While there is some stuff around this that sounds like PR (probably some intermediate results and/or SOTA on that field as of today could help reaching that res

While there is some stuff around this that sounds like PR (probably some intermediate results and/or SOTA on that field as of today could help reaching that result), the process seems interesting. Seems like it launches ahem "agents" that do simulations, verifications, refines and iterates over different options... and it takes a while: https://research.google/blog/accelerating-scientific-breakthroughs-with-an-ai-co-scientist/ https://research.google/blog/accelerating-scientific-breakth...