3 ms·

It is crazy how the misleading marketing of calling data-driven algorithms "AI" has led to people thinking this stuff might be actually intelligent. We have no

It is crazy how the misleading marketing of calling data-driven algorithms "AI" has led to people thinking this stuff might be actually intelligent. We have not made any significant progress at all in the field of general artificial intelligence in the last what, 20, 30 years? The field is pretty much dead. Yeah, ChatGPT can sound very impressive but then again even good old ELIZA from the 60s was able to fool some people into seeing it as a therapist with just a bit of pattern matching. Many data driven solutions have become practical in recent years not because of breakthroughs in research but because it is simply more feasibly to acquire the huge amounts of training data and processing power that those models require. Thinking those models will one day magically achieve general intelligence by just becoming really good is akin to thinking a chess master will one day become so good at chess that they can run a marathon. That is not how it works.

circuit10 4y ago
“ We have not made any significant progress at all in the field of general artificial intelligence in the last what, 20, 30 years?” Well clearly you haven’t been following the progress at all because there’s been an incredible amount of progress in the last few years alone
Aaron2222 4y ago
Plenty of animals aren't very intelligent, but we still have ethics standards regarding the treatment of them, because we have the belief that they have the ability to suffer. Note that they do say "Most reinforcement learners in operation today likely do not have significant moral weight, but this could very well change as AI research develops", whatever they mean by 'moral weight'.
trabant00 4y ago
I agree with the marketing part and that refining learning models will not lead to AI. But I disagree that the field is dead. Testing all directions to exhaustion is the only way forward to achieving AI, if it will ever be achieved (note). The hype, while possibly misleading is what gets resources for exhausting the options we have now. (note) I for one do not see AI as being inevitable. It can well be that humans are not smart enough to create one, and that is only one possible way to fail in the quest.
hackeraccount 4y ago
It's like imagining that a novelist is going to write a character so well that they will literally jump off the page. That a magician will create an illusion so amazing that it will not just fool all who see it but the magician as well - that the road to actual magic is better and better slight of hand.

circuit10 4y ago

“ We have not made any significant progress at all in the field of general artificial intelligence in the last what, 20, 30 years?” Well clearly you haven’t been following the progress at all because there’s been an incredible amount of progress in the last few years alone

Aaron2222 4y ago

Plenty of animals aren't very intelligent, but we still have ethics standards regarding the treatment of them, because we have the belief that they have the ability to suffer. Note that they do say "Most reinforcement learners in operation today likely do not have significant moral weight, but this could very well change as AI research develops", whatever they mean by 'moral weight'.

trabant00 4y ago

I agree with the marketing part and that refining learning models will not lead to AI. But I disagree that the field is dead. Testing all directions to exhaustion is the only way forward to achieving AI, if it will ever be achieved (note). The hype, while possibly misleading is what gets resources for exhausting the options we have now. (note) I for one do not see AI as being inevitable. It can well be that humans are not smart enough to create one, and that is only one possible way to fail in the quest.

hackeraccount 4y ago

It's like imagining that a novelist is going to write a character so well that they will literally jump off the page. That a magician will create an illusion so amazing that it will not just fool all who see it but the magician as well - that the road to actual magic is better and better slight of hand.