Replying to @null and @SimonDeDeo, @nattyover, @JannaLevin, @tegmark, @ylecun
Simon: Ah, but are those not meatspace assumptions? Does further alterity not increase the risk for conflict?
Thread · 5 tweets · 10 Sept 2017
Kin altruism might not mean much if you can reprogram yourself at will. Assuming it would mean anything long at a 50µs reproductive rate!
I would further assume that a superintelligent AI would understand evolution at least as well as we do. Could it not benefit from it…
…through internal simulations better than through extrinsic reproduction?
There are good scaling/distribution reasons why in meatspace twice of you beats you twice as big — I’m not sure they hold so well for AI.