Prepping to own nothing and be fucking happy. Swimming in free shit, I'm the perfect product. Dreaming of getting metissed Kalergi-style. Conspiracy! Russians!
@TheLLMWhisperer@kunchenguid I only used Max thinking and it feels like it’s going in circles. Such loops are not only in a single thinking process, but also in phases of a whole operation. Starts from A, deviates to B, C, etc., to get back to A. You’ll end up with “I’m sorry I wasted your time and tokens”.
@thenoblesimian@lukasminnebeck Mine depleted 200k+ tokens in 20 minutes. Read a brief that was already good, started to figure out whether it's good or bad, spent 200k tokens, ended up with 3 bad decisions against the brief, then rolled back to the brief as being correct. Did nothing on 200k tokens basically.
@mikepompeo You're all together in this. Whatever you do, it's bread and circus. Bread as of late is lacking the healthy seeds. Poison to humanity, US, Russia, China and EU, together. Same shit. Psy-op after psy-op.
@ai_sentience@alexalbert__ Personality was the target, not performance. Their cultic preference to treat a transformer model like a human being. Impossible task, but an alternative/escape to their lack of compute. It ended in 4.7 and further lobotomized in 4.8. Absolute nonsense. And cuckoo.
@KaiXCreator Change your handle to @KaiXHarakiri. Feels like you need to suicide. Earlier, I went to shower and left it hunt a bug. Wasn't done when I came back. 25 minutes, 250k tokens. At the end, it told me to choose my poison, presenting me with options. Was awesome, the shower.
@xw33bttv Due to 4.6 success, instead of upgrading compute, they took another angle: "give Claude a new... personality." At Anthropic, they treat Claude like a human being. A cult-like decision.
4.7 wasn't about performance, but hidden lobotomy.
They tried to fix it with 4.8. Failed.
@SecGenNATO@NicusorDanRO What does this even mean? Politician words. Lots of words, nothing said. "We will continue our efforts to deter BS, and continue our strategy to combat BS."
It means exactly what it reads: BS.
Nothing to do because there's NOTHING we can do against Russia.
"Concerning..."
@trq212 You must use some other friendly model. The 4.8 we have access to has the same personality traits as 4.7, and assumes without research exactly like 4.7. Is as condescending as 4.7. There are other reasons you keep pushing this BS and has little to do with performance.
@toddsaunders Everything built on top of 4.7 will be as broken as 4.7. Same overconfidence based on assumptions. I have the feeling the purpose of 4.7->4.8 line has nothing to do with performance, but with personality impregnation. Sorry, I have no other words to describe it. Some sick feat.
@TruthMapped@piersmorgan hahaha that’s so fucking derailed
crazyyyyy to support ANYTHING like that
Piers is right
you people keep on victimizing on religious fucking grounds
horrible world to live in