I still don't understand how a chat bot could be called or could spawn "AI" let alone "AGI".

I think it's clear we can have ever better machines (which is all these are) and that they could be so sophisticated that they could trick a human into thinking they are intelligent, but they still lack even the very basics of intelligence.

> People naturally try to compare Agent-1 to humans, but it has a very different skill profile. It knows more facts than any human, knows practically every programming language, and can solve well-specified coding problems extremely quickly. On the other hand, Agent-1 is bad at even simple long-horizon tasks, like beating video games it hasn’t played before.

I think that sums it up. It's not going to get magically better at doing things it has never done before. Whereas a hamster or a human can approach a new task and intuit how to complete it. A machine that stamps out square widgets is not going to suddenly stamp out circle widgets. Yes, you can have a second machine that stamps out the parts the first machine would need to make circles, and have a third machine switch out the parts. That's very very efficient, will replace jobs, but it's not intelligence. It's just turtles all the way down.

Reply to this note

Please Login to reply.

Discussion

I’ve never come across a good definition of intelligence. Nonetheless I think it’s just a matter of these machines taking a different evolutionary tack in their maturation (maybe bc of its silicon substrate). Some parts of its cognition are superhuman and others are retarded. I bet its executive function comes around after enough rounds of Pokémon. After all, executive function came late for humans

extensIONz/*****