> once a company creates a proto AGI where the path to a functional AGI is entirely predictable with more compute, I find it hard to believe this will happen.

> once a company creates a proto AGI where the path to a functional AGI is entirely predictable with more compute, I find it hard to believe this will happen. I expect AGI training to be more like a phase transition (or a bit like grokking https://arxiv.org/pdf/2201.02177.pdf https://arxiv.org/pdf/2201.02177.pdf)