are ppl really using claude or gemini over gpt 4? 

if so, have you tried the native iOS / Mac gpt apps?

can you share an example of gpt 4 clearly outmatched by a competitor?

design @splits | /latenightcrew /event-pass /feud

I'm currently Claude first, and it meets my needs. I found a number of examples where it would make better code changes than GPT-4, but I don't have them organized as proof rn, sorry.

co founder of @metamask . Making computers behave for us.

I think I was pushing its limits a bit: I would provide maybe a specification doc, maybe an example of interfacing with another API, and maybe a style guide. GPT would drop one of those components in its approach. Claude is better at holding all the relevant context into the final implementation, it seemed like.