Thread 106278937

And if he calls himself the smartest AI influencer, why isn't he able to figure out and admit that OpenAI used him to rug pull a bunch of people into paying for the new watered down ChatGPT which turned out to be dumbed down on purpose by OpenAI? Is he a shill or did he really get tricked? And why can't he figure out what went wrong if he's as smart as he claims to be when it comes to AI?

All pro-AI "people" are retarded

>>106278937 (OP)
Who?

>>106278993
He's like the Lex Friedman, but of AI. The Joe Rogan of AI of sorts. But now I can't tell if he's a shill and pretending to be dumb about being used or if he's in on it.

>>106279015
>the who of whos
That doesn't help me much but thanks for trying.

>>106279028
Pretending you don't know who Joe Rogan is won't make you sound cool and mysterious anon

>>106279028
He's like the Linus Tech Tips or the Gamers Nexus, but of AI.

This board was spammed with people claiming GPT-5 is "AGI", meanwhile it repeatedly failed to fix a basic ass mistake it made. I gave the problem to DeepSeek and it one shotted it. Then I gave GPT-5 all of DeepSeek's "thinking" to analyze how it fucked up so bad. This shit is undercooked.

>>106278993
some youtuber who keeps starting threads about himself
he got caught jerking off in a chuck e cheese a few months ago

>>106279720
>he got caught jerking off in a chuck e cheese a few months ago
Source?

>>106279666
Damn, did Deepseak reach AGI already?

>>106278993
some gay webshitter youtuber

>>106280219
Nta but I find deepseek to be more "clever" than a lot of models that it supposedly loses to in benchmarks. Claude in particular gets autistically fixated on stuff and often gets stuck as a result, and I'll switch the agent over to deepseek who usually finds an elegant and simpler solution.

I've asked deepseek why this is the case and it claims that its creators trained it to be more exploratory and less rigid. Which might also explain why it also feels more "schizo" than other models, since its fine going off into weird territory.

All of the models have their own unique feel and different strengths, I've actually found it beneficial to use a lot of different ones.

>>106278937 (OP)
imo he fucked up, after some initial stumbles gpt5 is way smarter than 4o was. he should have stuck to his guns, because even if he was a paid shill, he was correct. now it looks like he is both a paid shill, and he folded due to mob pressure, which makes his opinion basically worthless.

i watched his videos briefly because the ai normie i was subscribed to had one too many thumbnails with colored irises and was looking for alternatives.