@herbiebradley I'm imagining that AIs have better judgment than humans. I agree things will be bottlenecked on compute and potentially data. But I don't think humans will be adding any value once AIs are superintelligent. Maybe they won't hurt either, depending on how strongly they intervene.
I haven't yet seen a good argument for why the pace of overall progress wouldn't be bottlenecked by eg compute & data, such that there is no efficiency penalty (and potentially significant benefit via judgement) to keeping humans in the loop? So seems reasonable
Overall, I also think the picture described in Machines of Loving Grace is describing a substantially weaker type of AI than described in AI 2027, which is potentially illustrative of Anthropic's view on where the two scenarios come apart.
