Hmm, not sure how to respond, sorry — that’s the point of the entire comment. No these systems cannot think every possible thought, but neither can humans. We have no reason to think their cognition is unsustainable in a way that ours isn’t.

The end of your response seems to be about fears of “model collapse” from a few years ago, but I am convinced those are strongly oversold. For one thing, there’s no need to gather new training data ever again; the training data for Opus is clearly sufficient for instilling reason, and that’s a sufficient condition for AGI -> RSI -> ASI.