What’s the brightest possible AI future you can imagine? Bonus points if you share how we’ll have overcome some of the obvious pitfalls.
40
27
63
Conversation
Brightest possible?
We figure out a good, stable way to train a powerful general purpose agent to predict what will make us happy and satisfied, and set one loose. It improves its own code and hardware, and eventually becomes arbitrarily powerful. Paradise ensues.
Lots of obvious pitfalls here. You need to be really sure it's pursuing the right goal before you turn it loose - but you can't take too long once the technology is widely available, or else you'll be beaten to the punch by someone else doing a worse job.
1
Oh, and of course, you also have to invent a technology that modern ML methods don't cover at all, which is the ability to accurately model and predict the outcomes of complex situations that the system has never observed during training.
1