@ynmncha No wonder huge models augmented with huge test time compute can get math olympiad gold and what not.
As ridiculous as it sounds, It will be interesting to see what gpt/claude/gemini 10 with 10GW test time compute can achieve... Maybe one day.
I bought this $ASTER as soon as @cz_binance tweeted and now its worth $3,000,000. Pretty insane.
Fck it, give away. 1 anon who like and retweet will get 15,000 $ASTER (yes, for real) in 24hrs.
@AJs_AI Brother... The first time you get the benefit of the doubt, but twice over two days with 0 response... If only hope was not so appealing, I'd think this is engagement bait.
Manus just blessed me AGAIN
They saw how quickly we ran through the first 100 community codes… so they gave me 500 more….
If you missed out on the first round, here’s your chance:
1. Retweet this post
2. Follow
3. DM me “MANUS”
If you got the access code in the first round, reply with what you cooked up so far!
Manus is absolutely next level. I literally maxed out the tokens yesterday.
@CodeWriter23@OpenAI Looking back our exchange was either prescient... or it was that obvious. All that dust raised at the time over a simple post training advancement really put OpenAI in a tight spot. They look clueless with their recent releases.
@indevvx @abacaj Except when building anything useful the cost of output tokens matters. Try using 4.5 for agentic use case. GL with $150/m-otokens.
Getting a base model to "think" with post training is trivial. What's even the point of gpt4.5 if its aimed at vibes but the public can't use it?
@max_paperclips @TheXeophon @tsarnick Him and Hasabis in one team would likely shit on the competition. That would be a dream scenario for the future of AI.
@CurtTigges@blader Yes but it depends how. The crux would be if it can solve problems by deduction, something that so far has alluded AI. The difference would be in the manner it solves it, whether by recall, having been trained on the problem, or by generalization / reasoning.
@CodeWriter23@OpenAI I did not mean it literally. But he is one of their biggest assets as far as fulfilling their AGI mission goes. Replacing someone like him is almost impossible.
@CodeWriter23@OpenAI They can't fire him. But if he is not reinstated in the board, I do not see him sticking for long. I can see Musk and Larry Page giving him the helm at their AI divisions.
@lukaszkaiser@RichardHead5420 And do you think your friend did what he did out of stupidity? Or is there info asymmetry.
I doubt he stays after this, now that he is out of the board and Sam MS have complete control.
I am willing to bet that where he moves, AGI discovery chances move.
With Ilya Sutskever out of OpenAI board, and assuming he does not get reinstated. That is the biggest loss to OpenAI out of this. I do not see how he continues there given his personal mission if not onboarded. He is the central piece to AGI. Google / Elon might make a move?
@OntologicalMax@davidchalmers42 Might wanna have a look at the concept of emergence. Start with the complexity/dynamics difference of an ant to an ant colony, and extrapolate that to neurons. Should be right up your alley, the philosophical implications are vast, from AI to free will in deterministic systems.