As a software engineer myself (though not an expert in AI by any means), I remain unconvinced that AGI is just around the corner. Quantum computing may very well be a game changer in that regard but there's still the problem of creating the software that makes it possible.
You are viewing a single comment's thread from:
That is a comforting bit of information to hear, now that I've read half of the first article provided in this post. I can't help but hope there is a limit, not an infinite upward curve, to what a machine can know. The proof is in the feelings - can a computer ever know those? I hope not. And I hope humans do not come to the day when we no longer find feelings of value.
AGI is going to take some time, no question. Once we get there ASI will literally be "around the corner". Right now almost all our AI is focused on one task and do it well. The step to all tasks is a huge leap and all efforts we have now are very focused on one task type of training.
I'm not even sure ASI would be around the corner though I understand the argument. I guess the question for me is how much more computing power would an ASI require over an AGI? If it's not much, then sure, it would be right around the corner. If it's orders of magnitude more, then it will take a while. I think there may be some physical constraints that limit the speed at which that happens.
The thing is, with AGI it can solve problems it wasn't trained on and becomes more human-like in that it can adapt to any problem thrown at it. With the raw processing power of endless computing, it can perform far better and faster than humans (although not smarter, that's when it turns into ASI). In other words, it can play out more scenarios faster than we can.