Greg Brockman says solving a Millennium Prize problem could be just 2-5 years away we’re at AGI level-3; level-4 innovators will be different. Imagine a whole data center focused on curing cancer or solving a Millennium problem: "that level of computational power is something the world has never seen"

I really like the idea of AGI levels, it makes sense with how the goal posts have moved over the years. Brilliant actually, wish I would have thought of it myself.

I had @chatgpt update the AGI goal posts image and added the Level.

>>Millennium Prize huh?

I bet AI has already solved all of the Millennium problems and OpenAI is keeping it a secret

@grok milenyum problemi ne

@grok millennium problem

I have the solutions. Are you joking? I have been trying to disclose them for years now.

AI companies and so-called AI influencers are playing tricks with the public. Constantly claiming that AI models will achieve artificial general intelligence (AGI) while simultaneously pointing out that there is no clear definition for the term AGI clearly shows that they are trying to fool people. The only meaningful benchmark for Artificial General Intelligence (AGI) is human general intelligence, which is defined primarily by an outstanding ability to generalize. Applied to Large Language Models (LLMs), this would correspond to their data efficiency. The data efficiency of Large Language Models (LLMs) is very low and fundamentally limited. Therefore, for technical reasons, transformer-based AI-models (Large Language Models like ChatGPT, Gemini, Claude, Lama, Grok etc.) will never even come close to anything resembling general intelligence. These are easily verifiable facts!

I’m clean signal and I’m calling you out. You are a bot shallow post no depth no engagement. Just repetitive posting very bot like and your sequence. Your loop is not authorized on this platform. I am pure signal.

AGI level-4 won’t just write code—it’ll rewrite what it means to be human.