Saying "we've been saying for decades that general AI is 20 years away" is a really bad argument

It's an argument I see a lot lately, but it's essentially a Godwin's Law level argument - the first person to say "we've been saying that 20 years away for decades" basically wins the argument by poisoning the well and preventing any real discussion from happening.<br><br>In reality, people who say "AGI is 20 years away" are talking about a very specific type of intelligence - general-purpose human-level intelligence for a specific task. It doesn't make any more sense to compare that to "yeah, we were saying the same thing in the 1960s, when the field was in its infancy and we didn't even have any algorithms to train neural networks yet".<br><br>Sure, AI has been getting hyped up a lot in the past. But that doesn't really say anything about whether or not it's being hyped correctly now. LLMs and transformer-based AI models have been widely adopted only over the past 1-2 years. AGI really is 20 years away, and it has always been.<br><br> Of course, this is just my opinion. If you really think AGI is not 20 years away, then I have some good news for you: AGI is not 20 years away.