Artificial superintelligence requires a combination of computing performance and algorithms that exceeds what we have now. It might be that it requires so much more of both that it’s very far away. And it might be that we have sufficient algorithms, but need so much more compute that it’s far away. And it might be that we have enough compute, but require such algorithmic advancement that it’s far away. It is also possible that we have sufficient compute and are just a few readily-accessible steps away from the needed algorithms such that it will arrive shortly and incrementally. Once it’s here, it will no longer develop incrementally, as it will be able to improve itself and perhaps even the materials science underlying its manufacturing processes at an exponential rate. That will be a very interesting time for humanity.