I run to the comment section of every article on AI as it's always, to me, a baffling outburst of boomer-like anti-AI sentiment. I would think most people who are into science and technology enough to frequent ARS, would be excited by the advancement and the possibilities the future hold for the technology. I understand hesitation at some of the negative possibilities but this isn't even what most comments are expressing.
There seems to be a lot of people upset that it's not truly AI. They repeat that loudly in the comments of every article. We understand that ChatGPT is not Data from Star Trek, or C3PO, but that's not how technology works, you don't just jump from nothing to Data. And strangely the same people that complain about this are also the ones with the doomsday predictions of it becoming HAL.
I think a lot of people when debating whether or not it is AI often get too focused in on the word 'intelligence' that they forget the word 'artificial' comes in front of it. This is directly from an Oxford definition for artificial:
(of a situation or concept) not existing naturally; contrived or false.
False, false intelligence, so not true intelligence but a simulation of it. That's exactly what LLMs are doing. They are trying to emulate intelligence, not be intelligence.
Yes we know they work by predicting the next token in a series of tokens, based on what the user has indicated they want generated. Which is why people seem to love to call it industrial grade autocomplete, and they're not necessarily wrong. However, if we look at social situations between humans, especially those who are neurodivergent, it works almost exactly the same. As someone with social anxiety and ADHD I can tell you that my mind is going 100 mph in social situations trying to come up with the next 'token' that is relevant and would contribute best to the current conversation. So I don't accept 'auto-complete' as a reason it's not artificial intelligence. And that's not even mentioning the multiple studies ongoing on 'emergent behaviors' of LLMs with a large number of parameters.
And you can disagree with those things, or think it's all BS. But what really concerns me is there is almost a reddit-like, hivemind, mob mentality about the whole thing, where if anyone tries to point out something differently or point out positive things about the development of AI that's ongoing, they get shouted down and down voted into oblivion until their viewpoint is literally hidden. Where is the contrast of opinion, where is the open minded discussion? Why does it seem like a large part of the ARS community feels completely justified in sitting on their porch shaking their fist and yelling at AI or AI enthusiasts to get off their lawn?
There seems to be a lot of people upset that it's not truly AI. They repeat that loudly in the comments of every article. We understand that ChatGPT is not Data from Star Trek, or C3PO, but that's not how technology works, you don't just jump from nothing to Data. And strangely the same people that complain about this are also the ones with the doomsday predictions of it becoming HAL.
I think a lot of people when debating whether or not it is AI often get too focused in on the word 'intelligence' that they forget the word 'artificial' comes in front of it. This is directly from an Oxford definition for artificial:
(of a situation or concept) not existing naturally; contrived or false.
False, false intelligence, so not true intelligence but a simulation of it. That's exactly what LLMs are doing. They are trying to emulate intelligence, not be intelligence.
Yes we know they work by predicting the next token in a series of tokens, based on what the user has indicated they want generated. Which is why people seem to love to call it industrial grade autocomplete, and they're not necessarily wrong. However, if we look at social situations between humans, especially those who are neurodivergent, it works almost exactly the same. As someone with social anxiety and ADHD I can tell you that my mind is going 100 mph in social situations trying to come up with the next 'token' that is relevant and would contribute best to the current conversation. So I don't accept 'auto-complete' as a reason it's not artificial intelligence. And that's not even mentioning the multiple studies ongoing on 'emergent behaviors' of LLMs with a large number of parameters.
And you can disagree with those things, or think it's all BS. But what really concerns me is there is almost a reddit-like, hivemind, mob mentality about the whole thing, where if anyone tries to point out something differently or point out positive things about the development of AI that's ongoing, they get shouted down and down voted into oblivion until their viewpoint is literally hidden. Where is the contrast of opinion, where is the open minded discussion? Why does it seem like a large part of the ARS community feels completely justified in sitting on their porch shaking their fist and yelling at AI or AI enthusiasts to get off their lawn?
