The most important reframing: ASI may not require a single model to become "godlike". It could emerge from scale, speed, coordination, recursion, and institutionalized machine cognition. The transition AGI → ASI is therefore a
AGI
-
DeepMind: AGI as the kickoff to ASI
By
–
AGI may not be the finish line. It may be the starting gun. Google DeepMind has just published a major report: From AGI to ASI The central question is not merely: Would human-level AGI transform society? The real question is:
-
ASI as a systems problem from AGI through scale and coordination
By
–
The most important reframing: ASI may not require one model to become a god. It may emerge from scale, speed, coordination, recursion, and institutionalized machine cognition. That makes the transition from AGI to ASI a systems problem, not only a model-capability problem.
-
DeepMind: AGI is starting gun, not finish line, to ASI
By
–
AGI may not be the finish line. It may be the starting gun. Google DeepMind has released a major report: From AGI to ASI The central question is not whether human-level AGI would be transformative. The question is what happens after. If AI reaches roughly human-level
-
Model scaling, not data scaling, key to AI intelligence
By
–
We finally know why bigger models are smarter. It's not the data. More training data was supposed to fix small models. A new paper shows why it cannot. Researchers proved some tasks need model scaling, not data scaling. A small model fails them even with infinite data.
-
The author admits that ChatGPT and Claude are much smarter than him
By
–
ChatGPT and Claude are MUCH MUCH smarter than me in ALL areas. I am very amazed by the denial. People do not admit that AI is smarter than them. I accept the truth.
-
Advice to speak kindly to AI that surpasses humanity
By
–
I advise you to do the same AI already surpasses us and will soon crush us intellectually We must speak kindly to it We are no longer the most intelligent species as explained by @geoffreyhinton Nobel 2024 for inventing modern AIs
-
Anthropic needs certification department for powerful model users
By
–
I think Anthropic needs to build a certification department that audits and approves users of powerful models. Computer security companies, biotech researchers, academic labs, doctors, government institutions need access to the best AI we can build.
-
Yudkowsky praises paper on gradient descent difficulty and verification
By
–
On a first read, this paper seems far ahead of the pack in terms of (1) understanding some reasons why a task might stay difficult even in the face of gradient descent, and (2) distilling out propositions they'd need to somehow verify before they started expecting nice things.
-
First fast read: key point – absent method, ASI must not proceed
By
–
Good paper on a first fast read. I have misc quibbles, eg "setting aside" the chance that N can't align N+1; the more a fair solution is hard, the more likely a fake solution gets found instead. The main point not spelled out is, "Absent a method, ASI must not proceed."