[ioulaum]

The exponential part may be iffy, but it is self improving.

And this same RL is also creating improvements in small model performance.

So, more LLMs are about to rise in quality.

[immibis]

It's self-improving? So, we can ask AI how to improve AI, and the suggestions actually work?

[sesm]

It's more like Intel in early days using their CPUs to compute layout for bigger CPUs.

[ethbr1]

Effectively: is the limiting factor to improvement addressable by the thing being improved?

If yes, then you get exponential increases very trivially. If no, then something external continues to bottleneck progress.