Meta “programmed it to simply not answer questions,” but it did anyway.
Hallucinating is a fancy term for BEING WRONG.
Unreliable bullshit generator is still unreliable. Imagine that!
AI doesn’t know what’s wrong or correct. It hallucinates every answer. It’s up to the supervisor to determine whether it’s wrong or correct.
Mathematically verifying the correctness of these algorithms is a hard problem. It’s intentional and the trade-off for the incredible efficiency.
Besides, it can only “know” what it has been trained on. It shouldn’t be suprising that it cannot answer about the Trump shooting. Anyone who thinks otherwise simply doesn’t know how to use these models.
It is impossible to mathematically determine if something is correct. Literally impossible.
At best the most popular answer, even if it is narrowed down to reliable sources, is what it can spit out. Even that isn’t the same thing is consensus, because AI is not intelligent.
If the ‘supervisor’ has to determine if it is right and wrong, what is the point of AI as a source of knowledge?
It is impossible to mathematically determine if something is correct. Literally impossible.
No, you’re wrong. You can indeed prove the correctness of a neural network. You can also prove the correctness of many things. It’s the most integral part of mathematics and computer-science.
For example a very simple proof: with the conjecture that an even number is 2k of a number k, then you can prove that the addition of two even numbers is again an even number (and that prove is definite): 2a+2b=2(a+b), since a+b=k for some k.
Obviously, proving more complex mathematical problems like AI is more involved. But that’s why we have scientists that work on that.
At best the most popular answer, even if it is narrowed down to reliable sources, is what it can spit out. Even that isn’t the same thing is consensus, because AI is not intelligent.
That is correct. But it’s not a limitation. It’s by design. It’s the tradeoff for the efficiency of the models. It’s like lossy JPG compression. You accept some artifacts but in return you get much smaller images and much faster loading times.
But there are indeed "AI"s and neural networks that have been proven correct. This is mostly applied to safety critical applications like airplane collision avoidance systems or DAS. But a language model is not safety critical; so we take full advantage.
If the ‘supervisor’ has to determine if it is right and wrong, what is the point of AI as a source of knowledge?
You’re completely misunderstanding the whole thing. The only reason why it’s so incredibly good in many applications is because it’s bad in others. It’s intentionally designed that way. There are exact algorithms and there approximation algorithms. The latter tend to be much more efficient and usable in practice.
You can prove some things are correct, like math problems (assuming the axioms they are based on are also correct).
You can’t prove that things like events having happened are correct. That’s even a philosophical issue with human memory. We can’t prove anything in the past actually happened. We can hope that our memory of events is accurate and reliable and work from there, but it can’t actually be proven. In theory everything before could have just been implanted into our minds. This is incredibly unlikely (as well as not useful at best), but it can’t be ruled out.
If we could prove events in the past are true we wouldn’t have so many pseudo-historians making up crazy things about the pyramids, or whatever else. We can collect evidence and make inferences, but we can’t prove it because it is no longer happening. There’s a chance that we miss something or some information can’t be recovered.
LLMs are algorithms that use large amounts of data to identify correlations. You can tune them to give more unique answers or more consistent answers (and other conditions) but they aren’t intelligent. They are, at best, correlation finders. If you give it bad data (internet conversations) or incomplete data then it at best will (usually confidently) give back bad information. People who don’t understand how they work assume they’re actually intelligent and can do more than this. This is dangerous and should be dispelled quickly, or they believe any garbage it spits out, like the example from this post.
You can’t prove that things like events having happened are correct.
You can’t so solidly that this shouldn’t even be discussed.
What should be is whether you can make a machine capable of reasoning.
There’s symbolic logic, so you can maybe some day make a machine that makes correct syllogisms, detects incorrect syllogisms and such.
People who don’t understand how they work assume they’re actually intelligent and can do more than this. This is dangerous and should be dispelled quickly, or they believe any garbage it spits out, like the example from this post.
Sadly there’s that archetype of “the narrow-minded not cool scientist against the cool brave inventor” which means that actively dispelling that may do harm. People who don’t understand will match the situation with that archetype and it will reinforce their belief.
Well but this kind of correctness applies to everything. By thag logic, you can’t believe anything. I’m talking about an entirely different correctness. Like resistance against certain adversarial attacks. Of course, proving that the model is always correct, is as complicated as modelling the entire reality. That’s infeasible. But it’s also infeasible for every other software.
This sounds like an overly pedantic view of “prove”
It’s not pedantic. You can mathematically prove math.
You can’t mathematically/algorithmically prove an event happened or did not happen.
Adding “mathematically/algorithmically” in front of the word “prove” as if it were always implicitly there, and suggesting that it’s the only way we should be using the word “prove” seems pretty darned pedantic to me.
You can prove mathematical logic and you can (not 1-to-1) tie that to symbolic logic, but since it’s not 1-to-1, because of ambiguity of symbols, there will be much more complexity. I personally think that the future of various machine assistants lies there, and what LLM’s now do is going to be used in auxiliary roles for that.
No. It’s just pure math and logic. And LLMs are nothing more than billions of additions and multiplications. Literally. You can prove certain things on it just like you can prove theorems in mathematics. It’s an ongoing ressearch field.
It’s just pure math and logic. And LLMs are nothing more than billions of additions and multiplications.
Okay: using additions and multiplications prove the assassination attempt on Donald Trump happened
Your proof example is a proof from your discrete structures class. That’s very different than “proving” something like “the Trump assassination attempt was a conspiracy.”
Otherwise we could have gotten rid of courts a long time ago.
Well obviously. But that was not at all what I said or claimed. I just said that you can prove certain properties of neural networks because others said that you can’t. And others also misunderstood LLMs in general. They believe it’s an information retrival service, which is wrong.
Besides, your argument, as you’ve written it, applies to everything. Literally. From Wikipedia, to News, even up to your eyesight. What can you actually prove? I don’t understand the point you’re making and how that is related to LLMs.
Just like us. Sometimes it’s better to have bullshit predictions than none.
The only reason why it’s so incredibly good in many applications is because it’s bad in others. It’s intentionally designed that way.
lolwut
It’s designed in a ways that’ll make it inherently incorrect. Even on a physical basis (due to numeric issues). It’s not a problem of the algorithm because it has been designed that way. The problem is that you don’t know how to correctly use it.
I can’t explain it any differently without getting overly technical. You wouldn’t understand it anyways, judging by your comment “lolwut”. If you want to learn how LLMs work specifically, there are plenty of ressources on the internet.
It’s designed in a ways that’ll make it inherently incorrect. Even on a physical basis (due to numeric issues). It’s not a problem of the algorithm because it has been designed that way. The problem is that you don’t know how to correctly use it.
“It doesn’t make a good source of knowledge.”
“Yeah, but it is designed to be inherently wrong”
How does that make any sense when trying to use something for knowledge? Being inherently wrong is the opposite of helpful for knowledge.
AI is great at pattern recognition, but knowledge isn’t pattern recognition. Needing to know when it gives false information requires the “supervisor” to already have that knowledge. That makes the AI less useful than a simple reference because at least the reference can come from a trusted source.
If people stopped trying to jam AI into situations where being correct is important it wouldn’t be a problem. But excusing that because it is designed to be inherently wrong deserves another LOLWUT.
How does that make any sense when trying to use something for knowledge? Being inherently wrong is the opposite of helpful for knowledge.
It was never designed to reproduce knowledge. It was designed to do reasoning and natural language processing and generation. You’re using it wrong.
LULWUT
If you don’t know what you’re talking about and don’t have any capacity to learn something new, it’s sometimes best to stop talking. Especially when you’re starting to get rude to knowlegable people that try to explain it to you.
It’s designed in a ways that’ll make it inherently incorrect. Even on a physical basis (due to numeric issues). It’s not a problem of the algorithm because it has been designed that way. The problem is that you don’t know how to correctly use it.
So it is bad at things like giving or finding factual information. I agree, companies need to stop cramming it into everything (like search engines) for tasks that it is specifically bad at because it is not designed for it.
Can you recommend any for resource to start with? (If I can be picky, then something I can consume after a whole day of being a patent because there is no energy for much else.)
That is, unless you define correct in mathematical terms. Which no one has done yet.
It also wouldn’t be a source of knowledge. It would be a shitty calculator.
We should understand that 99.9% of what wee say and think and believe is what feels good to us and we then rationalize using very faulty reasoning, and that’s only when really challenged! You know how I came up with these words? I hallucinated them. It’s just a guided hallucination. People with certain mental illnesses are less guided by their senses. We aren’t magic and I don’t get why it is so hard for humans to accept how any individual is nearly useless for figuring anything out. We have to work as agents too, so why do we expect an early days LLM to be perfect? It’s so odd to me. Computer is trying to understand our made up bullshit. A logic machine trying to comprehend bullshit. It is amazing it even appears to understand anything at all.
You know how I came up with these words? I hallucinated them. It’s just a guided hallucination.
The the word hallucination means literally anything you want it to. Cool, cool. Very valiant of you.
Uhm. Have you ever talked to a human being.
Human beings are not infallible either.
That’s like saying car crash is just a fancy word for accident, or cat is just a fancy term for animal.
Hallucination is a technical term for this type of AI, and it’s inherent to how it works at it’s core.
And now I’ll let you get back to your hating.
Hallucination is also wildly misleading. The AI does not believe something that isn’t real, it was incorrect in the words it guessed would be appropriate.
The funny thing is we hallucinate all our answers too. I don’t know where these words are coming from and I am not reasoning about them other than construction of a grammatically correct sentence. Why did I type this? I don’t have a fucking clue. 😂
We map our meanings onto whatever words we see fit. If I had a dollar for every time I’ve heard a Republican call Obama a Marxist still blows my mind.
Thank you for saying something too. Better than I could do. I’ve been thinking about AI since I was a little kid. I’ve watched it go from at best some heuristic pathfinding in video games all the way to what we have now. Most people just weren’t ever paying attention. It’s been incredible to see that any of this was even possible.
I watched Two Minute Papers from back when he was mostly doing light transport simulation (raytracing). It’s incredible where we are, but baffling people can’t see the tech as separate from good old capitalism and the owner class. It just so happens it takes a fuckton of money to build stuff like this, especially at first. This is super early.
Kaplan noted that AI chatbots “are not always reliable when it comes to breaking news or returning information in real time,” because “the responses generated by large language models that power these chatbots are based on the data on which they were trained, which can at times understandably create some issues when AI is asked about rapidly developing real-time topics that occur after they were trained.”
If you’re expecting a glorified autocomplete to know about things it doesn’t have in its training data, you’re an idiot.
There are definitely idiots, but these idiots don’t get their ideas of how the world works out of thin air. These AI chatbot companies push the cartoon reality that this is a smart robot that knows things hard in their advertisements, and to learn otherwise you have to either listen to smart people or read a lot of text.
I just assumed that its bs at first, but I also once nearly went unga bunga caveman against a computer from 1978. So I probably have a deeper understanding of how dumb computers can be.
Some services will use glorified RAG to put more current info in the context.
But yeah, if it’s just the raw model, I’m not sure what they were expecting.
Sir, are you telling me AI isn’t a panacea for conveying facts? /s
The shooting happened after the end of the training date. Like asking windows 95 clippy about 9/11 and it saying it didn’t happen.
Clippy being a 9/11 conspiracy theorist is now canon
maybe Meta AI is into something
Is it wrong to root this on simply because I hate that shitbag?
Hatred is a path to the dark side.
As evidenced by you now rooting for misinformation.
Oh I’m far to pragmatic to believe that. If truth isn’t working, then what choice do you really have?
Well, if the chatbot learned anything from Dementia Don the racist rapist with 34 felonies that can’t complete a coherent sentence, it learned that you never tell the truth.
Does this AI work with real time info?
Does the AI consistently say that, no matter who asks?
Because if so, that’s not a hallucination.
