VessOnSecurity on Nostr: I'd gladly use a Chinese model, if it were any good. Unfortunately, they just ain't. ...

I'd gladly use a Chinese model, if it were any good. Unfortunately, they just ain't. Nothing can compare with Claude. Nothing. Kimi is usable (unlike things like ChatGPT, Gemini, or Grok, which are totally useless crap, at least for programming tasks) - but there is a huge difference between "usable" and "as good as Claude". And, yes, I've tried GLM 5.2 too - and it, too, failed to impress me.