Today's Large Language Models are Essentially BS Machines
Today's Large Language Models are Essentially BS Machines
You're viewing a single thread.
And everyone in tech who has worked on ML before collectively says "yeah that's what we've been trying to tell you". Don't get me wrong, LLMs are a huge leap, but god did it show how greedy corporations are, just immediately jumping to "how quick can we lay people off?". The tech is not to that spec. Yet. It will get there, but goddamn do we need to be demanding some regulations now
The tech is not to that spec. Yet.
I'm not sure it will. At least, not this tech, not this approach to the problem. From my understanding there's fundamentally no comprehension; it's not bugged, broken, or incomplete, it's just not there... it's missing from the design.
We don't know that for sure yet, we saw a lot of emergent intelligent properties appear as we scaled up, and we're nowhere near done scaling LLM's, I'm not saying it will be solved, just that we don't know one way or the other yet.
LLMs are fundamentally different from human consciousness. It isn't a problem of scale, but kind.
They are like your phone's autocomplete, but very very good. But there's no level of "very good" for autocomplete that makes it a human, or will give it sentience, or allow it to understand the words it is suggesting. It simply returns the next most-likely word in a response.
If we want computerized intelligence, LLMs are a dead end. They might be a good way for that intelligence to speak pretty sentences to us, but they will never be that themselves.
You're guessing, you don't actually know that for sure, it seems intuitively correct, but we simply do not know enough about cognition to make that assumption.
Perhaps our ability to reason exclusively comes from our ability to predict, and by scaling up the ability to predict, we become more and more able to reason.
These are guesses, all we have now are guesses, you can say "it doesn't reason" and "it's just autocorrect" all you want, but if that were the case why did scaling it up eventually enable it to perform basic math? Why did scaling it up improve its ability to problemsolve significantly (gpt3 vs gpt4), there's so many unknowns in this field, to just say "nah, can't be, it works differently from us" doesn't mean it can't do the same things as us given enough scale.
I'm not guessing. When I say it's a difference of kind, I really did mean that. There is no cognition here; and we know enough about cognition to say that LLMs are not performing anything like it.
Believing LLMs will eventually perform cognition with enough hardware is like saying, "if we throw enough hardware at a calculator, it will eventually become alive." Even if you throw all the hardware in the world at it, there is no emergent property of a calculator that would create sentience. So too LLMs, which really are just calculators that can speak English. But just like calculators they have no conception of what English is and they do not think in any way, and never will.
I’m not guessing. When I say it’s a difference of kind, I really did mean that. There is no cognition here; and we know enough about cognition to say that LLMs are not performing anything like it.
We do not know that, I challenge you to find a source for that, in fact, i've seen sources showing the opposite, they seem to reason in tokens, for example, LLM's perform significantly better at tasks when asked to give a step by step reasoned explanation, this indicates that they are doing a form of reasoning, and their reasoning is limited by what I have no better term for than laziness.
https://blog.research.google/2022/05/language-models-perform-reasoning-via.html
It is your responsibility to prove your assertion that if we just throw enough hardware at LLMs they will suddenly become alive in any recognizable sense, not mine to prove you wrong.
You are anthropomorphizing LLMs. They do not reason and they are not lazy. The paper discusses a way to improve their predictive output, not a way to actually make them reason.
But don't take my word for it. Go talk to ChatGPT. Ask it anything like this:
"If an LLM is provided enough processing power, would it eventually be conscious?"
"Are LLM neural networks like a human brain?"
"Do LLMs have thoughts?"
"Are LLMs similar in any way to human consciousness?"
Just always make sure to check the output of LLMs. Since they are complicated autosuggestion engines, they will sometimes confidently spout bullshit, so must be examined for correctness. (As my initial post discussed.)
You're assuming i'm saying something that i'm not, and then arguing with that, instead of my actual claim.
I'm saying we don't know for sure what they will be able to do when they're scaled up. That's the end of my assertion. I don't have to prove that they will suddenly come alive, i'm not claiming they will, i'm just claiming we don't know what will happen when they're scaled, and they seem to have emergent properties as they scale up. Nobody has devised a way of predicting what emergent properties happen when, nobody has made any progress whatsoever on knowing what scaling up accomplishes.
Can they reason? Yes, but poorly right now, will that get better? Who knows.
The end of my claim is that we don't know what'll happen when they scale up, and that you can't just write it off like you are.
If you want proof that they reason, see the research article I linked. If they can do that in their rudimentary form that we've created with very little time, we can't write off the possibility that they will scale.
Whether or not they reason LIKE HUMANS is irrelevant if they can do the job.
And i'm not anthropomorphizing them without reason, there aren't terms for this already, what would you call this behavior of answering questions significantly better when asked to fully explain reasoning? I would say it is taking the easiest option that still meets the qualifications of what it is requested to do, following the path of least resistance, I don't have a better word for this than laziness.
Furthermore predictive power is just another way of achieving reasoning, better predictive power IS better reasoning, because you can't predict well without reasoning.
It’s your job to prove your assertion that we know enough about cognition to make reasonable comparisons.
Are you just fucking around here? C'mon. In your hypothetical scenario, the emergent property would not be "of a calculator".
So for context, I am an applied mathematician, and I primarily work in neural computation. I have an essentially cursory knowledge of LLMs, their architecture, and the mathematics of how they work.
I hear this argument, that LLMs are glorified autocomplete and merely statistical inference machines and are therefore completely divorced from anything resembling human thought.
I feel the need to point out that not only is there no compelling evidence that any neural computation that humans do anything different from a statistical inference machine, there's actually quite a bit of evidence that that is exactly what real, biological neural networks do.
Now, admittedly, real neurons and real neural networks are way more sophisticated than any deep learning network module, real neural networks are extremely recurrent and extremely nonlinear, with some neural circuits devoted to simply changing how other neural circuits process signals without actually processing said signals on their own. And in the case of humans, several orders of magnitude larger than even the largest LLM.
All that said, it boils down to an insanely powerful statistical machine.
There are questions of motivation and input: we all want to stay alive (ish), avoid pain, and have constant feedback from sensory organs while a LLM just produces what it was supposed to. But in an abstraction the ideas of wants and needs and rewards aren't substantively different from prompts.
Anyway. I agree that modern AI is a poor substitute for real human intelligence, but the fundamental reason is a matter of complexity, not method.
Some reading:
Large scale neural recordings call for new insights to link brain and behavior
A unifying perspective on neural manifolds and circuits for cognition
a comparison of neuronal population dynamics measured with calcium imaging and electrophysiology
If you truly believe humans are simply autocompletion engines then I just don't know what to tell you. I think most reasonable people would disagree with you.
Humans have actual thoughts and emotions; LLMs do not. The neural networks that LLMs use, while based conceptually in biological neural networks, are not biological neural networks. It is not a difference of complexity, but of kind.
Additionally, no matter how many statistics, CPU power, or data you give an LLM, it will not develop cognition because it is not designed to mimic cognition. It is designed to link words together. It does that and nothing more.
A dog is more sentient than an LLM in the same way that a human is more sentient than a toaster.
In a more diplomatic reading of your post, I'll say this: Yes, I think humans are basically incredibly powerful autocomplete engines. The distinction is that an LLM has to autocomplete a single prompt at a time, with plenty of time between the prompt and response to consider the best result, while living animals are autocompleting a continuous and endless barrage of multimodal high resolution prompts and doing it quickly enough that we can manipulate the environment (prompt generator) to some level.
Yeah biocomputers are fucking wild and put silicates to shame. The issue I have is with considering biocomputation as something that fundamentally cannot be be done by any computational engine, and as far as neural computation is understood, it's a really sophisticated statistical prediction machine
We all want to believe that humans, or indeed animals as a whole, have some secret special sauce that makes us fundamentally distinguishable from statistical algorithms that approximate a best fit function according to some cost metric, but the fact of the matter is we don't.
There is no science to support the idea that biological neurons are particularly special, and there are reams and reams of papers suggestin that real neural cognition is little more than an extremely powerful statistical machine.
I don't care about what "most reasonable people" think. "Most reasonable people" don't have an opinion about the axiom of choice, or the existence of central pattern generators. That's not to devalue them but their opinions on things this far outside of their expertise are worth about as much as my opinions on the concept of art. I am a professional in neural computation, and I put it to you to even hypothesize about how animal neural computation is fundamentally distinct from LLM computation.
Like I said, we are wildly more capable than GPT, because our hardware is wildly more complex than any ANN, but the fundamental computing strategy is not all that different.
"most reasonable people" - indirect ad hominem is still ad hominem. You are making a fool of yourself.
I am picking up a hint of the autocompletion you describe, in your writing.
I think I write well :) I am not an LLM though.
I don't believe in scaling as a way to discover understanding. Doing that is just praying that the machine comes alive... these machines weren't programmed to come alive in that way. That's my fundamental argument, the design of LLMs ignores understanding of the content... it doesn't matter how much content it's been scaled up to.
If I teach a real AI about fishing, it should be able to reason about fishing and it shouldn't need to have read a supplementary knowledge of mankind to do it.
What the LLMs seem to be moving towards is more of a search and summary engine (for existing content). That's a very similar and potentially quite useful thing, but it's not the same thing as understanding.
It's the difference between the kid that doesn't know much but is really good at figuring it out based on what they know vs the kid that's read all the text books front to back and can't come up with anything original to save their life but can quickly regurgitate and summarize anything they've ever read.
If I teach a real AI about fishing, it should be able to reason about fishing and it shouldn’t need to have read a supplementary knowledge of mankind to do it.
This is a faulty assumption.
In order for you to learn about fishing, you had to learn a shitload about the world. Babies don't come out of the womb able to do such tasks, there is a shitload of prerequisite knowledge in order to fish, it's unfair to expect an ai to do this without prerequisite knowledge.
Furthermore, LLM's have been shown to do many things that aren't in their training data, so the notion that it's a stochastic parrot is also false.
Furthermore, LLM’s have been shown to do many things that aren’t in their training data, so the notion that it’s a stochastic parrot is also false.
And (from what I've seen) they get things wrong with extreme regularity, increasingly so as thing diverge from the training data. I wouldn't say they're a "stochastic parrot" but they don't seem to be much better when things need to be correct... and again, based on my (admittedly limited) understanding of their design, I don't anticipate this technology (at least without some kind of augmented approach that can reason about the substance) overcoming that.
In order for you to learn about fishing, you had to learn a shitload about the world. Babies don’t come out of the womb able to do such tasks, there is a shitload of prerequisite knowledge in order to fish, it’s unfair to expect an ai to do this without prerequisite knowledge.
That's missing the forest for the trees. Of course an AI isn't going to go fishing. However, I should be able to assert some facts about fishing and it should be able to reason based on those assertions. e.g. a child can work off of facts presented about fishing, "fish are hard to catch in muddy water" -> "the water is muddy, does that impact my chances of a catching a bluegill?" -> "yes, it does, bluegill are fish, and fish don't like muddy water".
There are also "teachings" brought about by how these are programmed that make the flaws less obvious, e.g., if I try to repeat the experiment in the post here Google's Bard outright refuses to continue because it doesn't have information about Ryan McGee. I've also seen Bard get notably better as it's been scaled up, early on I tried asking it about RuneScape and it spewed absolute nonsense. Now... It's reasonable-ish.
I was able to reproduce a nonsense response (once again) by asking about RuneScape. I asked how to get 99 firemaking, and it invented a mechanic that doesn't exist "Using a bonfire in the Charred Stump: The Charred Stump is a bonfire located in the Wilderness. It gives 150% Firemaking experience, but it is also dangerous because you can be attacked by other players." This is a novel (if not creative) invention of Bard likely derived from advice for training Prayer (which does have something in the Wilderness which gives 350% experience).
And (from what I’ve seen) they get things wrong with extreme regularity, increasingly so as thing diverge from the training data. I wouldn’t say they’re a “stochastic parrot” but they don’t seem to be much better when things need to be correct… and again, based on my (admittedly limited) understanding of their design, I don’t anticipate this technology (at least without some kind of augmented approach that can reason about the substance) overcoming that.
Keep in mind, you're talking about a rudimentary, introductory version of this, my argument is that we don't know what will happen when they've scaled up, we know for certain hallucinations become less frequent as the model size increases (see the statistics on gpt3 vs 4 on hallucinations), perhaps this only occurs because they haven't met a critical size yet? We don't know.
There's so much we don't know.
That’s missing the forest for the trees. Of course an AI isn’t going to go fishing. However, I should be able to assert some facts about fishing and it should be able to reason based on those assertions. e.g. a child can work off of facts presented about fishing, “fish are hard to catch in muddy water” -> “the water is muddy, does that impact my chances of a catching a bluegill?” -> “yes, it does, bluegill are fish, and fish don’t like muddy water”.
https://blog.research.google/2022/05/language-models-perform-reasoning-via.html
they do this already, albeit imperfectly, but again, this is like, a baby LLM.
and just to prove it:
https://chat.openai.com/share/54455afb-3eb8-4b7f-8fcc-e144a48b6798
And we're nowhere near dome scalimg LLM's
I think we might be, I remember hearing openAI was training on so much literary data that they didn't and couldn't find enough for testing the model. Though I may be misrememberimg.
No that's definitely the case. However, Microsoft is now working making LLM's more dependent on several high quality sources. For example: encyclopedias will be more important sources than random reddit posts.
Microsoft is also using LinkedIn to help as well, getting users to correct articles generated by AI.
Cunningham's Law may be very helpful in this respect.
"the best way to get the right answer on the internet is not to ask a question; it's to post the wrong answer."
I was mostly posting this because the last time LLMs came up, people kept on going on and on about how much their thoughts are like ours and how they know so much information. But as this article makes clear, they have no thoughts and know no information.
In many ways they are simply a mathematical party trick; formulas trained on so much language, they can produce language themselves. But there is no “there” there.
have no thoughts
True
know no information
False. There's plenty of information stored in the models, and plenty of papers that delve into how it's stored, or how to extract or modify it.
I guess you can nitpick over the work "know", and what it means, but as someone else pointed out, we don't actually know what that means in humans anyway. But LLMs do use the information stored in context, they don't simply regurgitate it verbatim. For example (from this article):
If you ask an LLM what's near the Eiffel Tower, it'll list location in Paris. If you edit its stored information to think the Eiffel Tower is in Rome, it'll actually start suggesting you sights in Rome instead.
They only use words in context, which is their problem. It doesn't know what the words mean or what the context means; it's glorified autocomplete.
I guess it depends on what you mean by "information." Since all of the words it uses are meaningless to it (it doesn't understand anything of what it either is asked or says), I would say it has no information and knows nothing. At least, nothing more than a calculator knows when it returns 7 + 8 = 15. It doesn't know what those numbers mean or what it represents; it's simply returning the result of a computation.
So too LLMs responding to language.
Why is that a problem?
For example, I've used it to learn the basics of Galois theory, and it worked pretty well.
- The information is stored in the model, do it can tell me the basics
- The interactive nature of taking to LLM actually helped me learn better than just reading.
- And I know enough general math so I can tell the rare occasions (and they indeed were rare) when it makes things up.
- Asking it questions can be better than searching Google, because Google needs exact keywords to find the answer, and the LLM can be more flexible (of course, neither will answer if the answer isn't in the index/training data).
So what if it doesn't understand Galois theory - it could teach it to me well enough. Frankly if it did actually understand it, I'd be worried about slavery.
Basically the problem is point 3.
You obviously know some of what it's telling you is inaccurate already. There is the possibility it's all bullshit. Granted a lot of it probably isn't, but it will tell you the bullshit with the exact same level of confidence as actual facts... because it doesn't know Galois theory and it isn't teaching it to you, it's simply stringing sentences together in response to your queries.
If a human were doing this we would rightly proclaim the human a bad teacher that didn't know their subject, and that you should go somewhere else to get your knowledge. That same critique should apply to the LLM as well.
That said it definitely can be a useful tool. I just would never fully trust knowledge I gained from an LLM. All of it needs to be reviewed for correctness by a human.
That same critique should apply to the LLM as well.
No, it shouldn't. Instead, you should compare it to the alternatives you have on hand.
The fact is,
- Using LLM was a better experience for me then reading a textbook.
- And it was also a better experience for me then watching recorded video lectures.
So, if I have to learn something, I have enough background to spot hallucinations, and I don't have a teacher (having graduated college, that's always true), I would consider using it, because it's better then the alternatives.
I just would never fully trust knowledge I gained from an LLM
There are plenty of cases where you shouldn't fully trust knowledge you gained from a human, too.
And there are, actually, cases where you can trust the knowledge gained from an LLM. Not because it sounds confident, but because you know how it behaves.
Obviously you should do what you think is right, so I mean, I'm not telling you you're living wrong. Do what you want.
The reason to not trust a human is different from the reasons not to trust an LLM. An LLM is not revealing to you knowledge it understands. Or even knowledge it doesn't understand. It's literally completing sentences based on word likelihood. It doesn't understand any of what it's saying, and none of it is rooted in any knowledge of the subject of any kind.
I find that concerning in terms of learning from it. But if it worked for you, then go for it.
Sadly we don't even know what "knowing" is, considering human memory changes every time it is accessed. We might just need language and language only. Right now they're testing if generating verbalized trains of thought helps (it might?). The question might change to: Does the sum total of human language have enough consistency to produce behavior we might call consciousness? Can we brute force the Chinese room with enough data?
They are the perfect embodiment of the internet.
They know everything, but understand nothing
I've been unemployed for 7 months. Every online job I see that's been posted for at least 6 hours has over 200 applications. I'm a senior Dev with 30 years experience, and I can't find work.
I'd say generative AI is an existential threat as bad as offshoring was for steel in the early 80s. I'm now left with the prospect of spending the last 20 years of my work life at or near minimum wage.
After all, I can't afford to spend $250,000 on a new bachelor's degree, and a community college degree might get me to $25/hr, and still costs thousands. This is causing impoverishment on a massive scale.
Ignore this threat at your peril.
Your issue sounds more like a capitalism issue. FANG companies lay off thousands of employees to cut costs and prepare for changes in the economy. AI didn't make them lay off all those employees, just corporate greed. Until AI can gather requirements, accurately produce code with at least 80%, can compile the software itself, it isn't a threat.
Edit fix autocorrect
and 100% accuracy. Only a fool would trust something coming out of AI and slap it right into production right now.
I agree though I was following the 80/20 rule. if the softwares essentially free and does 80% of your business needs businesses would be happy. Either way AI is nowhere near that since it requires someone with the knowledge currently to get it anywhere close to a complete project.
I understand how it works. The fact remains that companies already laid off people because of AI, and until now, I have never been unemployed more than 2 months.
I've also never seen a market in which most job posting garner 200 to 500 applications within 24 hours. It is armageddon out here.
I'm a senior dev too, and at first I thought the same, but really it's a market downturn. Companies are just afraid to hire right now. I'd look into generative AI, try to understand how it works. That's how I've been spending my time, and yeah, it's intuitive the way they do it but the more you understand how it works the more you realize that it's not ready to take our jobs. Yet. Again maybe someday, but there is a lot of work that needs to be done to get something semi up and running, and the models that Google uses are not going to be usable for every company. (Take a look at all the specialized models already).
Our job never goes away, but it does constantly evolve. This is just another point where we have to learn new skills, and that may be that we all need to be model tuners some day. At the end of the day the user still needs to correctly describe what they want to have happen on the screen, and there are currently no ways to take what they describe into a full piece of software.
Hard to believe a senior dev can't find work. Those positions are the most needed. Also 25 an hour is 50k a year. No where in the US are senior devs paid that little. I suppose you may not be US based, but your cost for college seems to imply US, albeit at an expensive school.
I was not saying 25 for a dev job, i was saying that for other kinds of work i might be able to get without getting a new degree.
And everyone in tech who has worked on ML before collectively says “yeah that’s what we’ve been trying to tell you”.
Everybody in tech would even have a passing understanding of the technology was collectively saying that. We understand the limits of technology and can feel out the bounds easily. But, too many of these dumbasses with dollar signs in their eyes are all "to the moon!", and tripping and failing on implementing the tech in unreasonable ways.
It was never a factoid machine, like some people wanted to believe. It was always about creatively writing something, and only one with so much attention.
It was never a factoid machine
Funny tidbit about the word "factoid": its original meaning was "an item of unreliable information that is reported and repeated so often that it becomes accepted as fact", but the modern usage is "a brief or trivial item of news or information".
This means that the modern usage of "factoid" is in itself a factoid, and that in the old sense LLMs sort of are factoid machines.
Note that I'm not saying the modern use is wrong. Languages evolve, and words taking on new meanings doesn't mean the new meanings are "wrong" (and surprisingly words changing to mean the opposite of what they used to mean isn't all that uncommon either.)
I disagree, a lot of white collar work is simply writing bullshit.