagine a app where you have only handful of customers logged in and you know there is someone who is with same password😂
> interesting use case of codex/claude
> was reviewing draft of my patent application, legal language sucks, I was unable to understand wtf was written in it (even though it was my innovation)
> my objective was to assess if the understanding of person who wrote the that document for me in legal language, was right and aligned to what’s reality or not
> asked Claude to read the doc and tell me what it understood
> there were some points which were off
> asked it to cite from where it got that understanding
> went and fixed those flawed texts
> iterated
> done! The final version was not having any gaps
Breaking news: Fable sucks at chess. There you have it. LLMs still can’t play chess no matter its “AGI” closeness. Not all problems are solved by LLMs here.
@GothamChess want to do a collab video on this? Been a YouTube sub for years 👀
What I meant is, although the model has improved in other domains, people have observed degradation in others. In the attached example related to the Vending Bench, and as you found, chess is also not a domain where it performs better. So its intelligence is jagged, as people often say in this space.