Google co-founder Larry Page predicted back in 2007 that AI would come from raw computation, not clever algorithms.

Full video from @Google:

Sorry to say he was basically wrong. Or you could say he was half right. It was Google itself that published the transformer algorithm. In summary it's what's powering the foundation frontier models right now. And there are more clever you could call them algorithms coming out soon that use a far less processing power and energy. Clever algorithms and a lot of processing power are required

And I have discovered the "mathematical organism", first of its kind... an organism not coded but arising by simple rules like Collatz structures. The An(n) function yields fascinating graphs that look like organisms. Tail and head with dynamics internally From simply An(1)

2007 Page calling "scale eats algorithms" while everyone else was still hand-tuning SVMs — annoyingly correct

Μετά έβαλαν τον Ινδό CEO ο οποίος κάνει όλο και χειρότερη τη Google και τον κρατάνε στη θέση του επειδή φέρνει λεφτά στα ταμεία πρήζοντάς μας με διαφημίσεις και πώληση των data μας

Now he’s saying it’s clever algorithms not compute. I think it’s both

@xenoaesthetics

Algorithms still matter, but compute is what unlocked everything. He nailed the direction early.

🍎

Thanks. This is really helpful.

But that’s not true. Artificial neural networks with the transformer long-range dependency represent a staggeringly successful & profound learning algorithm.

The future of education will be proper Ai usage. Optimization for the human algorithm..

No wonder he created Google.

All computing is just "a lot of computations".