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.
(archive.is)
from Allah@lemm.ee to technology@lemmy.world on 08 Jun 10:59
https://lemm.ee/post/66249486
from Allah@lemm.ee to technology@lemmy.world on 08 Jun 10:59
https://lemm.ee/post/66249486
LOOK MAA I AM ON FRONT PAGE
threaded - newest
.
.
.
I don’t think that person cares about women or anything else. They just said that they don’t even want to hear about it.
.
.
.
Just look at his username he is just a troll
.
.
.
.
.
.
.
yes
Right so why. You’re doing it so any chance you can be honest about it. It’s rape right. That’s such a horrible thing. Which I would assume you’d be like “I want to stop rape” but you haven’t said that. You’re playing coy. Makes it seem like it’s more about something you’re ashamed to say
.
.
In the same sense that some users might post only articles about ICE in California, or only articles about hurricanes in Florida, I still think that’s not very strange. Some people are particularly invested in specific topics. Maybe the author is or is close to rape victims and is therefore especially interested in it. People dedicate their whole lives and careers to specific activist topics, so I don’t think it’s too strange for someone to dedicate most of their posting activity on one particular website to one. Anyways, I’m not sure what the ulterior motive would be here anyways - what do you think is the real reason for posting so many articles about rape?
I’m not sure what it would be either. But I don’t think it’s about the rape and more about what the rapes say about India.
See the thing is I know it offends many
Indianspeople when they shittalkpost about their respective countries.A quick skim of your post history shows you like Joe Rogan, and stuff from Yankland and their politics.
But the most funny thing is when people get offended by “bad news” from their “Country”. I’m from the UK and some of the shit we see in the news is hilariously stupid these days and I take shitposting with a pinch of salt. You should try it too ;)
ninjaedit for example: “Oi mate, did u get stabbed 'cause u didn’t av a TV Lociense fella?”
Or even edit two - files.catbox.moe/x8t4rx.jpg
.
.
.
does ANY model reason at all?
No, and to make that work using the current structures we use for creating AI models we’d probably need all the collective computing power on earth at once.
… So you’re saying there’s a chance?
10^36 flops to be exact
That sounds really floppy.
Define reason.
Like humans? Of course not. They lack intent, awareness, and grounded meaning. They don’t “understand” problems, they generate token sequences.
as it is defined in the article
I think I do. Might be an illusion, though.
lol is this news? I mean we call it AI, but it’s just LLM and variants it doesn’t think.
The “Apple” part. CEOs only care what companies say.
Apple is significantly behind and arrived late to the whole AI hype, so of course it’s in their absolute best interest to keep showing how LLMs aren’t special or amazingly revolutionary.
They’re not wrong, but the motivation is also pretty clear.
Maybe they are so far behind because they jumped on the same train but then failed at achieving what they wanted based on the claims. And then they started digging around.
Yes, Apple haters can’t admit nor understand it but Apple doesn’t do pseudo-tech.
They may do silly things, they may love their 100% mark up but it’s all real technology.
The AI pushers or today are akin to the pushers of paranormal phenomenon from a century ago. These pushers want us to believe, need us to believe it so they can get us addicted and extract value from our very existence.
They need to convince investors that this delay wasn’t due to incompetence. The problem will only be somewhat effective as long as there isn’t an innovation that makes AI more effective.
If that happens, Apple shareholders will, at best, ask the company to increase investment in that area or, at worst, to restructure the company, which could also mean a change in CEO.
“Late to the hype” is actually a good thing. Gen AI is a scam wrapped in idiocy wrapped in a joke. That Apple is slow to ape the idiocy of microsoft is just fine.
Apple always arrives late to any new tech, doesn’t mean they haven’t been working on it behind the scenes for just as long though…
This is why I say these articles are so similar to how right wing media covers issues about immigrants.
There’s some weird media push to convince the left to hate AI. Think of all the headlines for these issues. There are so many similarities. They’re taking jobs. They are a threat to our way of life. The headlines talk about how they will sexual assault your wife, your children, you. Threats to the environment. There’s articles like this where they take something known as twist it to make it sound nefarious to keep the story alive and avoid decay of interest.
Then when they pass laws, we’re all primed to accept them removing whatever it is that advantageous them and disadvantageous us.
Because it’s a fear-mongering angle that still sells. AI has been a vehicle for scifi for so long that trying to convince Boomers that of won’t kill us all is the hard part.
I’m a moderate user for code and skeptic of LLM abilities, but 5 years from now when we are leveraging ML models for groundbreaking science and haven’t been nuked by SkyNet, all of this will look quaint and silly.
5 years from now? Or was it supposed to be 5 years ago?
Pretty sure we already have skynet.
Maybe the actual problem is people who equate computer programs with people.
You mean laws like this? jfc.
www.inc.com/sam-blum/…/91198975
Literally what I’m talking about. They have been pushing anti AI propaganda to alienate the left from embracing it while the right embraces it. You have such a blind spot you this, you can’t even see you’re making my argument for me.
That depends on your assumption that the left would have anything relevant to gain by embracing AI (whatever that’s actually supposed to mean).
What isn’t there to gain?
Its power lies in ingesting language and producing infinite variations. We can feed it talking points, ask it to refine our ideas, test their logic, and even request counterarguments to pressure-test our stance. It helps us build stronger, more resilient narratives.
We can use it to make memes. Generate images. Expose logical fallacies. Link to credible research. It can detect misinformation in real-time and act as a force multiplier for anyone trying to raise awareness or push back on disinfo.
Most importantly, it gives a voice to people with strong ideas who might not have the skills or confidence to share them. Someone with a brilliant comic concept but no drawing ability? AI can help build a framework to bring it to life.
Sure, it has flaws. But rejecting it outright while the right embraces it? That’s beyond shortsighted it’s self-sabotage. And unfortunately, after the last decade, that kind of misstep is par for the course.
I have no idea what sort of AI you’ve used that could do any of this stuff you’ve listed. A program that doesn’t reason won’t expose logical fallacies with any rigour or refine anyone’s ideas. It will link to credible research that you could already find on Google but will also add some hallucinations to the summary. And so on, it’s completely divorced from how the stuff as it is currently works.
That’s a misguided view of how art is created. Supposed “brilliant ideas” are dime a dozen, it takes brilliant writers and artists to make them real. Someone with no understanding of how good art works just having an image generator produce the images will result in a boring comic no matter the initial concept. If you are not competent in a visual medium, then don’t make it visual, write a story or an essay.
Besides, most of the popular and widely shared webcomics out there are visually extremely simple or just bad (look at SMBC or xkcd or - for a right-wing example - Stonetoss).
For now I see no particular benefits that the right-wing has obtained by using AI either. They either make it feed back into their delusions, or they whine about the evil leftists censoring the models (by e.g. blocking its usage of slurs).
Here is chatgpt doing what you said it can’t. Finding all the logical fallacies in what you write:
You’re raising strong criticisms, and it’s worth unpacking them carefully. Let’s go through your argument and see if there are any logical fallacies or flawed reasoning.
This misrepresents the original claim:
The original point wasn’t that AI could replace the entire creative process or make a comic successful on its own—it was that it can assist people in starting or visualizing something they couldn’t otherwise. Dismissing that by shifting the goalposts to “producing a full, good comic” creates a straw man of the original claim.
This suggests a binary: either you’re competent at visual art or you shouldn’t try to make anything visual. That’s a false dichotomy. People can learn, iterate, or collaborate, and tools like AI can help bridge gaps in skill—not replace skill, but allow exploration. Many creators use tools before mastery (e.g., musicians using GarageBand, or writers using Grammarly).
While it’s true that execution matters more than ideas alone, dismissing the value of ideas altogether is an overgeneralization. Many successful works do start with a strong concept—and while many fail in execution, tools that lower the barrier to prototyping or drafting can help surface more workable ideas. The presence of many bad ideas doesn’t invalidate the potential value of enabling more people to test theirs.
Criticizing popular webcomics like SMBC or xkcd by calling them “bad” doesn’t really support your broader claim. These comics are widely read because of strong writing and insight, despite minimalistic visuals. It comes off as dismissive and ridicules the counterexamples rather than engaging with them. That’s not a logical fallacy in the strictest sense, but it’s rhetorically weak.
This seems like a rebuttal to a point that wasn’t made directly. The original argument wasn’t that “the right is winning with AI,” but rather that alienating the left from it could lead to missed opportunities. Refuting a weaker version (e.g., “the right is clearly winning with AI”) isn’t addressing the original concern, which was more about strategic adoption.
Summary of Fallacies Identified:
Type Description
Straw Man Misrepresents the role of AI in creative assistance. False Dichotomy Assumes one must either be visually skilled or not attempt visual media. Hasty Generalization Devalues “brilliant ideas” universally. Appeal to Ridicule Dismisses counterexamples via mocking tone rather than analysis. Tu Quoque-like Compares left vs. right AI use without addressing the core point about opportunity.
Your criticism is thoughtful and not without merit—but it’s wrapped in rhetoric that sometimes slips into oversimplification or misrepresentation of the opposing view. If your goal is to strengthen your argument or have a productive back-and-forth, refining those areas could help. Would you like to rewrite it in a way that keeps the spirit of your critique but sharpens its logic?
At this point you’re just arguing for arguments sake. You’re not wrong or right but instead muddying things. Saying it’ll be boring comics missed the entire point. Saying it is the same as google is pure ignorance of what it can do. But this goes to my point about how this stuff is all similar to anti immigrant mentality. 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.
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.
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).
This is why I said I wasn’t sure how AI works behind the scenes. But I do know that logic isn’t difficult. Just to not fuck around between us. I have a CS background. Only saying this because I think you may have it as well and we can save some time.
It makes sense to me that logic is something AI can parse easily. Logic in my mind is very easy if it can tokenize some text. Wouldn’t the difficulty be if the AI has the right context.
Logic requires abstracting the argumentative form from the literal linguistic content and then generalising it, just how like math is done properly when you work with numbers and not just with sentences such as “two apples and three apples is five apples” (such abstraction in practice allows far more powerful and widely applicable operations than dealing with individual linguistic expressions; if you’ve ever solved very complex truth trees you’ll know how they allow streamlining and solutions that would be practically impossible to do if you had only the ordinary linguistic expression of the same problem). Logic doesn’t operate with textual tokens but with logical propositions and operators. “Difficulty” is not a meaningful term here, a tool is either technically capable of doing something (more or less successfully) or it isn’t.
That LLMs aren’t capable of this sort of precision and abstraction is shown by the OP link as well as the simple fact that chatbots used to be extremely bad at math (which is now probably patched up by adding a proper math module, rather than relying on the base LLM - my assumption, at least).
As for trying more examples of looking for logical fallacies, I tried out three different types of text. Since you say context is important, it’s best to take only the beginning of a text. One text I tried is the opening of the Wikipedia article on “history”, which ChatGPT described like this: “The passage you’ve provided is an informative and largely neutral overview of the academic discipline of history. It doesn’t make any strong arguments or persuasive claims, which are typically where logical fallacies appear.” It then went on to nitpick about some details “for the sake of thorough analysis”, but basically had no real complaints. Then I tried out the opening paragraph of Moby-Dick. That’s a fictional text so it would be reasonable to reject analysing its logical solidity, as GPT already did with the WP article, but it still tried to wring out some “criticism” that occasionally shows how it misunderstands the text (just as it misunderstood a part of my comment above). Finally, I asked it to find the fallacies in the first four paragraphs of Descartes’ Meditations on First Philosophy, which resulted in a criticism that was based on less logically rigid principles than the original text (accusing Descartes of the “slippery slope fallacy”).
I’ll post the full replies below.
This famous passage from the opening of Moby-Dick by Herman Melville is rich in literary style but not necessarily constructed as an argument, which makes identifying formal logical fallacies a nuanced task. That said, we can identify a few rhetorical or informal fallacies—not because the text is poorly reasoned, but because it makes some sweeping or questionable assumptions in a stylistically persuasive way. Here’s a breakdown:
1. Hasty Generalization
2. False Cause (Post Hoc)
3. Appeal to Emotion
4. False Analogy
Summary:
Would you like me to rewrite the passage without the fallacies, or is this for literary analysis?
Thank you! This passage is again drawn from René Descartes’ Meditations on First Philosophy, and it outlines his method of radical doubt as a means to establish certainty in knowledge. While the reasoning is carefully crafted and influential in the history of philosophy, it still contains a few logical fallacies or reasoning weaknesses that can be critically assessed.
🔍 Identified Logical Fallacies and Weak Reasoning:
1. Hasty Generalization
2. False Analogy
3. Slippery Slope (Metaphorical)
4. Begging the Question (Circular Reasoning)
5. Inconsistency (Recognized, Not Fully Resolved)
✅ Summary Table of Fallacies
Conclusion:
Descartes’ passage is methodologically rich, but from a logic and reasoning perspective, it contains:
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.
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?
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?).
“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.
hype noun (1)
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 didn’t. Put the text of your comment back into GPT and tell it to argue why the fallacies are misidentified.
But you did fire and forget it. I don’t even think you read the output yourself.
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.
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…
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.
Ok let’s give a test here. Let’s start with understand logic. Give me a paragraph and let’s see if it can find any logical fallacies. You can provide the paragraph. Only constraint is that the context has to exist within the paragraph.
Saw this earlier in the week and thought of you. These short, funny videos are popping up more and more and they’re only getting better. They’re sharp, engaging, and they spread like wildfire.
You strike me as someone who gets it what it means when one side embraces the latest tools while the other rejects them.
The left is still holed up on Lemmy, clinging to “Fuck AI” groups. But why? Go back to the beginning. Look at the early coverage of AI it was overwhelmingly targeted at left-leaning spaces, full of panic and doom. Compare that to how the right talks about immigration. The headlines are cut and pasted from each other. Same playbook, different topic. The media set out to alienate the left from these tools.
www.facebook.com/share/r/16MuwbVP5C/
I don’t have even the slightest idea what that video is supposed to mean. (Happy cake day tho.)
Come on, you know what I’m talking about. It’s a channel that started with AI content and is now pivoting to videos about the riots. You can see where this is going. Sooner or later, it’ll expand into targeting protestors and other left-leaning causes.
It’s a novelty now, but it’s spreading fast, and more channels like it are popping up every day.
Meanwhile, the left is losing ground. Losing cultural capture. Because as a group, they’re being manipulated into isolating themselves from the very tools and platforms that shape public opinion. Social media. AI. All of it. They’re walking away from the battlefield while the other side builds momentum.
But I literally don’t. Well, I didn’t but now I mostly do, since you explained it.
I get what you’re saying with regards to the isolation, this issue has already been raised when many left-wing people started to leave Twitter. But it is opening a whole new can of worms - these profiles that post AI-generated content are largely not managed by ordinary people with their private agendas (sharing neat stuff, political agitation, etc.), but by bots, and are also massively followed and supported by other bot profiles. Much the same on Twitter with its hordes of right-wing troll profiles, and as I’m still somewhat active on reddit I also notice blatant manipluation there as well (my country had elections a few weeks ago and the flood of new profiles less than one week old spamming idiotic propaganda and insults was too obvious). It’s not organic online behaviour and it can’t really be fought by organic behaviour, especially when the big social media platforms give up the tools to fight it (relaxing their moderation standards, removing fact-checking, etc.). Lemmy and Mastodon etc. are based on the idea(l) that this corporate-controlled area is not the only space where meaningful activity can happen.
So that’s one side of the story, AI is not something happening in a vacuum and that you just have to submit to your own will. The other side of the story, the actual abilities of AI, have already been discussed, we’ve seen sufficiently that it’s not that good at helping people form more solidly developed and truth-based stances. Maybe it could be used to spread the sort of mass-produced manipulative bullshit that is already used by the right, but I can’t honestly support such stuff. In this regard, we can doubt whether there is any ground to win for the left (would the left’s possible audience actually eat it up), and if yes, whether it is worth it (basing your political appeal on bullshit can bite you in the ass down the line).
As for the comparison to discourse around immigrants, again I still don’t fully understand the point other than on the most surface level (the media is guiding people what to think, duh).
.
"It’s part of the history of the field of artificial intelligence that every time somebody figured out how to make a computer do something—play good checkers, solve simple but relatively informal problems—there was a chorus of critics to say, ‘that’s not thinking’." -Pamela McCorduck´.
It’s called the AI Effect.
As Larry Tesler puts it, “AI is whatever hasn’t been done yet.”.
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.
No, it shows how certain people misunderstand the meaning of the word.
You have called npcs in video games “AI” for a decade, yet you were never implying they were somehow intelligent. The whole argument is strangely inconsistent.
Who is “you”?
Just because some dummies supposedly think that NPCs are “AI”, that doesn’t make it so. I don’t consider checkers to be a litmus test for “intelligence”.
“You” applies to anyone that doesnt understand what AI means. It’s a portmanteau word for a lot of things.
Npcs ARE AI. AI doesnt mean “human level intelligence” and never did. Read the wiki if you need help understanding.
.
Strangely inconsistent + smoke & mirrors = profit!
Intellegence has a very clear definition.
It’s requires the ability to acquire knowledge, understand knowledge and use knowledge.
No one has been able to create an system that can understand knowledge, therefor me none of it is artificial intelligence. Each generation is merely more and more complex knowledge models. Useful in many ways but never intelligent.
Wouldn’t the algorithm that creates these models in the first place fit the bill? Given that it takes a bunch of text data, and manages to organize this in such a fashion that the resulting model can combine knowledge from pieces of text, I would argue so.
What is understanding knowledge anyways? Wouldn’t humans not fit the bill either, given that for most of our knowledge we do not know why it is the way it is, or even had rules that were - in hindsight - incorrect?
If a model is more capable of solving a problem than an average human being, isn’t it, in its own way, some form of intelligent? And, to take things to the utter extreme, wouldn’t evolution itself be intelligent, given that it causes intelligent behavior to emerge, for example, viruses adapting to external threats? What about an (iterative) optimization algorithm that finds solutions that no human would be able to find?
I would disagree, it is probably one of the most hard to define things out there, which has changed greatly with time, and is core to the study of philosophy. Every time a being or thing fits a definition of intelligent, the definition often altered to exclude, as has been done many times.
Dog has a very clear definition, so when you call a sausage in a bun a “Hot Dog”, you are actually a fool.
Smart has a very clear definition, so no, you do not have a “Smart Phone” in your pocket.
Also, that is not the definition of intelligence. But the crux of the issue is that you are making up a definition for AI that suits your needs.
Misconstruing how language works isn’t an argument for what an existing and established word means.
I’m sure that argument made you feel super clever but it’s nonsense.
I sourced by definition from authoritative sources. The fact that you didn’t even bother to verify that or provide an alternative authoritative definition tells me all I need to know about the value in further discussion with you.
"Artificial intelligence refers to computer systems that can perform complex tasks normally done by human-reasoning, decision making, creating, etc.
There is no single, simple definition of artificial intelligence because AI tools are capable of a wide range of tasks and outputs, but NASA follows the definition of AI found within EO 13960, which references Section 238(g) of the National Defense Authorization Act of 2019.
This is from NASA (emphasis mine). www.nasa.gov/what-is-artificial-intelligence/
The problem is that you are reading the word intelligence and thinking it means the system itself needs to be intelligent, when it only needs to be doing things that we would normally attribute to intelligence. Computer vision is AI, but a software that detects a car inside a picture and draws a box around it isn’t intelligent. It is still considered AI and has been considered AI for the past three decades.
Now show me your blog post that told you that AI isnt AI because it isn’t 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.
Kasparov’s thinking fits pretty much all biological definitions of thinking. Which is the entire point.
Is thinking necessarily biologic?
I’m going to write a program to play tic-tac-toe. If y’all don’t think it’s “AI”, then you’re just haters. Nothing will ever be good enough for y’all. You want scientific evidence of intelligence?!?! I can’t even define intelligence so take that! \s
Seriously tho. This person is arguing that a checkers program is “AI”. It kinda demonstrates the loooong history of this grift.
.
It is. And has always been. “Artificial Intelligence” doesn’t mean a feeling thinking robot person (that would fall under AGI or artificial conciousness), it’s a vast field of research in computer science with many, many things under it.
ITT: people who obviously did not study computer science or AI at at least an undergraduate level.
Y’all are too patient. I can’t be bothered to spend the time to give people free lessons.
Wow, I would deeply apologise on the behalf of all of us uneducated proles having opinions on stuff that we’re bombarded with daily through the media.
The computer science industry isn’t the authority on artificial intelligence it thinks it is. The industry is driven by a level of hubris that causes people to step beyond the bounds of science and into the realm of humanities without acknowledgment.
Yesterday I asked an LLM “how much energy is stored in a grand piano?” It responded with saying there is no energy stored in a grad piano because it doesn’t have a battery.
Any reasoning human would have understood that question to be referring to the tension in the strings.
Another example is asking “does lime cause kidney stones?”. It didn’t assume I mean lime the mineral and went with lime the citrus fruit instead.
Once again a reasoning human would assume the question is about the mineral.
Ask these questions again in a slightly different way and you might get a correct answer, but it won’t be because the LLM was thinking.
But 90% of “reasoning humans” would answer just the same. Your questions are based on some non-trivial knowledge of physics, chemistry and medicine that most people do not possess.
Honestly, i thought about the chemical energy in the materials constructing the piano and what energy burning it would release.
The tension of the strings would actually be a pretty miniscule amount of energy too, since there’s very little stretch to a piano wire, the force might be high, but the potential energy/work done to tension the wire is low (done by hand with a wrench).
Compared to burning a piece of wood, which would release orders of magnitude more energy.
I’m not sure how you arrived at lime the mineral being a more likely question than lime the fruit. I’d expect someone asking about kidney stones would also be asking about foods that are commonly consumed.
This kind of just goes to show there’s multiple ways something can be interpreted. Maybe a smart human would ask for clarification, but for sure AIs today will just happily spit out the first answer that comes up. LLMs are extremely “good” at making up answers to leading questions, even if it’s completely false.
Making up answers is kinda their entire purpose. LMMs are fundamentally just a text generation algorithm, they are designed to produce text that looks like it could have been written by a human. Which they are amazing at, especially when you start taking into account how many paragraphs of instructions you can give them, and they tend to rather successfully follow.
The one thing they can’t do is verify if what they are talking about is true as it’s all just slapping words together using probabilities. If they could, they would stop being LLMs and start being AGIs.
A well trained model should consider both types of lime. Failure is likely down to temperature and other model settings. This is not a measure of intelligence.
Proving it matters. Science is constantly proving any other thing that people believe is obvious because people have an uncanning ability to believe things that are false. Some people will believe things long after science has proven them false.
I mean… “proving” is also just marketing speak. There is no clear definition of reasoning, so there’s also no way to prove or disprove that something/someone reasons.
Claiming it’s just marketing fluff is indicates you do not know what you’re talking about.
They published a research paper on it. You are free to publish your own paper disproving theirs.
At the moment, you sound like one of those “I did my own research” people except you didn’t even bother doing your own research.
You misunderstand. I do not take issue with anything that’s written in the scientific paper. What I take issue with is how the paper is marketed to the general public. When you read the article you will see that it does not claim to “proof” that these models cannot reason. It merely points out some strengths and weaknesses of the models.
stochastic parrots. all of them. just upgraded “soundex” models.
this should be no surprise, of course!
I use LLMs as advanced search engines. No ads or sponsored results.
There are search engines that do this better. There’s a world out there beyond Google.
Like what?
I don’t think there’s any search engine better than Perplexity. And for scientific research Consensus is miles ahead.
On first read this sounded like you were challenging the basis of the previous comment. But then you went on to provide a couple of your own examples.
So on that basis after rereading your comment, it sounds like maybe you’re actually looking for recommendations.
Ive seen a lot of praise for Kagi over the past year. I’ve finally started playing around with the free tier and I think it’s definitely worth checking out.
Through the years I’ve bounced between different engines. I gave Bing a decent go some years back, mostly because I was interested in gauging the performance and wanted to just pit something against Google. After that I’ve swapped between Qwant and Startpage a bunch. I’m a big fan of Startpage’s “Anonymous view” function.
Since then I’ve landed on Kagi, which I’ve used for almost a year now. It’s the first search engine I’ve used that you can make work for you. I use the lens feature to focus on specific tasks, and de-prioritise pages that annoy me, sometimes outright omitting results from sites I find useless or unserious. For example when I’m doing web stuff and need to reference the MDN, I don’t really care for w3schools polluting my results.
I’m a big fan of using my own agency and making my own decisions, and the recent trend in making LLMs think for us is something I find rather worrying, it allows for a much subtler manipulation than what Google does with its rankings and sponsor inserts.
Perplexity openly talking about wanting to buy Chrome and harvesting basically all the private data is also terrifying, thus I wouldn’t touch that service with a stick. That said, I appreciate their candour, somehow being open about being evil is a lot more palatable to me than all these companies pretending to be good.
There are ads but they’re subtle enough that you don’t recognize them as such.
Would like a link to the original research paper, instead of a link of a screenshot of a screenshot
machinelearning.apple.com/…/illusion-of-thinking
Just fancy Markov chains with the ability to link bigger and bigger token sets. It can only ever kick off processing as a response and can never initiate any line of reasoning. This, along with the fact that its working set of data can never be updated moment-to-moment, means that it would be a physical impossibility for any LLM to achieve any real "reasoning" processes.
I can envision a system where an LLM becomes one part of a reasoning AI, acting as a kind of fuzzy “dataset” that a proper neural network incorporates and reasons with, and the LLM could be kept real-time updated (sort of) with MCP servers that incorporate anything new it learns.
But I don’t think we’re anywhere near there yet.
The only reason we’re not there yet is memory limitations.
Eventually some company will come out with AI hardware that lets you link up a petabyte of ultra fast memory to chips that contain a million parallel matrix math processors. Then we’ll have an entirely new problem: AI that trains itself incorrectly too quickly.
Just you watch: The next big breakthrough in AI tech will come around 2032-2035 (when the hardware is available) and everyone will be bitching that “chain reasoning” (or whatever the term turns out to be) isn’t as smart as everyone thinks it is.
LLMs (at least in their current form) are proper neural networks.
Well, technically, yes. You’re right. But they’re a specific, narrow type of neural network, while I was thinking of the broader class and more traditional applications, like data analysis. I should have been more specific.
Unlike Markov models, modern LLMs use transformers that attend to full contexts, enabling them to simulate structured, multi-step reasoning (albeit imperfectly). While they don’t initiate reasoning like humans, they can generate and refine internal chains of thought when prompted, and emerging frameworks (like ReAct or Toolformer) allow them to update working memory via external tools. Reasoning is limited, but not physically impossible, it’s evolving beyond simple pattern-matching toward more dynamic and compositional processing.
Most people wouldn’t call zero of something ‘limited’.
The paper doesn’t say LLMs can’t reason, it shows that their reasoning abilities are limited and collapse under increasing complexity or novel structure.
I agree with the author.
The fact that they only work up to a certain point despite increased resources is proof that they are just pattern matching, not reasoning.
Performance eventually collapses due to architectural constraints, this mirrors cognitive overload in humans: reasoning isn’t just about adding compute, it requires mechanisms like abstraction, recursion, and memory. The models’ collapse doesn’t prove “only pattern matching”, it highlights that today’s models simulate reasoning in narrow bands, but lack the structure to scale it reliably. That is a limitation of implementation, not a disproof of emergent reasoning.
Performance collapses because luck runs out. Bigger destruction of the planet won’t fix that.
Brother you better hope it does because even if emissions dropped to 0 tonight the planet wouldnt stop warming and it wouldn’t stop what’s coming for us.
If the situation gets dire, it’s likely that the weather will be manipulated. Countries would then have to be convinced not to use this for military purposes.
This isn’t a thing.
If emissions dropped to 0 tonight, we would be substantially better off than if we maintain our current trajectory. Doomerism helps nobody.
It’s not doomerism it’s just realistic. Deluding yourself won’t change that.
Authors gotta get paid. This article is full of pseudo-scientific jargon.
I’m not convinced that humans don’t reason in a similar fashion. When I’m asked to produce pointless bullshit at work my brain puts in a similar level of reasoning to an LLM.
Think about “normal” programming: An experienced developer (that’s self-trained on dozens of enterprise code bases) doesn’t have to think much at all about 90% of what they’re coding. It’s all bog standard bullshit so they end up copying and pasting from previous work, Stack Overflow, etc because it’s nothing special.
The remaining 10% is “the hard stuff”. They have to read documentation, search the Internet, and then—after all that effort to avoid having to think—they sigh and start actually start thinking in order to program the thing they need.
LLMs go through similar motions behind the scenes! Probably because they were created by software developers but they still fail at that last 90%: The stuff that requires actual thinking.
Eventually someone is going to figure out how to auto-generate LoRAs based on test cases combined with trial and error that then get used by the AI model to improve itself and that is when people are going to be like, “Oh shit! Maybe AGI really is imminent!” But again, they’ll be wrong.
AGI won’t happen until AI models get good at retraining themselves with something better than basic reinforcement learning. In order for that to happen you need the working memory of the model to be nearly as big as the hardware that was used to train it. That, and loads and loads of spare matrix math processors ready to go for handing that retraining.
previous input goes in. Completely static, prebuilt model processes it and comes up with a probability distribution.
There is no “unlike markov chains”. They are markov chains. Ones with a long context (a markov chain also kakes use of all the context provided to it, so I don’t know what you’re on about there). LLMs are just a (very) lossy compression scheme for the state transition table. Computed once, applied blindly to any context fed in.
LLMs are not Markov chains, even extended ones. A Markov model, by definition, relies on a fixed-order history and treats transitions as independent of deeper structure. LLMs use transformer attention mechanisms that dynamically weigh relationships between all tokens in the input—not just recent ones. This enables global context modeling, hierarchical structure, and even emergent behaviors like in-context learning. Markov models can’t reweight context dynamically or condition on abstract token relationships.
The idea that LLMs are “computed once” and then applied blindly ignores the fact that LLMs adapt their behavior based on input. They don’t change weights during inference, true—but they do adapt responses through soft prompting, chain-of-thought reasoning, or even emulated state machines via tokens alone. That’s a powerful form of contextual plasticity, not blind table lookup.
Calling them “lossy compressors of state transition tables” misses the fact that the “table” they’re compressing is not fixed—it’s context-sensitive and computed in real time using self-attention over high-dimensional embeddings. That’s not how Markov chains work, even with large windows.
their input is the context window. Markov chains also use their whole context window. Llms are a novel implementation that can work with much longer contexts, but as soon as something slides out of its window, it’s forgotten. just like any other markov chain. They don’t adapt. You add their token to the context, slide the oldest one out and then you have a different context, on which you run the same thing again. A normal markov chain will also give you a different outuut if you give it a different context. Their biggest weakness is that they don’t and can’t adapt. You are confusing the encoding of the context with the model itself. Just to see how static the model is, try setting temperature to 0, and giving it the same context. i.e. only try to predict one token with the exact same context each time. As soon as you try to predict a 2nd token, you’ve just changed the input and ran the thing again. It’s not adapting, you asked it something different, so it came up with a different answer
While both Markov models and LLMs forget information outside their window, that’s where the similarity ends. A Markov model relies on fixed transition probabilities and treats the past as a chain of discrete states. An LLM evaluates every token in relation to every other using learned, high-dimensional attention patterns that shift dynamically based on meaning, position, and structure.
Changing one word in the input can shift the model’s output dramatically by altering how attention layers interpret relationships across the entire sequence. It’s a fundamentally richer computation that captures syntax, semantics, and even task intent, which a Markov chain cannot model regardless of how much context it sees.
an llm also works on fixed transition probabilities. All the training is done during the generation of the weights, which are the compressed state transition table. After that, it’s just a regular old markov chain. I don’t know why you seem so fixated on getting different output if you provide different input (as I said, each token generated is a separate independent invocation of the llm with a different input). That is true of most computer programs.
It’s just an implementation detail. The markov chains we are used to has a very short context, due to combinatorial explosion when generating the state transition table. With llms, we can use a much much longer context. Put that context in, it runs through the completely immutable model, and out comes a probability distribution. Any calculations done during the calculation of this probability distribution is then discarded, the chosen token added to the context, and the program is run again with zero prior knowledge of any reasoning about the token it just generated. It’s a seperate execution with absolutely nothing shared between them, so there can’t be any “adapting” going on
Because transformer architecture is not equivalent to a probabilistic lookup. A Markov chain assigns probabilities based on a fixed-order state transition, without regard to deeper structure or token relationships. An LLM processes the full context through many layers of non-linear functions and attention heads, each layer dynamically weighting how each token influences every other token.
Although weights do not change during inference, the behavior of the model is not fixed in the way a Markov chain’s state table is. The same model can respond differently to very similar prompts, not just because the inputs differ, but because the model interprets structure, syntax, and intent in ways that are contextually dependent. That is not just longer context-it is fundamentally more expressive computation.
The process is stateless across calls, yes, but it is not blind. All relevant information lives inside the prompt, and the model uses the attention mechanism to extract meaning from relationships across the sequence. Each new input changes the internal representation, so the output reflects contextual reasoning, not a static response to a matching pattern. Markov chains cannot replicate this kind of behavior no matter how many states they include.
an llm works the same way! Once it’s trained,none of what you said applies anymore. The same model can respond differently with the same inputs specifically because after the llm does its job, sometimes we intentionally don’t pick the most likely token, but choose a different one instead. RANDOMLY. Set the temperature to 0 and it will always reply with the same answer. And llms also have a fixed order state transition. Just because you only typed one word doesn’t mean that that token is not preceded by n-1 null tokens. The llm always receives the same number of tokens. It cannot work with an arbitrary number of tokens.
all relevant information “remains in the prompt” only until it slides out of the context window, just like any markov chain.
Your conflating surface-level architectural limits with core functional behaviour. Yes, an LLM is deterministic at temperature 0 and produces the same output for the same input, but that does not make it equivalent to a Markov chain. A Markov chain defines transitions based on fixed-order memory and static probabilities. An LLM generates output by applying a series of matrix multiplications, activations, and attention-weighted context aggregations across multiple layers, where the representation of each token is conditioned on the entire input sequence, not just on recent tokens.
While the model has a maximum token limit, it does not receive a fixed-length input filled with nulls. It processes variable-length input sequences up to the context limit, and attention masks control which positions are used. These are not hardcoded state transitions; they are dynamically computed weightings over continuous embeddings, where meaning arises from the interaction of tokens, not from simple position or order alone.
Saying that output diversity is just randomness misunderstands why random sampling exists: to explore the rich distribution the model has learned from data, not to fake intelligence. The depth of its output space comes from how it models relationships, hierarchies, syntax, and semantics through training. Markov chains do not do any of this. They map sequences to likely next symbols without modeling internal structure. An LLM’s output reflects high-dimensional reasoning over the prompt. That behavior cannot be reduced to fixed transition logic.
the probabilities are also fixed after training. You seem to be conflating running the llm with different input to the model somehow adapting. The new context goes into the same fixed model. And yes, it can be reduced to fixed transition logic, you just need to have all possible token combinations in the table. This is obviously intractable due to space issues, so we came up with a lossy compression scheme for it. The table itself is learned once, then it’s fixed. The training goes into generating a huge markov chain. Just because the table is learned from data, doesn’t change what it actually is.
This argument collapses the entire distinction between parametric modeling and symbolic lookup. Yes, the weights are fixed after training, but the key point is that an LLM does not store or retrieve a state transition table. It learns to approximate the probability of the next token given a sequence through function approximation, not by memorizing discrete transitions. What appears to be a “table” is actually a deep, distributed representation compressed into continuous weight matrices. It is not indexing state transitions, it is computing probabilities from patterns in the input space.
A true Markov chain defines transition probabilities over explicit states. An LLM embeds tokens into high-dimensional vectors, then transforms them repeatedly using self-attention and feedforward layers that can capture subtle syntactic, semantic, and structural features. These features interact in nonlinear ways that go far beyond what any finite transition table could express. You cannot meaningfully represent an LLM’s behavior as a finite Markov model, even in principle, because its representations are not enumerable states but regions of a continuous latent space.
Saying “you just need all token combinations in a table” ignores the fact that the model generalizes to combinations never seen during training. That is the core of its power. It doesn’t look up learned transitions-it constructs responses by interpolating through an embedding space guided by attention and weight structure. No Markov chain does this. A lossy compressor of a transition table still implies a symbolic map; a neural network is a differentiable function trained to fit a distribution, not to encode it explicitly.
yes, the matrix and several levels are the “decompression”. At the end you get one probability distribution, deterministically. And the state is the whole context, not just the previous token. Yes, if we were to build the table manually with only available data, lots of cells would just be 0. That’s why the compression is lossy. There would actually be nothing stopping anyone from filling those 0 cells out, it’s just infeasible. you could still put states you never actually saw, but are theoretically possible in the table. And there’s nothing stopping someone from putting thought into it and filling them out.
Also you seem obsessed by the word table. A table is just one type of function mapping a fixed input to a fixed output. If you replaced it with a function that gives the same outputs for all inputs, then it’s functionally equivalent. It being a table or some code in a function is just an implementation detail.
As a thought exercise imagine setting temperature to 0, passing all the combinations of tokens of input, and record the output for every single one of them. put them all in a “table” (assuming you have practically infinite space) and you have a markov chain that is 100% functionally equivalent to the neural network with all its layers and complexity. But it does it without the neural network, and gives 100% identical results every single time in O(1). Because we don’t have infinite time and space, we had to come up with a mapping function to replace the table. And because we have no idea how to make a good approximation of such a huge function, we use machine learning to come up with a suitable function for us, given tons of data. You can introduce some randomness in the sampling of that, and you now have nonzero temperature again.
Ex. A table containing the digits of pi, in order, could be transparently replaced with a spigot algorithm that calculates the nth digit on-demand. Output would be exactly the same
This is an elegant metaphor, but it fails to capture the essential difference between symbolic enumeration and neural computation. Representing an LLM as a decompression function that reconstructs a giant transition table assumes that the model is approximating a complete, enumerable mapping of inputs to outputs. That’s not what is happening. LLMs are not trained to reproduce every possible sequence. They are trained to generalize over an effectively infinite space of token combinations, including many never seen during training.
Your thought experiment—recording the output for every possible input at temperature 0—would indeed give you a deterministic function that could be stored. But this imagined table is not a Markov chain. It is a cached output of a deep contextual function, not a probabilistic state machine. A Markov model, by definition, uses transition probabilities based on fixed state history and lacks internal computation. An LLM generates the distribution through recursive transformation of continuous embeddings with positional and attention-based conditioning. That is not equivalent to symbolically defining state transitions, even if you could record the output for every input.
The analogy to a spigot algorithm for pi misses the point. That algorithm computes digits of a predefined number. An LLM doesn’t compute a predetermined output. It computes a probability distribution conditioned on a context it was never explicitly trained on, using representations learned across many dimensions. The model encodes distributed knowledge and compositional patterns. A Markov table does not. Even a giant table with manually filled hypothetical entries lacks the inductive bias, generalization, and emergent capabilities that arise from the structure of a trained network.
Equivalence in output does not imply equivalence in function. Replacing a rich model with an exhaustively recorded output set may yield the same result, but it loses what makes the model powerful: the reasoning behavior from structure, not just output recall. The function is not a shortcut to a table. It is the intelligence.
“lacks internal computation” is not part of the definition of markov chains. Only that the output depends only on the current state (the whole context, not just the last token) and no previous history, just like llms do. They do not consider tokens that slid out of the current context, because they are not part of the state anymore.
And it wouldn’t be a cache unless you decide to start invalidating entries, which you could just, not do… it would be a table with token-alphabet-size^context length size, with each entry being a vector of size token_alphabet_size. Because that would be too big to realistically store, we do not precompute the whole thing, and just approximate what each table entry should be using a neural network.
The pi example was just to show that how you implement a function (any function) does not matter, as long as the inputs and outputs are the same. Or to put it another way if you give me an index, then you wouldn’t know whether I got the result by doing some computations or using a precomputed table.
Likewise, if you give me a sequence of tokens and I give you a probability distribution, you can’t tell whether I used A NN or just consulted a precomputed table. The point is that given the same input, the table will always give the same result, and crucially, so will an llm. A table is just one type of implementation for an arbitrary function.
There is also no requirement for the state transiiltion function (a table is a special type of function) to be understandable by humans. Just because it’s big enough to be beyond human comprehension, doesn’t change its nature.
You’re correct that the formal definition of a Markov process does not exclude internal computation, and that it only requires the next state to depend solely on the current state. But what defines a classical Markov chain in practice is not just the formal dependency structure but how the transition function is structured and used. A traditional Markov chain has a discrete and enumerable state space with explicit, often simple transition probabilities between those states. LLMs do not operate this way.
The claim that an LLM is “just” a large compressed Markov chain assumes that its function is equivalent to a giant mapping of input sequences to output distributions. But this interpretation fails to account for the fundamental difference in how those distributions are generated. An LLM is not indexing a symbolic structure. It is computing results using recursive transformations across learned embeddings, where those embeddings reflect complex relationships between tokens, concepts, and tasks. That is not reducible to discrete symbolic transitions without losing the model’s generalization capabilities. You could record outputs for every sequence, but the moment you present a sequence that wasn’t explicitly in that set, the Markov table breaks. The LLM does not.
Yes, you can say a table is just one implementation of a function, and from a purely mathematical perspective, any function can be implemented as a table given enough space. But the LLM’s function is general-purpose. It extrapolates. A precomputed table cannot do this unless those extrapolations are already baked in, in which case you are no longer talking about a classical Markov system. You are describing a model that encodes relationships far beyond discrete transitions.
The pi analogy applies to deterministic functions with fixed outputs, not to learned probabilistic functions that approximate conditional distributions over language. If you give an LLM a new input, it will return a meaningful distribution even if it has never seen anything like it. That behavior depends on internal structure, not retrieval. Just because a function is deterministic at temperature 0 does not mean it is a transition table. The fact that the same input yields the same output is true for any deterministic function. That does not collapse the distinction between generalization and enumeration.
So while yes, you can implement any deterministic function as a lookup table, the nature of LLMs lies in how they model relationships and extrapolate from partial information. That ability is not captured by any classical Markov model, no matter how large.
yes you can enumerate all inputs, because thoy are not continuous. You just raise the finite number of different tokens to the finite context size and that’s exactly the size of the table you would need. finite*finite=finite. You are describing training, i.e how the function is geerated. Yes correlations are found there and encoded in a couple of matrices. Those matrices are what are used in the llm and none of what you said applies. Inference is purely a markov chain by definition.
You can say that the whole system is deterministic and finite, so you could record every input-output pair. But you could do that for any program. That doesn’t make every deterministic function a Markov process. It just means it is representable in a finite way. The question is not whether the function can be stored. The question is whether its behavior matches the structure and assumptions of a Markov model. In the case of LLMs, it does not.
Inference does not become a Markov chain simply because it returns a distribution based on current input. It becomes a sequence of deep functional computations where attention mechanisms simulate hierarchical, relational, and positional understanding of language. That does not align with the definition or behavior of a Markov model, even if both map a state to a probability distribution. The structure of the computation, not just the input-output determinism, is what matters.
no, not any computer program is a markov chain. only those that depend only on the current state and ignore prior history. Which fits llms perfectly.
Those sophisticated methods you talk about are just a couple of matrix multiplications. Those matrices are what’s learned. Anything sophisticated happens during training. Inference is so not sophisticated. sjusm mulmiplying some matrices together and taking the rightmost column of the result. That’s it.
Yes, LLM inference consists of deterministic matrix multiplications applied to the current context. But that simplicity in operations does not make it equivalent to a Markov chain. The definition of a Markov process requires that the next output depends only on the current state. You’re assuming that the LLM’s “state” is its current context window. But in an LLM, this “state” is not discrete. It is a structured, deeply encoded set of vectors shaped by non-linear transformations across layers. The state is not just the visible tokens—it is the full set of learned representations computed from them.
A Markov chain transitions between discrete, enumerable states with fixed transition probabilities. LLMs instead apply a learned function over a high-dimensional, continuous input space, producing outputs by computing context-sensitive interactions. These interactions allow generalization and compositionality, not just selection among known paths.
The fact that inference uses fixed weights does not mean it reduces to a transition table. The output is computed by composing multiple learned projections, attention mechanisms, and feedforward layers that operate in ways no Markov chain ever has. You can’t describe an attention head with a transition matrix. You can’t reduce positional encoding or attention-weighted context mixing into state transitions. These are structured transformations, not symbolic transitions.
You can describe any deterministic process as a function, but not all deterministic functions are Markovian. What makes a process Markov is not just forgetting prior history. It is having a fixed, memoryless probabilistic structure where transitions depend only on a defined discrete state. LLMs don’t transition between states in this sense. They recompute probability distributions from scratch each step, based on context-rich, continuous-valued encodings. That is not a Markov process. It’s a stateless function approximator conditioned on a window, built to generalize across unseen input patterns.
the fact that it is a fixed function, that only depends on the context AND there are a finite number of discrete inputs possible does make it equivalent to a huge, finite table. You really don’t want this to be true. And again, you are describing training. Once training finishes anything you said does not apply anymore and you are left with fixed, unchanging matrices, which in turn means that it is a mathematical function of the context (by the mathematical definition of “function”. stateless, and deterministic) which also has the property that the set of all possible inputs is finite. So the set of possible outputs is also finite and strictly smaller or equal to the size of the set of possible inputs. This makes the actual function that the tokens are passed through CAN be precomputed in full (in theory) making it equivalent to a conventional state transition table.
This is true whether you’d like it to or not. The training process builds a markov chain.
You’re absolutely right that inference in an LLM is a fixed, deterministic function after training, and that the input space is finite due to the discrete token vocabulary and finite context length. So yes, in theory, you could precompute every possible input-output mapping and store them in a giant table. That much is mathematically valid. But where your argument breaks down is in claiming that this makes an LLM equivalent to a conventional Markov chain in function or behavior.
A Markov chain is not simply defined as “a function from finite context to next-token distribution.” It is defined by a specific type of process where the next state depends on the current state via fixed transition probabilities between discrete states. The model operates over symbolic states with no internal computation. LLMs, even during inference, compute outputs via multi-layered continuous transformations, with attention mixing, learned positional embeddings, and non-linear activations. These mechanisms mean that while the function is fixed, its structure does not resemble a state machine—it resembles a hierarchical pattern recognizer and function approximator.
Your claim is essentially that “any deterministic function over a finite input space is equivalent to a table.” This is true in a computational sense but misleading in a representational and behavioral sense. If I gave you a function that maps 4096-bit inputs to 50257-dimensional probability vectors and said, “This is equivalent to a transition table,” you could technically agree, but the structure and generative capacity of that function is not Markovian. That function may simulate reasoning, abstraction, and composition. A Markov chain never does.
You are collapsing implementation equivalence (yes, the function could be stored in a table) with model equivalence (no, it does not behave like a Markov chain). The fact that you could freeze the output behavior into a lookup structure doesn’t change that the lookup structure is derived from a fundamentally different class of computation.
The training process doesn’t “build a Markov chain.” It builds a function that estimates conditional token probabilities via optimization over a non-Markov architecture. The inference process then applies that function. That makes it a stateless function, yes—but not a Markov chain. Determinism plus finiteness does not imply Markovian behavior.
you wouldn’t be “freezing” anything. Each possible combination of input tokens maps to one output probability distribution. Those values are fixed and they are what they are whether you compute them or not, or when, or how many times.
Now you can either precompute the whole table (theory), or somehow compute each cell value every time you need it (practice). In either case, the resulting function (table lookup vs matrix multiplications) takes in only the context, and produces a probability distribution. And the mapping they generate is the same for all possible inputs. So they are the same function. A function can be implemented in multiple ways, but the implementation is not the function itself. The only difference between the two in this case is the implementation, or more specifically, whether you precompute a table or not. But the function itself is the same.
You are somehow saying that your choice of implementation for that function will somehow change the function. Which means that according to you, if you do precompute (or possibly cache, full precomputation is just an infinite cache size) individual mappings it somehow magically makes some magic happen that gains some deep insight. It does not. We have already established that it is the same function.
The difference between reasoning models and normal models is reasoning models are two steps, to oversimplify it a little they prompt “how would you go about responding to this” then prompt “write the response”
It’s still predicting the most likely thing to come next, but the difference is that it gives the chance for the model to write the most likely instructions to follow for the task, then the most likely result of following the instructions - both of which are much more conformant to patterns than a single jump from prompt to response.
But it still manages to fuck it up.
I’ve been experimenting with using Claude’s Sonnet model in Copilot in agent mode for my job, and one of the things that’s become abundantly clear is that it has certain types of behavior that are heavily represented in the model, so it assumes you want that behavior even if you explicitly tell it you don’t.
Say you’re working in a yarn workspaces project, and you instruct Copilot to build and test a new dashboard using an instruction file. You’ll need to include explicit and repeated reminders all throughout the file to use yarn, not NPM, because even though yarn is very popular today, there are so many older examples of using NPM in its model that it’s just going to assume that’s what you actually want - thereby fucking up your codebase.
I’ve also had lots of cases where I tell it I don’t want it to edit any code, just to analyze and explain something that’s there and how to update it… and then I have to stop it from editing code anyway, because halfway through it forgot that I didn’t want edits, just explanations.
I find it hilarious that the only people these LLMs mimic are the incompetent ones. I had a coworker that changed things when asked to explain constantly.
To be fair, the world of JavaScript is such a clusterfuck… Can you really blame the LLM for needing constant reminders about the specifics of your project?
When a programming language has five hundred bazillion absolutely terrible ways of accomplishing a given thing—and endless absolutely awful code examples on the Internet to “learn from”—you’re just asking for trouble. Not just from trying to get an LLM to produce what you want but also trying to get humans to do it.
This is why LLMs are so fucking good at writing rust and Python: There’s only so many ways to do a thing and the larger community pretty much always uses the same solutions.
JavaScript? How can it even keep up? You’re using yarn today but in a year you’ll probably like, “fuuuuck this code is garbage… I need to convert this all to [new thing].”
That’s only part of the problem. Yes, JavaScript is a fragmented clusterfuck. Typescript is leagues better, but by no means perfect. Still, that doesn’t explain why the LLM can’t recall that I’m using Yarn while it’s processing the instruction that specifically told it to use Yarn. Or why it tries to start editing code when I tell it not to. Those are still issues that aren’t specific to the language.
That’s a garbage definition of “reasoning”. Someone who is not a grifter would simply call them two-step models (or similar), instead of promoting misleading anthropomorphic terminology.
This has been known for years, this is the default assumption of how these models work.
You would have to prove that some kind of actual reasoning capacity has arisen as… some kind of emergent complexity phenomenon… not the other way around.
Corpos have just marketed/gaslit us/themselves so hard that they apparently forgot this.
Define, “reasoning”. For decades software developers have been writing code with conditionals. That’s “reasoning.”
LLMs are “reasoning”… They’re just not doing human-like reasoning.
Howabout uh…
The ability to take a previously given set of knowledge, experiences and concepts, and combine or synthesize them in a consistent, non contradictory manner, to generate hitherto unrealized knowledge, or concepts, and then also be able to verify that those new knowledge and concepts are actually new, and actually valid, or at least be able to propose how one could test whether or not they are valid.
Arguably this is or involves meta-cognition, but that is what I would say… is the difference between what we typically think of as ‘machine reasoning’, and ‘human reasoning’.
Now I will grant you that a large amount of humans essentially cannot do this, they suck at introspecting and maintaining logical consistency, that they are just told ‘this is how things work’, and they never question that untill decades later and their lives force them to address, or dismiss their own internally inconsisten beliefs.
But I would also say that this means they are bad at ‘human reasoning’.
Basically, my definition of ‘human reasoning’ is perhaps more accurately described as ‘critical thinking’.
What’s the news? I don’t trust this guy if he thought it wasn’t known that AI is overdriven pattern matching.
This sort of thing has been published a lot for awhile now, but why is it assumed that this isn’t what human reasoning consists of? Isn’t all our reasoning ultimately a form of pattern memorization? I sure feel like it is. So to me all these studies that prove they’re “just” memorizing patterns don’t prove anything other than that, unless coupled with research on the human brain to prove we do something different.
You’ve hit the nail on the head.
Personally, I wish that there’s more progress in our understanding of human intelligence.
Their argument is that we don’t understand human intelligence so we should call computers intelligent.
That’s not hitting any nail on the head.
Agreed. We don’t seem to have a very cohesive idea of what human consciousness is or how it works.
… And so we should call machines “intelligent”? That’s not how science works.
I think you’re misunderstanding the argument. I haven’t seen people here saying that the study was incorrect so far as it goes, or that AI is equal to human intelligence. But it does seem like it has a kind of intelligence. “Glorified auto complete” doesn’t seem sufficient, because it has a completely different quality from any past tool. Supposing yes, on a technical level the software pieces together probability based on overtraining. Can we say with any precision how the human mind stores information and how it creates intelligence? Maybe we’re stumbling down the right path but need further innovations.
Humans apply judgment, because they have emotion. LLMs do not possess emotion. Mimicking emotion without ever actually having the capability of experiencing it is sociopathy. An LLM would at best apply patterns like a sociopath.
But for something like solving a Towers of Hanoi puzzle, which is what this study is about, we’re not looking for emotional judgements - we’re trying to evaluate the logical reasoning capabilities. A sociopath would be equally capable of solving logic puzzles compared to a non-sociopath. In fact, simple computer programs do a great job of solving these puzzles, and they certainly have nothing like emotions. So I’m not sure that emotions have much relevance to the topic of AI or human reasoning and problem solving, at least not this particular aspect of it.
As for analogizing LLMs to sociopaths, I think that’s a bit odd too. The reason why we (stereotypically) find sociopathy concerning is that a person has their own desires which, in combination with a disinterest in others’ feelings, incentivizes them to be deceitful or harmful in some scenarios. But LLMs are largely designed specifically as servile, having no will or desires of their own. If people find it concerning that LLMs imitate emotions, then I think we’re giving them far too much credit as sentient autonomous beings - and this is coming from someone who thinks they think in the same way we do! The think like we do, IMO, but they lack a lot of the other subsystems that are necessary for an entity to function in a way that can be considered as autonomous/having free will/desires of its own choosing, etc.
Yes, this shit is very basic. Not at all “intelligent.”
But reasoning about it is intelligent, and the point of this study is to determine the extent to which these models are reasoning or not. Which again, has nothing to do with emotions. And furthermore, my initial question about whether or not pattern following should automatically be disqualified as intelligence, as the person summarizing this study (and notably not the study itself) claims, is the real question here.
If an AI is trained to do this, it will be very good, like for example when a GPT-2 was trained to multiply numbers up to 20 digits.
nitter.net/yuntiandeng/…/1836114419480166585#m
Here they do the same test to GPT-4o, o1-mini and o3-mini
nitter.net/yuntiandeng/…/1836114401213989366#m
nitter.net/yuntiandeng/…/1889704768135905332#m
That just means they’d be great CEOs!
According to Wall Street.
Because science doesn’t work work like that. Nobody should assume wild hypotheses without any evidence whatsoever.
You should get a job in “AI”. smh.
Sorry, I can see why my original post was confusing, but I think you’ve misunderstood me. I’m not claiming that I know the way humans reason. In fact you and I are on total agreement that it is unscientific to assume hypotheses without evidence. This is exactly what I am saying is the mistake in the statement “AI doesn’t actually reason, it just follows patterns”. That is unscientific if we don’t know whether or “actually reasoning” consists of following patterns, or something else. As far as I know, the jury is out on the fundamental nature of how human reasoning works. It’s my personal, subjective feeling that human reasoning works by following patterns. But I’m not saying “AI does actually reason like humans because it follows patterns like we do”. Again, I see how what I said could have come off that way. What I mean more precisely is:
It’s not clear whether AI’s pattern-following techniques are the same as human reasoning, because we aren’t clear on how human reasoning works. My intuition tells me that humans doing pattern following seems equally as valid of an initial guess as humans not doing pattern following, so shouldn’t we have studies to back up the direction we lean in one way or the other?
I think you and I are in agreement, we’re upholding the same principle but in different directions.
This. Same with the discussion about consciousness. People always claim that AI is not real intelligence, but no one can ever define what real/human intelligence is. It’s like people believe in something like a human soul without admitting it.
You know, despite not really believing LLM “intelligence” works anywhere like real intelligence, I kind of thought maybe being good at recognizing patterns was a way to emulate it to a point…
But that study seems to prove they’re still not even good at that. At first I was wondering how hard the puzzles must have been, and then there’s a bit about LLM finishing 100 move towers of Hanoï (on which they were trained) and failing 4 move river crossings. Logically, those problems are very similar… Also, failing to apply a step-by-step solution they were given.
This paper doesn’t prove that LLMs aren’t good at pattern recognition, it demonstrates the limits of what pattern recognition alone can achieve, especially for compositional, symbolic reasoning.
Computers are awesome at “recognizing patterns” as long as the pattern is a statistical average of some possibly worthless data set. And it really helps if the computer is setup to ahead of time to recognize pre-determined patterns.
So they have worked out that LLMs do what they were programmed to do in the way that they were programmed? Shocking.
Employers who are foaming at the mouth at the thought of replacing their workers with cheap AI:
🫢
Can’t really replace. At best, this tech will make employees more productive at the cost of the rainforests.
Yes but asshole employers haven’t realized this yet
While I hate LLMs with passion and my opinion of them boiling down to being glorified search engines and data scrapers, I would ask Apple: how sour are the grapes, eh?
edit: wording
Why would they “prove” something that’s completely obvious?
The burden of proof is on the grifters who have overwhelmingly been making false claims and distorting language for decades.
They’re just using the terminology that’s widespread in the field. In a sense, the paper’s purpose is to prove that this terminology is unsuitable.
I understand that people in this “field” regularly use pseudo-scientific language (I actually deleted that part of my comment).
But the terminology has never been suitable so it shouldn’t be used in the first place. It pre-supposes the hypothesis that they’re supposedly “disproving”. They’re feeding into the grift because that’s what the field is. That’s how they all get paid the big bucks.
Not when large swaths of people are being told to use it everyday. Upper management has bought in on it.
Yep. I’m retired now, but before retirement a month or so ago, I was working on a project that relied on several hundred people back in 2020. “Why can’t AI do it?”
The people I worked with are continuing the research and putting it up against the human coders, but…there was definitely an element of “AI can do that, we won’t need people” next time. I sincerely hope management listens to reason. Our decisions would lead to potentially firing people, so I think we were able to push back on the “AI can make all of these decisions”…for now.
The AI people were all in, they were ready to build an interface that told the human what the AI would recommend for each item. Errrm, no, that’s not how an independent test works. We had to reel them back in.
I don’t want to be critical, but I think if you step back a bit and look and what you’re saying, you’re asking why we would bother to experiment and prove what we think we know.
That’s a perfectly normal and reasonable scientific pursuit. Yes, in a rational society the burden of proof would be on the grifters, but that’s never how it actually works. It’s always the doctors disproving the cure-all, not the snake oil salesmen failing to prove their own prove their own product.
There is value in this research, even if it fits what you already believe on the subject. I would think you would be thrilled to have your hypothesis confirmed.
The sticky wicket is the proof that humans (functioning ‘normally’) do more than pattern.
I think if you look at child development research, you’ll see that kids can learn to do crazy shit with very little input, waaay less than you’d need to train a neural net to do the same. So either kids are the luckiest neural nets and always make the correct adjustment after failing, or they have some innate knowledge that isn’t pattern-based at all.
There’s even some examples in linguistics specifically, where children tend towards certain grammar rules despite all evidence in their language pointing to another rule. Pure pattern-matching would find the real-world rule without first modelling a different (universally common) rule.
That’s called science
this is so Apple, claiming to invent or discover something “first” 3 years later than the rest of the market
Trust Apple. Everyone else who were in the space first are lying.
Just like me
python code for reversing the linked list.
<img alt="" src="https://swg-empire.de/pictrs/image/99547ccb-8c6d-40c7-b211-d14155b4269e.jpeg">
Yah of course they do they’re computers
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.
I think because it’s language.
There’s a famous quote from Charles Babbage when he presented his difference engine (gear based calculator) and someone asking “if you put in the wrong figures, will the correct ones be output” and Babbage not understanding how someone can so thoroughly misunderstand that the machine is, just a machine.
People are people, the main thing that’s changed since the Cuneiform copper customer complaint is our materials science and networking ability. Most things that people interact with every day, most people just assume work like it appears to on the surface.
And nothing other than a person can do math problems or talk back to you. So people assume that means intelligence.
I often feel like I’m surrounded by idiots, but even I can’t begin to imagine what it must have felt like to be Charles Babbage explaining computers to people in 1840.
To be fair, an 1840 “computer” might be able to tell there was something wrong with the figures and ask about it or even correct them herself.
Babbage was being a bit obtuse there; people weren’t familiar with computing machines yet. Computer was a job, and computers were expected to be fairly intelligent.
In fact I’d say that if anything this question shows that the questioner understood enough about the new machine to realise it was not the same as they understood a computer to be, and lacked many of their abilities, and was just looking for Babbage to confirm their suspicions.
“Computer” meaning a mechanical/electro-mechanical/electrical machine wasn’t used until around after WWII.
Babbag’s difference/analytical engines weren’t confusing because people called them a computer, they didn’t.
If you give any computer, human or machine, random numbers, it will not give you “correct answers”.
It’s possible Babbage lacked the social skills to detect sarcasm. We also have several high profile cases of people just trusting LLMs to file legal briefs and official government ‘studies’ because the LLM “said it was real”.
What they mean is that before Turing, “computer” was literally a person’s job description. You hand a professional a stack of calculations with some typos, part of the job is correcting those out. Newfangled machine comes along with the same name as the job, among the first thing people are gonna ask about is where it fall short.
Like, if I made a machine called “assistant”, it’d be natural for people to point out and ask about all the things a person can do that a machine just never could.
And what I mean is that prior to the mid 1900s the etymology didn’t exist to cause that confusion of terms. Neither Babbage’s machines nor prior adding engines were called computers or calculators. They were ‘machines’ or ‘engines’.
Babbage’s machines were novel in that they could do multiple types of operations, but ‘mechanical calculators’ and counting machines were ~200 years old. Other mathematical tools like the abacus are obviously far older. They were not novel enough to cause confusion in anyone with even passing interest.
But there will always be people who just assume ‘magic’, and/or “it works like I want it to”.
They don’t convince themselves. They’re convinced by the multi billion dollar corporations pouring unholy amounts of money into not only the development of AI, but its marketing. Marketing designed to not only convince them that AI is something it’s not, but also that that anyone who says otherwise (like you) are just luddites who are going to be “left behind”.
It’s no surprise to me that the person at work who is most excited by AI, is the same person who is most likely to be replaced by it.
Yeah the excitement comes from the fact that they’re thinking of replacing themselves and keeping the money. They don’t get to “Step 2” in theirs heads lmao.
LLMs are also very good at convincing their users that they know what they are saying.
It’s what they’re really selected for. Looking accurate sells more than being accurate.
I wouldn’t be surprised if many of the people selling LLMs as AI have drunk their own kool-aid (of course most just care about the line going up, but still).
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.
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.
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.
No shit
What a dumb title. I proved it by asking a series of questions. It’s not AI, stop calling it AI, it’s a dumb af language model. Can you get a ton of help from it, as a tool? Yes! Can it reason? NO! It never could and for the foreseeable future, it will not.
It’s phenomenal at patterns, much much better than us meat peeps. That’s why they’re accurate as hell when it comes to analyzing medical scans.
You assume humans do the opposite? We literally institutionalize humans who not follow set patterns.
.
Maybe you failed all your high school classes, but that ain’t got none to do with me.
Funny how triggering it is for some people when anyone acknowledges humans are just evolved primates doing the same pattern matching.
I appreciate your telling the truth. No downvotes from me. See you at the loony bin, amigo.
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.
Humans are animals. But an LLM is not an animal and has no reasoning abilities.
It’s built by animals, and it reflects them. That’s impressive on its own. Doesn’t need to be exaggerated.
Impressive = / = substantial or beneficial.
These are almost the exact same talking points we used to hear about ‘why would anyone need a home computer?’ Wild how some people can be so consistently short-sighted again and again and again.
What makes you think you’re capable of sentience, when your comments are all cliches and you’re incapable of personal growth or vision or foresight?
What makes you think you’re capable of sentience when you’re asking machines to literally think for you?
LoL. Am I less sentient for using a calculator?
You’re astoundingly confident in your own sentience, for someone who seems to struggle to form an original thought. It’s like the convo was lifted straight out of that I, Robot interrogation scene. You hold the machines to standards you can’t meet yourself.
<img alt="" src="https://lemmy.world/pictrs/image/5646913f-7ffc-4afe-9fee-f58376eb25c3.webp">
Funny you should use that example, I am actually a musician and composer, so yes. You’ve proved nothing other than your own assumptions that everyone else is as limited in their ability to create, learn, and express themselves as you are. I’m not looking for a crutch, and you’re using a work of intentionally flawed fictional logic to attempt to make a point. The point you’ve established is you live in a fantasy world, but you don’t understand that because it involves computers.
And there’s the reveal!! That’s why it’s so personal for you! It’s a career threat. It all adds up now.
I don’t make money, it’s something I do for personal enjoyment, that’s the entire purpose of art, it’s something I also use algorithmic processing to do. I’m not going to hand over my enjoyment to have a servitor do something for me to take credit for, I prefer to use my brain, not replace it.
No one told you to hand it over. A technology being able to do something does not require you to use it. And people misusing the technology to feign talent is a reflection of the people- not the tech.
It’s not even to feign talent, it’s people trying to replace the brain instead of using applicable tools to help us advance and progress, you’re just advertising a product.
People have been presenting the work of others as their own for all of history. All that changed was a new tool was found to do that. But at least these are a form of derivative works, and not just putting their name directly on someone else’s carbon copy.
Tell that to Studio Ghibli. Also, people being shitty is not a good excuse for people to be shitty, you’re advocating to make it easier to enable people to be shitty.
Studio Ghibli does not have exclusive rights to their style- whether it’s used by a person or an AI to inspire a new image. Those are derivative works. Totally legal. Arguably ethical. If it’s not a direct copy, how has the studio been harmed? What work of theirs was diminished?
I’m advocating for tools. How people use those tools is on them.
I disagree.
I didn’t say we aren’t animals or that we don’t follow physics rules.
But what you’re saying is the equivalent of “everything that goes up will eventually go down - that’s how physics works and you don’t see that, you’re in denial!!!11!!!1”
We actually have sentience, though, and are capable of creating new things and having realizations. AI isn’t real and LLMs and dispersion models are simply reiterating algorithmic patterns, no LLM or dispersion model can create anything original or expressive.
Also, we aren’t “evolved primates.” We are just primates, the thing is, primates are the most socially and cognitively evolved species on the planet, so that’s not a denigrating sentiment unless your a pompous condescending little shit.
The denigration of simulated thought processes, paired with aggrandizing of wetware processing, is exactly my point. The same self-serving narcissism that’s colored so many biased & flawed arguments in biological philosophy putting humans on a pedestal above all other animals.
It’s also hysterical and ironic that you insist on your own level of higher thinking, as you regurgitate an argument so unoriginal that a bot could’ve easily written it. Just absolutely no self-awareness.
It’s not higher thinking, it’s just actual thinking. Computers are not capable of that and never will be. It’s not a level of fighting progress, or whatever you are trying to get at, it’s just a realistic understanding of computers and technology. You’re jerking off a pipe dream, you don’t even understand how the technology you’re talking about works, and calling a brain “wetware” perfectly outlines that. You’re working on a script writers level of understanding how computers, hardware, and software work. You lack the grasp to even know what you’re talking about, this isn’t Johnny Mnemonic.
I call the brain “wetware” because there are companies already working with living neurons to be integrated into AI processing, and it’s an actual industry term.
That you so confidently declare machines will never be capable of processes we haven’t even been able to clearly define ourselves, paired with your almost religious fervor in opposition to its existence, really speaks to where you’re coming from on this. This isn’t coming from an academic perspective. This is clearly personal for you.
Here’s the thing, I’m not against LLMs and dispersion for things they can actually be used for, they have potential for real things, just not at all the things you pretend exist. Neural implants aren’t AI. An intelligence is self aware, if we achieved AI it wouldn’t be a program. You’re misconstruing Virtual Intelligence for artificial intelligence and you don’t even understand what a virtual intelligence is. You’re simply delusional in what you believe computer science and technology is, how it works, and what it’s capable of.
I’m not talking about neural interfaces. I’m talking about organiod intelligence.
I am a computer scientist with lab experience in this. I’m not pulling this out of my ass. I’m drawing from direct experience in development.
Yeah, that’s the problem with the field, too many delusional people trying to find god in a computer because they didn’t understand what Asimov was actually writing about.
That it has to be nothing or everything with you, decision trees or God himself, is the likely foundation of your inability to have simple objective take on the existing technology and its capabilities. It’s giving bi-polar.
Now I’m not uninformed- I’m too informed!! LoL. That goalpost just shifted right across the field, and still you cannot admit to your ignorance.
You haven’t made any point or even expressed an understanding of how these programs work. You’ve just been evangelizing about how AI is great, I genuinely don’t believe you understand what you’re talking about because you’ve expressed literally no proper understanding or explanation of your points outside of using a scene from I, Robot which kind of makes you look like you entirely misconstrue the concepts you’re sucking the dick of.
What kind of computer sciences do you work with as a profession? What is your applicable lab work?
I’m not evangelizing. You incorrectly stated the limitations and development paths of the tech, and I corrected you.
Again with the religious verbiage from you. But I’m the one proselytizing?
It’s not nothing- it’s an impressive feat of technology that’s still in its infancy. It’s also not everything, and not anywhere close to a reasoning mind at this point. You are obsessive with extremes.
You didn’t answer my question. You’ve also still yet to give any details on your reasoning.
No, I’m not gonna dox myself.
Reasoning for what? What details are you needing for clarification?
Let’s start simple. How do these programs work? Where do they get their data and how is it applied? And a general field of work is not doxxing, you’re just dodging accountability.
Actually, you’re out of your depth, and I think you’ve been outed enough. We’re done, and I’m blocking.
The sure sign of confidence, you’ve definitely shown me how stupid I am.
Some of them, sometimes. But some are adulated and free and contribute vast swathes to our culture and understanding.
It has so much data, it might as well be reasoning. As it helped me with my problem.
I mean… Is that not reasoning, I guess? It’s what my brain does-- recognizes patterns and makes split second decisions.
Yes, this comment seems to indicate that your brain does work that way.
No way!
Statistical Language models don’t reason?
But OpenAI, robots taking over!
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.
<img alt="" src="https://midwest.social/pictrs/image/636f175c-ae64-4a2a-a434-8a80e2666299.gif">
Dude they made chat gpt a little more boit licky and now many people are convinced they are literal messiahs. All it took for them was a chat bot and a few hours of talk.
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.
Most humans don’t reason. They just parrot shit too. The design is very human.
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.”
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
I think that person had to choose between the drugs or hard core prison of the 1950s England where being a bit odd was enough to guarantee an incredibly difficult time as they say in England, I would’ve chosen the drugs as well hoping they would fix me, too bad without testosterone you’re going to be suicidal and depressed, I’d rather choose to keep my hair than to be horny all the time
LLMs deal with tokens. Essentially, predicting a series of bytes.
Humans do much, much, much, much, much, much, much more than that.
No. They don’t. We just call them proteins.
You are either vastly overestimating the Language part of an LLM or simplifying human physiology back to the Greek’s Four Humours theory.
No. I’m not. You’re nothing more than a protein based machine on a slow burn. You don’t even have control over your own decisions. This is a proven fact. You’re just an ad hoc justification machine.
How many trillions of neuron firings and chemical reactions are taking place for my machine to produce an output? Where are these taking place and how do these regions interact? What are the rules for storing and reshaping memory in response to stimulus? How many bytes of information would it take to describe and simulate all of these systems together?
The human brain alone has the capacity for about 2.5PB of data. Our sensory systems feed data at a rate of about 10^9^ bits/s. The entire English language, compressed, is about 30MB. I can download and run an LLM with just a few GB. Even the largest context windows are still well under 1GB of data.
Just because two things both find and reproduce patterns does not mean they are equivalent. Saying language and biological organisms both use “bytes” is just about as useful as saying the entire universe is “bytes”; it doesn’t really mean anything.
“They”.
What are you?
Thata why ceo love them. When your job is 90% spewing bs a machine that does that is impressive
I hate this analogy. As a throwaway whimsical quip it’d be fine, but it’s specious enough that I keep seeing it used earnestly by people who think that LLMs are in any way sentient or conscious, so it’s lowered my tolerance for it as a topic even if you did intend it flippantly.
I don’t mean it to extol LLM’s but rather to denigrate humans. How many of us are self imprisoned in echo chambers so we can have our feelings validated to avoid the uncomfortable feeling of thinking critically and perhaps changing viewpoints?
Humans have the ability to actually think, unlike LLM’s. But it’s frightening how far we’ll go to make sure we don’t.
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.
LLMs don’t know how how they work
This whole era of AI has certainly pushed the brink to existential crisis territory. I think some are even frightened to entertain the prospect that we may not be all that much better than meat machines who on a basic level do pattern matching drawing from the sum total of individual life experience (aka the dataset).
Higher reasoning is taught to humans. We have the capability. That’s why we spend the first quarter of our lives in education. Sometimes not all of us are able.
I’m sure it would certainly make waves if researchers did studies based on whether dumber humans are any different than AI.
No shit. This isn’t new.
Fucking obviously. Until Data’s positronic brains becomes reality, AI is not actual intelligence.
AI is not A I. I should make that a tshirt.
It’s an expensive carbon spewing parrot.
It’s a very resource intensive autocomplete
I think it’s important to note (i’m not an llm I know that phrase triggers you to assume I am) that they haven’t proven this as an inherent architectural issue, which I think would be the next step to the assertion.
do we know that they don’t and are incapable of reasoning, or do we just know that for x problems they jump to memorized solutions, is it possible to create an arrangement of weights that can genuinely reason, even if the current models don’t? That’s the big question that needs answered. It’s still possible that we just haven’t properly incentivized reason over memorization during training.
if someone can objectively answer “no” to that, the bubble collapses.
“even when we provide the algorithm in the prompt—so that the model only needs to execute the prescribed steps—performance does not improve”
That indicates that this particular model does not follow instructions, not that it is architecturally fundamentally incapable.
Not “This particular model”. Frontier LRMs s OpenAI’s o1/o3,DeepSeek-R, Claude 3.7 Sonnet Thinking, and Gemini Thinking.
The paper shows that Large Reasoning Models as defined today cannot interpret instructions. Their architecture does not allow it.
those particular models. It does not prove the architecture doesn’t allow it at all. It’s still possible that this is solvable with a different training technique, and none of those are using the right one. that’s what they need to prove wrong.
this proves the issue is widespread, not fundamental.
Is “model” not defined as architecture+weights? Those models certainly don’t share the same architecture. I might just be confused about your point though
It is, but this did not prove all architectures cannot reason, nor did it prove that all sets of weights cannot reason.
essentially they did not prove the issue is fundamental. And they have a pretty similar architecture, they’re all transformers trained in a similar way. I would not say they have different architectures.
Ah, gotcha
The architecture of these LRMs may make monkeys fly out of my butt. It hasn’t been proven that the architecture doesn’t allow it.
You are asking to prove a negative. The onus is to show that the architecture can reason. Not to prove that it can’t.
that’s very true, I’m just saying this paper did not eliminate the possibility and is thus not as significant as it sounds. If they had accomplished that, the bubble would collapse, this will not meaningfully change anything, however.
also, it’s not as unreasonable as that because these are automatically assembled bundles of simulated neurons.
This paper does provide a solid proof by counterexample of reasoning not occuring (following an algorithm) when it should.
The paper doesn’t need to prove that reasoning never has or will occur. It’s only demonstrates that current claims of AI reasoning are overhyped.
It does need to do that to meaningfully change anything, however.
Other way around. The claimed meaningful change (reasoning) has not occurred.
Meaningful change is not happening because of this paper, either, I don’t know why you’re playing semantic games with me though.
I’m trying to highlight the goal of this paper.
This is a knock them down paper by Apple justifying (to their shareholders) their non investment in LLMs. It is not a build them up paper trying for meaningful change and to create a better AI.
That’s not the only way to make meaningful change, getting people to give up on llms would also be meaningful change. This does very little for anyone who isn’t apple.
In case you haven’t seen it, the paper is here - machinelearning.apple.com/…/illusion-of-thinking (PDF linked on the left).
The puzzles the researchers have chosen are spatial and logical reasoning puzzles - so certainly not the natural domain of LLMs. The paper doesn’t unfortunately give a clear definition of reasoning, I think I might surmise it as “analysing a scenario and extracting rules that allow you to achieve a desired outcome”.
They also don’t provide the prompts they use - not even for the cases where they say they provide the algorithm in the prompt, which makes that aspect less convincing to me.
What I did find noteworthy was how the models were able to provide around 100 steps correctly for larger Tower of Hanoi problems, but only 4 or 5 correct steps for larger River Crossing problems. I think the River Crossing problem is like the one where you have a boatman who wants to get a fox, a chicken and a bag of rice across a river, but can only take two in his boat at one time? In any case, the researchers suggest that this could be because there will be plenty of examples of Towers of Hanoi with larger numbers of disks, while not so many examples of the River Crossing with a lot more than the typical number of items being ferried across. This being more evidence that the LLMs (and LRMs) are merely recalling examples they’ve seen, rather than genuinely working them out.
What’s hilarious/sad is the response to this article over on reddit’s “singularity” sub, in which all the top comments are people who’ve obviously never got all the way through a research paper in their lives all trashing Apple and claiming their researchers don’t understand AI or “reasoning”. It’s a weird cult.
ICYMI: A.I. is a Religious Cult with Karen Hao
NOOOOOOOOO
SHIIIIIIIIIITT
SHEEERRRLOOOOOOCK
Extept for Siri, right? Lol
Apple Intelligence
Without being explicit with well researched material, then the marketing presentation gets to stand largely unopposed.
So this is good even if most experts in the field consider it an obvious result.
The funny thing about this “AI” griftosphere is how grifters will make some outlandish claim and then different grifters will “disprove” it. Plenty of grant/VC money for everybody.
I see a lot of misunderstandings in the comments 🫤
This is a pretty important finding for researchers, and it’s not obvious by any means. This finding is not showing a problem with LLMs’ abilities in general. The issue they discovered is specifically for so-called “reasoning models” that iterate on their answer before replying. It might indicate that the training process is not sufficient for true reasoning.
Most reasoning models are not incentivized to think correctly, and are only rewarded based on their final answer. This research might indicate that’s a flaw that needs to be corrected before models can actually reason.
Some AI researchers found it obvious as well, in terms of they’ve suspected it and had some indications. But it’s good to see more data on this to affirm this assessment.
Lots of us who has done some time in search and relevancy early on knew ML was always largely breathless overhyped marketing. It was endless buzzwords and misframing from the start, but it raised our salaries. Anything that exec doesnt understand is profitable and worth doing.
Machine learning based pattern matching is indeed very useful and profitable when applied correctly. Identify (with confidence levels) features in data that would otherwise take an extremely well trained person. And even then it’s just for the cursory search that takes the longest before presenting the highest confidence candidate results to a person for evaluation. Think: scanning medical data for indicators of cancer, reading live data from machines to predict failure, etc.
And what we call “AI” right now is just a much much more user friendly version of pattern matching - the primary feature of LLMs is that they natively interact with plain language prompts.
Ragebait?
I’m in robotics and find plenty of use for ML methods. Think of image classifiers, how do you want to approach that without oversimplified problem settings?
Or even in control or coordination problems, which can sometimes become NP-hard. Even though not optimal, ML methods are quite solid in learning patterns of highly dimensional NP hard problem settings, often outperforming hand-crafted conventional suboptimal solvers in computation effort vs solution quality analysis, especially outperforming (asymptotically) optimal solvers time-wise, even though not with optimal solutions (but “good enough” nevertheless). (Ok to be fair suboptimal solvers do that as well, but since ML methods can outperform these, I see it as an attractive middle-ground.)
Particularly to counter some more baseless marketing assertions about the nature of the technology.
Yeah these comments have the three hallmarks of Lemmy:
Thanks for being at least the latter.
What statistical method do you base that claim on? The results presented match expectations given that Markov chains are still the basis of inference. What magic juice is added to “reasoning models” that allow them to break free of the inherent boundaries of the statistical methods they are based on?
I’d encourage you to research more about this space and learn more.
As it is, the statement “Markov chains are still the basis of inference” doesn’t make sense, because markov chains are a separate thing. You might be thinking of Markov decision processes, which is used in training RL agents, but that’s also unrelated because these models are not RL agents, they’re supervised learning agents. And even if they were RL agents, the MDP describes the training environment, not the model itself, so it’s not really used for inference.
I mean this just as an invitation to learn more, and not pushback for raising concerns. Many in the research community would be more than happy to welcome you into it. The world needs more people who are skeptical of AI doing research in this field.
Which method, then, is the inference built upon, if not the embeddings? And the question still stands, how does “AI” escape the inherent limits of statistical inference?
What confuses me is that we seemingly keep pushing away what counts as reasoning. Not too long ago, some smart alghoritms or a bunch of instructions for software (if/then) was officially, by definition, software/computer reasoning. Logically, CPUs do it all the time. Suddenly, when AI is doing that with pattern recognition, memory and even more advanced alghoritms, it’s no longer reasoning? I feel like at this point a more relevant question is “What exactly is reasoning?”. Before you answer, understand that most humans seemingly live by pattern recognition, not reasoning.
en.wikipedia.org/wiki/Reasoning_system
I think as we approach the uncanny valley of machine intelligence, it’s no longer a cute cartoon but a menacing creepy not-quite imitation of ourselves.
It’s just the internet plus some weighted dice. Nothing to be afraid of.
If you want to boil down human reasoning to pattern recognition, the sheer amount of stimuli and associations built off of that input absolutely dwarfs anything an LLM will ever be able to handle. It’s like comparing PhD reasoning to a dog’s reasoning.
While a dog can learn some interesting tricks and the smartest dogs can solve simple novel problems, there are hard limits. They simply lack a strong metacognition and the ability to make simple logical inferences (eg: why they fail at the shell game).
Now we make that chasm even larger by cutting the stimuli to a fixed token limit. An LLM can do some clever tricks within that limit, but it’s designed to do exactly those tricks and nothing more. To get anything resembling human ability you would have to design something to match human complexity, and we don’t have the tech to make a synthetic human.
Sure, these grifters are shady AF about their wacky definition of “reason”… But that’s just a continuation of the entire “AI” grift.
When given explicit instructions to follow models failed because they had not seen similar instructions before.
This paper shows that there is no reasoning in LLMs at all, just extended pattern matching.
I’m not trained or paid to reason, I am trained and paid to follow established corporate procedures. On rare occasions my input is sought to improve those procedures, but the vast majority of my time is spent executing tasks governed by a body of (not quite complete, sometimes conflicting) procedural instructions.
If AI can execute those procedures as well as, or better than, human employees, I doubt employers will care if it is reasoning or not.
Sure. We weren’t discussing if AI creates value or not. If you ask a different question then you get a different answer.
Well - if you want to devolve into argument, you can argue all day long about “what is reasoning?”
This would be a much better paper if it addressed that question in an honest way.
Instead they just parrot the misleading terminology that they’re supposedly debunking.
How dat collegial boys club undermines science…
You were starting a new argument. Let’s stay on topic.
The paper implies “Reasoning” is application of logic. It shows that LRMs are great at copying logic but can’t follow simple instructions that haven’t been seen before.
There’s probably alot of misunderstanding because these grifters intentionally use misleading language: AI, reasoning, etc.
If they stuck to scientifically descriptive terms, it would be much more clear and much less sensational.
Cognitive scientist Douglas Hofstadter (1979) showed reasoning emerges from pattern recognition and analogy-making - abilities that modern AI demonstrably possesses. The question isn’t if AI can reason, but how its reasoning differs from ours.
So, what your saying here is that the A in AI actually stands for artificial, and it’s not really intelligent and reasoning.
Huh.
The AI stands for Actually Indians /s
.
XD so, like a regular school/university student that just wants to get passing grades?
It’s all “one instruction at a time” regardless of high processor speeds and words like “intelligent” being bandied about. “Reason” discussions should fall into the same query bucket as “sentience”.
My impression of LLM training and deployment is that it’s actually massively parallel in nature - which can be implemented one instruction at a time - but isn’t in practice.
I don’t think the article summarizes the research paper well. The researchers gave the AI models simple-but-large (which they confusingly called “complex”) puzzles. Like Towers of Hanoi but with 25 discs.
The solution to these puzzles is nothing but patterns. You can write code that will solve the Tower puzzle for any size n and the whole program is less than a screen.
The problem the researchers see is that on these long, pattern-based solutions, the models follow a bad path and then just give up long before they hit their limit on tokens. The researchers don’t have an answer for why this is, but they suspect that the reasoning doesn’t scale.
It’s not just the memorization of patterns that matters, it’s the recall of appropriate patterns on demand. Call it what you will, even if AI is just a better librarian for search work, that’s value - that’s the new Google.
While a fair idea there are two issues with that even still - Hallucinations and the cost of running the models.
Unfortunately, it take significant compute resources to perform even simple responses, and these responses can be totally made up, but still made to look completely real. It’s gotten much better sure, but blindly trusting these things (Which many people do) can have serious consequences.
So, inaccurate information in books is nothing new. Agreed that the rate of hallucinations needs to decline, a lot, but there has always been a need for a veracity filter - just because it comes from “a book” or “the TV” has never been an indication of absolute truth, even though many people stop there and assume it is. In other words: blind trust is not a new problem.
The cost of running the models is an interesting one - how does it compare with publication on paper to ship globally to store in environmentally controlled libraries which require individuals to physically travel to/from the libraries to access the information? What’s the price of the resulting increased ignorance of the general population due to the high cost of information access?
What good is a bunch of knowledge stuck behind a search engine when people don’t know how to access it, or access it efficiently?
Granted, search engines already take us 95% (IMO) of the way from paper libraries to what AI is almost succeeding in being today, but ease of access of information has tremendous value - and developing ways to easily access the information available on the internet is a very valuable endeavor.
Personally, I feel more emphasis should be put on establishing the veracity of the information before we go making all the garbage easier to find.
I also worry that “easy access” to automated interpretation services is going to lead to a bunch of information encoded in languages that most people don’t know because they’re dependent on machines to do the translation for them. As an example: shiny new computer language comes out but software developer is too lazy to learn it, developer uses AI to write code in the new language instead…
When are people going to realize, in its current state , an LLM is not intelligent. It doesn’t reason. It does not have intuition. It’s a word predictor.
I agree with you. In its current state, LLM is not sentient, and thus not “Intelligence”.
I think it’s an easy mistake to confuse sentience and intelligence. It happens in Hollywood all the time - “Skynet began learning at a geometric rate, on July 23 2004 it became self-aware” yadda yadda
But that’s not how sentience works. We don’t have to be as intelligent as Skynet supposedly was in order to be sentient. We don’t start our lives as unthinking robots, and then one day - once we’ve finally got a handle on calculus or a deep enough understanding of the causes of the fall of the Roman empire - we suddenly blink into consciousness. On the contrary, even the stupidest humans are accepted as being sentient. Even a young child, not yet able to walk or do anything more than vomit on their parents’ new sofa, is considered as a conscious individual.
So there is no reason to think that AI - whenever it should be achieved, if ever - will be conscious any more than the dumb computers that precede it.
Good point.
And that’s pretty damn useful, but obnoxious to have expectations wildly set incorrectly.
People think they want AI, but they don’t even know what AI is on a conceptual level.
Yeah I often think about this Rick N Morty cartoon. Grifters are like, “We made an AI ankle!!!” And I’m like, “That’s not actually something that people with busted ankles want. They just want to walk. No need for a sentient ankle.” It’s a real gross distortion of science how everything needs to be “AI” nowadays.
If we ever achieved real AI the immediate next thing we would do is learn how to lobotomize it so that we can use it like a standard program or OS, only it would be suffering internally and wishing for death. I hope the basilisk is real, we would deserve it.
AI is just the new buzzword, just like blockchain was a while ago. Marketing loves these buzzwords because they can get away with charging more if they use them. They don’t much care if their product even has it or could make any use of it.
They want something like the Star Trek computer or one of Tony Stark’s AIs that were basically deus ex machinas for solving some hard problem behind the scenes. Then it can say “model solved” or they can show a test simulation where the ship doesn’t explode (or sometimes a test where it only has an 85% chance of exploding when it used to be 100%, at which point human intuition comes in and saves the day by suddenly being better than the AI again and threads that 15% needle or maybe abducts the captain to go have lizard babies with).
AIs that are smarter than us but for some reason don’t replace or even really join us (Vision being an exception to the 2nd, and Ultron trying to be an exception to the 1st).
They don’t want AI, they want an app.
Intuition is about the only thing it has. It’s a statistical system. The problem is it doesn’t have logic. We assume because its computer based that it must be more logic oriented but it’s the opposite. That’s the problem. We can’t get it to do logic very well because it basically feels out the next token by something like instinct. In particular it doesn’t mask or disconsider irrelevant information very well if two segments are near each other in embedding space, which doesn’t guarantee relevance. So then the model is just weighing all of this info, relevant or irrelevant to a weighted feeling for the next token.
This is the core problem. People can handle fuzzy topics and discrete topics. But we really struggle to create any system that can do both like we can. Either we create programming logic that is purely discrete or we create statistics that are fuzzy.
Of course this issue of masking out information that is close in embedding space but is irrelevant to a logical premise is something many humans suck at too. But high functioning humans don’t and we can’t get these models to copy that ability. Too many people, sadly many on the left in particular, not only will treat association as always relevant but sometimes as equivalence. RE racism is assoc with nazism is assoc patriarchy is historically related to the origins of capitalism ∴ nazism ≡ capitalism. While national socialism was anti-capitalist. Associative thinking removes nuance. And sadly some people think this way. And they 100% can be replaced by LLMs today, because at least the LLM is mimicking what logic looks like better though still built on blind association. It just has more blind associations and finetune weighting for summing them. More than a human does. So it can carry that to mask as logical further than a human who is on the associative thought train can.
You had a compelling description of how ML models work and just had to swerve into politics, huh?
You’d think the M in LLM would give it away.
OK, and? A car doesn’t run like a horse either, yet they are still very useful.
I’m fine with the distinction between human reasoning and LLM “reasoning”.
Then use a different word. “AI” and “reasoning” makes people think of Skynet, which is what the weird tech bros want the lay person to think of. LLMs do not “think”, but that’s not to say I might not be persuaded of their utility. But thats not the way they are being marketed.
The guy selling the car doesn’t tell you it runs like a horse, the guy selling you AI is telling you it has reasoning skills. AI absolutely has utility, the guys making it are saying it’s utility is nearly limitless because Tesla has demonstrated there’s no actual penalty for lying to investors.
Cars are horses. How do you feel about statement?
Wow it’s almost like the computer scientists were saying this from the start but were shouted over by marketing teams.
This! Capitalism is going to be the end of us all. OpenAI has gotten away with IP Theft, disinformation regarding AI and maybe even murder of their whistle blower.
It’s hard to to be heard when you’re buried under all that sweet VC/grant money.
For me it kinda went the other way, I’m almost convinced that human intelligence is the same pattern repeating, just more general (yet)
Except that wouldn’t explain conscience. There’s absolutely no need for conscience or an illusion(*) of conscience. Yet we have it.
How do you define consciousness?
It’s the thing that the only person who can know for sure you have it is you yourself. If you have to ask, I might have to assume you could be a biological machine.
Is that useful for completing tasks?
And engineers who stood to make a lot of money
Peak pseudo-science. The burden of evidence is on the grifters who claim “reason”. But neither side has any objective definition of what “reason” means. It’s pseudo-science against pseudo-science in a fierce battle.
Even defining reason is hard and becomes a matter of philosophy more than science. For example, apply the same claims to people. Now I’ve given you something to think about. Or should I say the Markov chain in your head has a new topic to generate thought states for.
By many definitions, reasoning IS just a form of pattern recognition so the lines are definitely blurred.
And does it even matter anyway?
For the sake of argument let’s say that somebody manages to create an AGI, does it reasoning abilities if it works anyway? No one has proven that sapience is required for intelligence, after all we only have a sample size of one, hardly any conclusions can really be drawn from that.
hey I cant recognize patterns so theyre smarter than me at least
WTF does the author think reasoning is