Top
Best
New

Posted by jalev 2 days ago

The LLMentalist Effect (2023)(softwarecrisis.dev)
230 points | 307 commentspage 2
tock 2 days ago|
These questions will never go anywhere because there is no good definition of intelligence or consciousness.
scotty79 2 days ago|
There is a good definition of intelligence. It's what intelligence tests measure. If you define it like that, you can make harder tests, but as long as both humans and AIs can attempt to solve them we have measure and something to compare.

Consciousness though might be fully illusory and meaningless as many philosophical concepts before ultimately turned out to be.

pixl97 2 days ago|||
Intelligent needs to bedefined with scales.

Cells are intelligent. Organs have intelligence on top of that, that cells do not have. Plants and animals have intelligence that their organs do not.

This is the real problem of intelligence, the definition one uses. It can cause them to be blind of all the things intelligence is capable of.

tock 2 days ago||||
> If you define it like that

:)

scotty79 2 days ago||
The point is, defining it like that has merit proven in decades before AI.
slopinthebag 2 days ago|||
> There is a good definition of intelligence. It's what intelligence tests measure.

what do intelligence tests measure?

scotty79 2 days ago||
Intelligence.

It's not an exact analogy, rather something to consider ...

How much momentum does a body have? You can measure it easily, just measure the velocity and the mass and multiply.

What exactly did you measure? What is the meaning of this? Nobody really knows. But we know we can measure this and it's a thing and it's conserved.

slopinthebag 1 day ago||
eh, momentum has a pretty rigorous definition and it's possible to define it qualitatively. perhaps a better example would have been hardness, where A is harder than B when it scratches B in a test. but it's borderline, and quite a bit different from true circular reasoning

test X measures I(x) I(x) is defined as the thing test X measures. Therefore test X measures the pre-existing psychological trait intelligence.

this is absurd

scotty79 1 day ago||
There was no pre-existing trait called intelligence to measure.

The was just a vague idea that some people might have a bit more of some underspecified general quality but there never was any good idea what exactly that might be. So making a test that measures something correlated with outcomes that people associate with this nebulous pre-existing idea of intelligence is honest and fair stab at something that eluded our grasp for millennia.

Intelligence as people were talking about it before we created intelligence tests is something like atoms of Democritus. And IQs are atoms as we mesured them today.

draftpunked 2 days ago||
I don't think it's a "con," but I find that using the mental model of LLMs being sophisticated, lossy search engines of knowledge can help us separate some of the factors more cleanly than imagining that they have cognition or intelligence.

It's hard for me to imagine a stateless operation as intelligence per se, though perhaps the chaining of such operations starts looking more like it?

mindcrime 2 days ago|
It's hard for me to imagine a stateless operation as intelligence per se

That's an interesting point. My take would be to say that we shouldn't think of the AI as being just the model, but should include the harness. At that level, clearly we can keep state / context and that is probably a more natural mapping to our intuitive understanding of "intelligence".

natbennett 2 days ago||
> Delegating your decision-making, ranking, assessment, strategising, analysis, or any other form of reasoning to a chatbot becomes the functional equivalent to phoning a psychic for advice.

Lots of comments talking about how recent accomplishments disprove the article but I think this bit holds up pretty well.

LLMs are very good at tricking people into thinking they have capabilities that they don’t.

kraf 2 days ago||
I keep being confused about how people's understanding of the models get stuck at next token prediction. Isn't this entirely neglecting the RL training? I might be misunderstanding something but to my mind it makes the issue way fuzzier than it's being painted here.
mindcrime 2 days ago|
I keep being confused about how people's understanding of the models get stuck at next token prediction.

Heh. A lot of anti-ai hucksters I see posting on LinkedIn just LOVE to use the phrase "next token prediction" and the word "autoregressive". They've almost become shibboleths that identify members of that camp. That and the classic rallying cry of "Linear Algebra isn't intelligent!"

The best take I've seen on that recently, was somebody who made the point "just think of the next token prediction part as the output layer". Which makes perfect sense.. if you're replying in natural language, at some point in the flow, you have to construct a sentence and starting at the head and predicting next tokens is perfectly reasonable. I'm doing it literally as I'm typing these characters, for crying out loud!

But the mistake is to think that LLM's only "predict next tokens" with no consideration of the possibility that they are actually constructing richer representations, building concepts, making analogies, doing abduction, induction, etc. My own (admittedly anecdotal) take on working with LLM's suggests to me that they do do those things, albeit probably not the same way humans do.

I think a lot of folks are missing the point by being overly reductive when they start talking about "next token prediction" and "autoregressive". It's like, can we say "Phil (me) isn't intelligent because there's nothing going on but some electrical impulses and chemistry happening inside his brain. Everybody knows electricity and chemistry aren't intelligent!"

NinjaTrance 2 days ago||
LLMs may not think, but they do reason.

I say that because the word "reason" originates from the Latin word "ratio," which means "calculation".

LLMs do calculations to produce their answers -- thus, they reason.

pixl97 2 days ago|
Honestly the word think is so poorly defined it cannot be used with any scientific rigor and is rather useless without a dissertation being posed with it on what you actually mean by that. Intelligence really needs one too.
ungreased0675 2 days ago||
It does seem like LLMs share the language of psychics.

The author is also correct that LLM evangelicals and believers in the occult speak about it similarly.

pixl97 2 days ago|
The converse is also true. Anti-AI people make bold non-scientifically backed claims that we are supposed to accept like it's written in a Bible.

The more I've learned about intelligence and intelligent behavior the more I realize I don't know and this rabbit hole goes deep.

mstank 2 days ago||
This is a strange newsletter post. It’s harping against AI but seems AI-written itself.

Its ultimate conclusion:

“I’ve come to the conclusion that a language model is almost always the wrong tool for the job.

I strongly advise against integrating an LLM or chatbot into your product, website, or organisational processes.”

Seems so obviously biased that I can only understand it with the context that the writer is trying to sell their book for €35

mstank 2 days ago||
Just realized this was written in 2023. Can’t believe how wrong we were about AI back then.
JoshTriplett 2 days ago||
Can't believe how right we were, too.
daishi55 2 days ago||
You seem to be winking a little bit and I’m not sure exactly what you’re trying to say, but I agree with the one you are replying to: the author of this article wrote a lot of words - like a lot a lot of words - only to look very foolish a short time later. And I think much ink was similarly spilled writing very similar piece smirking at people who thought LLMs could do real intellectual work. And they were all wrong.
JoshTriplett 2 days ago||
It is absolutely the case that some things people wrote about AI and LLMs in the past is no longer true; it's a fast-changing area.

It's also the case that some things remain true. If someone says "don't use them for X because their quality isn't good enough", that's a recommendation with an expiration date. But if someone says "don't use them for X" for any number of other good reasons, that recommendation is often timeless.

"Can it" is changing rapidly. Much more rapidly than "Should it".

rpdillon 2 days ago|||
> But if someone says "don't use them for X" for any number of other good reasons, that recommendation is often timeless.

Are the good reasons timeless? I guess it's hard to parse what you're saying without some concrete examples. Is there any advice about what to use them for from 2022 or 2023 that you feel is true today, and you expect will be timeless? You said "often", so I'm assuming there are a handful of examples.

Ultimately though, this is not an objective discussion, since we're talking about "should", so I'm mainly asking out of curiosity on your view of things. The main example I can think of today is that you shouldn't do professional prose writing with LLMs. Obviously the models have gotten better, but writing is one of those areas where it produces middling output, though even here I hesitate because with a 21% illiteracy rate, it writes better than many adults in the US.

daishi55 2 days ago|||
Can you give an example of some of these things?
ryuuseijin 2 days ago||
The post makes sense for 2023. Today, not so much.

Also, definitely not AI written if the date is accurate.

scotty79 2 days ago||
> Also, definitely not AI written if the date is accurate.

Which is hilarious.

pixl97 2 days ago||
Hacker News, successfully detecting 375 out of 4 AI written articles.
scotty79 2 days ago||
> The intelligence illusion is in the mind of the user and not in the LLM itself.

> Many AI critics, including myself, are firmly in the second camp.

How can you reconcile this with the fact that AI can solve a real, intelligence bound problem for me (with zero intelligent effort on my part) that you can't?

Is the solution also illusory?

pixl97 2 days ago|
Horseshoe theory. At each end of the shoe you tend to get loud crazies.
scotty79 2 days ago|
> The intelligence illusion is in the mind of the user and not in the LLM itself.

I think that might be the solution to consciousness illusion.

The consciousness might be purely in the mind of someone that believes some other entity to be conscious.

pixl97 2 days ago|
I like to think of it this way. Consciousness is the outcome of intelligence. You can't get consciousness without it. Conversely this also means things that are not conscious can still be intelligent.
More comments...