The bigger risk for AI isn’t that it starts killing all of us it’s that we as a species become so dependent on it that we can’t do anything without it and lose our critical thinking skills. Don’t forget despite all the hoopla about navier stokes the technique that lead to solving it came from human intuition first it’s the grunt work that the AI did. AI has not intuition or insight into any significant problems and we still need humans for this, but it makes testing, verifying, and implementing much much faster. The bubble team has made several field changing discoveries as well using these tools but we don’t talk about it to avoid getting our alpha scooped. But none of these came directly from AI at all it’s all in my big head first. Big risk in the future that kids growing up with this stuff become like worse than iPad babies and suffer from attention deficit disorder and below average critical thinking skills. Some of the models have already surpassed me on some subjects and most of the output I get now just looks like word salad. (Not hardware yet thankfully).
