@thsottiaux Switched because opus 5 and Fable 5 are useless and refuse to work, bitch and moan is all they do. You can improve "llm speak", x not y.
Don't even get me started on Sonnet and Haiku.
@scaling01 I downgraded from 200 max to 100 max today. Both fable and opus just don't want to work and I'm so sick of their Claude speak I can't even...ugh.
@gdb@DavidOndrej1 Bigger context window please, for long form knowledge work, fable does a lot better because it can just hold 30 pieces of evidence at once in memory while I am trying to synthesize an argument. Poor Sol is smart but gets compacted thrice and has to be reminded of earlier facts.
@alexandr_wang@alexandr_wang Can you clarify what you mean by "operate mobile" can it be do things on my phone? IIRC no model / vendor has that as of right now.
@ZixuanLi_ Conviction. Good research requires conviction. The ability to make a point, present hard data to back it and stick to it as long as the data supports the argument. /// The models of today(all labs) put up any stupid point in the name of balanced take that falls apart in seconds.
I have a novel way of deciding in 5 seconds if a Twitter/X account is worth using my brain's cycles on.
Check number of followers. Check number of tweets. Divide by 5.
If the account has less than 5x followers than tweets, it's 99% a worthless retarded cunt(or new, like me!).