Apple just proved AI "reasoning" models like Claude, DeepSeek-R1, and o3-mini don't actually reason at all. They just memorize patterns really well.
-
Yah of course they do they’re computers
Computers are better at logic than brains are. We emulate logic; they do it natively.
It just so happens there's no logical algorithm for "reasoning" a problem through.
-
That's not really a valid argument for why, but yes the models which use training data to assemble statistical models are all bullshitting. TBH idk how people can convince themselves otherwise.
They aren't bullshitting because the training data is based on reality. Reality bleeds through the training data into the model. The model is a reflection of reality.
-
Most humans don't reason. They just parrot shit too. The design is very human.
LLMs deal with tokens. Essentially, predicting a series of bytes.
Humans do much, much, much, much, much, much, much more than that.
-
Funny how triggering it is for some people when anyone acknowledges humans are just evolved primates doing the same pattern matching.
It's not that institutionalized people don't follow "set" pattern matches. That's why you're getting downvotes.
Some of those humans can operate with the same brain rules alright. They may even be more efficient at it than you and I may. The higher level functions is a different thing.
-
It's not that institutionalized people don't follow "set" pattern matches. That's why you're getting downvotes.
Some of those humans can operate with the same brain rules alright. They may even be more efficient at it than you and I may. The higher level functions is a different thing.
That’s absolutely what it is. It’s a pattern on here. Any acknowledgment of humans being animals or less than superior gets hit with pushback.
-
LOOK MAA I AM ON FRONT PAGE
No shit. This isn't new.
-
That’s absolutely what it is. It’s a pattern on here. Any acknowledgment of humans being animals or less than superior gets hit with pushback.
Humans are animals. But an LLM is not an animal and has no reasoning abilities.
-
Yeah I've always said the the flaw in Turing's Imitation Game concept is that if an AI was indistinguishable from a human it wouldn't prove it's intelligent. Because humans are dumb as shit. Dumb enough to force one of the smartest people in the world take a ton of drugs which eventually killed him simply because he was gay.
I've heard something along the lines of, "it's not when computers can pass the Turing Test, it's when they start failing it on purpose that's the real problem."
-
Fair, but the same is true of me. I don't actually "reason"; I just have a set of algorithms memorized by which I propose a pattern that seems like it might match the situation, then a different pattern by which I break the situation down into smaller components and then apply patterns to those components. I keep the process up for a while. If I find a "nasty logic error" pattern match at some point in the process, I "know" I've found a "flaw in the argument" or "bug in the design".
But there's no from-first-principles method by which I developed all these patterns; it's just things that have survived the test of time when other patterns have failed me.
I don't think people are underestimating the power of LLMs to think; I just think people are overestimating the power of humans to do anything other than language prediction and sensory pattern prediction.
You either an llm, or don't know how your brain works.
-
Thank you Captain Obvious! Only those who think LLMs are like "little people in the computer" didn't knew this already.
Yeah, well there are a ton of people literally falling into psychosis, led by LLMs. So it’s unfortunately not that many people that already knew it.
-
Did you even read this garbage? It’s just words strung together without any meaning. The things it’s claiming show a fundamental lack of understanding of what it is responding to.
This didn’t prove your point at all, quite the opposite. And it wasted everyone’s time in the process. Good job, this was worthless.
I did and it was because it didn't have the previous context. But it did find the fallacies as present. Logic is literally what a chat AI is going. A human still needs to review the output but it did what it was asked. I don't know AI programming well. But I can say that logic is algorithmic. An AI has no problem parsing an argument and finding the fallacies. It's a tool like any other.
-
Excellent, these "fallacies" are exactly as I expected - made up, misunderstanding my comment (I did not call SMBC "bad"), and overall just trying to look like criticism instead of being one. Completely worthless - but I sure can see why right wingers are embracing it!
It's funny how you think AI will help refine people's ideas, but you actually just delegated your thinking to it and let it do it worse than you could (if you cared). That's why I don't feel like getting any deeper into explaining why the AI response is garbage, I could just as well fire up GPT on my own and paste its answer, but it would be equally meaningless and useless as yours.
Saying it’ll be boring comics missed the entire point.
So what was the point exactly? I re-read that part of your comment and you're talking about "strong ideas", whatever that's supposed to be without any actual context?
Saying it is the same as google is pure ignorance of what it can do.
I did not say it's the same as Google, in fact I said it's worse than Google because it can add a hallucinated summary or reinterpretation of the source. I've tested a solid number of LLMs over time, I've seen what they produce. You can either provide examples that show that they do not hallucinate, that they have access to sources that are more reliable than what shows up on Google, or you can again avoid any specific examples, just expecting people to submit to the revolutionary tech without any questions, accuse me of complete ignorance and, no less, compare me with anti-immigrant crowds (I honestly have no idea what's supposed to be similar between these two viewpoints? I don't live in a country with particularly developed anti-immigrant stances so maybe I'm missing something here?).
The people who buy into it will get into these type of ignorant and short sighted statements just to prove things that just are not true. But they’ve bought into the hype and need to justify it.
"They’ve bought into the hype and need to justify it"? Are you sure you're talking about the anti-AI crowd here? Because that's exactly how one would describe a lot of the pro-AI discourse. Like, many pro-AI people literally BUY into the hype by buying access to better AI models or invest in AI companies, the very real hype is stoked by these highly valued companies and some of the richest people in the world, and the hype leads the stock market and the objectively massive investments into this field.
But actually those who "buy into the hype" are the average people who just don't want to use this tech? Huh? What does that have to do with the concept of "hype"? Do you think hype is simply any trend that doesn't agree with your viewpoints?
Hype flows in both directions. Right now the hype from most is finding issues with chatgpt. It did find the fallacies based on what it was asked to do. It worked as expected. You act like this is fire and forget. Given what this output gave me, I can easily keep working this to get better and better arguments. I can review the results and clarify and iterate. I did copy and paste just to show an example. First I wanted to be honest with the output and not modify it. Second is an effort thing. I just feel like you can't honestly tell me that within 10 seconds having that summary is not beneficial. I didn't supply my argument to the prompt, only yours. If I submitted my argument it would be better.
-
Yeah, well there are a ton of people literally falling into psychosis, led by LLMs. So it’s unfortunately not that many people that already knew it.
-
Of course, that is obvious to all having basic knowledge of neural networks, no?
I still remember Geoff Hinton's criticisms of backpropagation.
IMO it is still remarkable what NNs managed to achieve: some form of emergent intelligence.
-
Humans are animals. But an LLM is not an animal and has no reasoning abilities.
wrote on last edited by [email protected]It’s built by animals, and it reflects them. That’s impressive on its own. Doesn’t need to be exaggerated.
-
Hype flows in both directions. Right now the hype from most is finding issues with chatgpt. It did find the fallacies based on what it was asked to do. It worked as expected. You act like this is fire and forget. Given what this output gave me, I can easily keep working this to get better and better arguments. I can review the results and clarify and iterate. I did copy and paste just to show an example. First I wanted to be honest with the output and not modify it. Second is an effort thing. I just feel like you can't honestly tell me that within 10 seconds having that summary is not beneficial. I didn't supply my argument to the prompt, only yours. If I submitted my argument it would be better.
Right now the hype from most is finding issues with chatgpt
publicity
especially : promotional publicity of an extravagant or contrived kind
You're abusing the meaning of "hype" in order to make the two sides appear the same, because you do understand that "hype" really describes the pro-AI discourse much better.
It did find the fallacies based on what it was asked to do.
It didn't. Put the text of your comment back into GPT and tell it to argue why the fallacies are misidentified.
You act like this is fire and forget.
But you did fire and forget it. I don't even think you read the output yourself.
First I wanted to be honest with the output and not modify it.
Or maybe you were just lazy?
Personally I'm starting to find these copy-pasted AI responses to be insulting. It has the "let me Google that for you" sort of smugness around it. I can put in the text in ChatGPT myself and get the same shitty output, you know. If you can't be bothered to improve it, then there's absolutely no point in pasting it.
Given what this output gave me, I can easily keep working this to get better and better arguments.
That doesn't sound terribly efficient. Polishing a turd, as they say. These great successes of AI are never actually visible or demonstrated, they're always put off - the tech isn't quite there yet, but it's just around the corner, just you wait, just one more round of asking the AI to elaborate, just one more round of polishing the turd, just a bit more faith on the unbelievers' part...
I just feel like you can’t honestly tell me that within 10 seconds having that summary is not beneficial.
Oh sure I can tell you that, assuming that your argumentative goals are remotely honest and you're not just posting stupid AI-generated criticism to waste my time. You didn't even notice one banal way in which AI misinterpreted my comment (I didn't say SMBC is bad), and you'd probably just accept that misreading in your own supposed rewrite of the text. Misleading summaries that you have to spend additional time and effort double checking for these subtle or not so subtle failures are NOT beneficial.
-
They aren't bullshitting because the training data is based on reality. Reality bleeds through the training data into the model. The model is a reflection of reality.
wrote on last edited by [email protected]An approximation of a very small limited subset of reality with more than a 1 in 20 error rate who produces massive amounts of tokens in quick succession is a shit representation of reality which is in every way inferior to human accounts to the point of being unusable for the industries in which they are promoted.
And that Error Rate can only spike when the training data contains errors itself, which will only grow as it samples its own content.
-
I did and it was because it didn't have the previous context. But it did find the fallacies as present. Logic is literally what a chat AI is going. A human still needs to review the output but it did what it was asked. I don't know AI programming well. But I can say that logic is algorithmic. An AI has no problem parsing an argument and finding the fallacies. It's a tool like any other.
That was a roundabout way of admitting you have no idea what logic is or how LLMs work. Logic works with propositions regardless of their literal meaning, LLMs operate with textual tokens irrespective of their formal logical relations. The chatbot doesn't actually do the logical operations behind the scenes, it only produces the text output that looks like the operations were done (because it was trained on a lot of existing text that reflects logical operations in its content).
-
Yeah I've always said the the flaw in Turing's Imitation Game concept is that if an AI was indistinguishable from a human it wouldn't prove it's intelligent. Because humans are dumb as shit. Dumb enough to force one of the smartest people in the world take a ton of drugs which eventually killed him simply because he was gay.
Yeah we’re so stupid we’ve figured out advanced maths, physics, built incredible skyscrapers and the LHC, we may as individuals be less or more intelligent but humans as a whole are incredibly intelligent
-
That entire paragraph is much better at supporting the precise opposite argument. Computers can beat Kasparov at chess, but they're clearly not thinking when making a move - even if we use the most open biological definitions for thinking.
By that metric, you can argue Kasparov isn't thinking during chess, either. A lot of human chess "thinking" is recalling memorized openings, evaluating positions many moves deep, and other tasks that map to what a chess engine does. Of course Kasparov is thinking, but then you have to conclude that the AI is thinking too. Thinking isn't a magic process, nor is it tightly coupled to human-like brain processes as we like to think.
