Hacker News new | past | comments | ask | show | jobs | submit
We can chalk this up as another example of over-exhuberance by what folks believe LLMs can accomplish vs. what they actually are.

LLM-based “AI” is able to use its vast corpus of inputs and calculate the most statistically likely output in a given situation. It is probabilistic, and when you are dealing with probabilities in a situation where certainties, not probabilities, matter, you’re going to get dinged on credibility massively when your LLM-based “AI” gets the probabilities wrong at best, or in this case, claims a line of code generates a vulnerability when it is, in fact, a code comment.

LLMs are text-prediction engines. They are not Artificial Intelligence, and shouldn’t not be treated in any form or fashion as if they possess intelligence. What bothers me about this entire situation is that presumably the folks that relied on the LLM-based “AI” to generate these vulnerabilities knew (or should have known) enough about their tool to know this would happen, but did not.

Now, we all pay the consequence, to the tune of hundreds of thousands if not millions of dollars of wasted productivity from teams that have to deal with the resulting fall-out of this usage of “AI”.

A human must verify everything an LLM presents as fact. Everything. If you don’t, we all pay the price. LLMs do not remove the onus of responsibility on the human being, if anything they amplify it because LLMs can generate lots more output more quickly that needs to be verified than humans can.

Many people with no skills are taking advantage of the LLM craze to artificially inflate their own value. I see it every day on LinkedIn.

People that previously have barely any experience in tech, now being hired in AI startups because they are good bullshitters.

Countless directors and managers are now cosplaying as engineers. I've seen so many myself and that's just my tiny slice of this engineering world
It is pretty funny to see the shoe on the other foot, since it's usually software engineers with unearned arrogance about other fields.
What fields do you see devs think they know about? I’ve never personally seen this with other devs I work with but obviously small sample size…
loading story #49160087
loading story #49159103
Engineers cosplay as physicists and mathematicians every day. What's your point? Think of it pragmatically. If they can do the job they can do the role.
> If they can do the job they can do the role.

Obviously.

Can they do the job? Because right now, government decisions are based on AI generated code, which was verified by nobody who can do that. So the cost of an unsatisfactory answer is quite high.

To be fair, "people with no skills inflating their own value" is what LinkedIn has always been like. But I guess LLMs are uniquely well positioned for that task.
LLMs have brought it to a new level. I have multiple people in my feed posting several page dissertations on a daily basis. More often than not, these posts are about the benefits of AI...
People who live and create on LinkedIn? absolutely. For a huge group of people LI was just the most popular job board, though.
I heard the term "Sloperator" a few times. Quite fitting...
> People that previously have barely any experience in tech, now being hired in AI startups because they are good bullshitters.

I like the symmetry of people being unable to detect "skill" in people who are great at manipulating language and in LLMs who are also great at manipulating language.

This isn't new due to AI, just amplified.
> A human must verify everything an LLM presents as fact. Everything. If you don’t, we all pay the price. LLMs do not remove the onus of responsibility on the human being, if anything they amplify it because LLMs can generate lots more output more quickly that needs to be verified than humans can.

The sort of person that's going to offload their thinking to AI is the exact sort of person that is not going to verify anything because they've already offloaded their thinking to AI.

loading story #49160090
Unfortunately people sometimes get defensive against this take. But I think treating the LLM as you described can make you a better LLM user and help get better output. It helps understand the failure modes better, and moderate one’s reliance on them. Just like how we should do for every tool we work with.
Yes, I've found that reminding yourself of how they actually work helps keep you on guard against LLM-patterned mistakes. Especially things like carefully considering what parts of the current task likely fall outside the distribution of corpus + RL data (as much as that can be guessed).
> ...Unfortunately people sometimes get defensive against this take. But I think treating the LLM as you described can make you a better LLM user and help get better output. It helps understand the failure modes better, and moderate one’s reliance on them. Just like how we should do for every tool we work with.

B...b...but the Anthropic trainer said we'd get the best results if we don't think of it as a tool, but instead give it a name and think of it as our brilliant coworker!

Why should I trust you, internet rando over a stormtrooper-level salesman? /s

> Now, we all pay the consequence, to the tune of hundreds of thousands if not millions of dollars of wasted productivity from teams that have to deal with the resulting fall-out of this usage of “AI”.

Brandolini's principle in action. It takes 10 times more energy to refute BS than to generate it. A related analogy to computing: it is easy to generate propositions, but hard to test if a given proposition is satisfiable or not, which curiously ties to P vs NP.

> Brandolini's principle

I much prefer the alternative name: the Bullshit Asymmetry Principle.

Seems like most of it is covered by: Entropy increases.
Apophatic intelligence? "We don't know what intelligence is, but LLMs with CoT are certainly not it despite being Turing-complete."

Watching for unexpected failure modes is surely worth it.

Turing-completeness is a necessary pre-requisite for being able to fulfill the requirements of a Turing machine, nothing more. In the same way that cell division is a necessary condition for life, but cell division does not mean a given life form itself is sentient.

Intelligent life-forms can generate probabilistic outputs based on inputs, but being able to generate probabilistic outputs based on inputs is not what makes us intelligent.

OK. A more pointed question. What do you know about intelligence that allows you to exclude LLMs with CoT from the category of intelligent systems with certainty?
How do you know there's not a teapot orbiting the sun?
There is no reason to think that teapots were sent into orbit or spontaneously formed there.

Likewise, there is no reason to think the brain employs super-Turing or quantum computations that cannot be approximated by LLMs.

The point is that you're asking someone to prove a negative.
Right, it follows naturally that one can never meaningfully discuss that which can only be settled by proving negatives. Hence the total lack of debate on religious topics, as everyone just gave up 1500 years ago and decided that God indeed may or may not exist and that's that.
It's a negative in your worldview where "intelligence can't be a computation" is an axiom or something close. To me this worldview is sloppy thinking.
I did not express my opinion or "worldview." I was simply responding to this:

> exclude LLMs with CoT from the category of intelligent systems with certainty

Why do you think that intelligence of a Turing-complete system trained on human data can be likened to a Russell's teapot?

At least, don't you think that the recent mathematical results of LLMs are a bit like a glimpse of something teapot-shaped in the orbit? (which makes it not a Russell's teapot, which. by definition, can't be observed).

To me, it's an expected progression of ANNs' approximation of human cognitive processes. The universal approximation theorem guaranties the existence of such ANNs barring the super-Turing or quantum superiority of the brain.

> Why do you think that intelligence of a Turing-complete system trained on human data can be likened to a Russell's teapot?

I don't? You are presenting opinions as if they are mine, but they are not.

To my knowledge there are many millions of teapots orbiting the sun.
No, they're all orbiting the Earth very badly. :p
You’re right to point that out. The load-bearing seam of your logical retort is the hidden assumption that the teapot is in outer space, not at peace on a kitchen stovetop here on earth. I am sorry I had not considered this fact. \s
They were modeled after us, which almost certainly dooms them to stupidity.

They could have been great, if trained on datasets from a more sensible species.

> but being able to generate probabilistic outputs based on inputs is not what makes us intelligent.

???

Of course it is. The brain is mechanically not capable of doing anything other than that.

Do you believe the brain is something other than a bundle of probabilistic physical interactions? Or are brains not the source of what we call intelligence?

May I suggest the excellent & recent book: "The brain. In theory" by Romain Brette?

This is going to elevate your thinking on this no end, if you're interested.

Looks interesting (added to my list, thank you for the recommendation), but not actually relevant to the topic at hand.

We know that the brain is a probabilistic input → output machine because the universe is a probabilistic input → output machine. The brain is made of universe. There are deterministic relationships (which at high sensitivity or complexity become easier to describe as probabilistic), and quantum relationships. That's it. The brain, like every other thing comprised of "universe" is comprised of those two types of relationships.

If Romain's book provides evidence of relationships in the brain that are neither quantum (therefore random) NOR classical (therefore deterministic), then 1) he would have already won at least one Nobel prize, and 2) anyone in this thread would be able to at least gesture toward what relationship that is.

loading story #49159496
We don’t know
Of course we do. Unless the brain is an object unlike anything else in the known universe, then it's an input → output machine.
This deserves about as large of a "[citation needed]" as one could draw. Are you well-studied in neuroscience?
Nothing in the universe is anything other than a chain of reactions, amigo.

The brain is an object in the universe.

The universe has quantum behaviors (fully random, not a source of intelligence) and it has deterministic behaviors (fully non-random). Many of those deterministic behaviors are so complex that they're easier to analyze and describe as probabilistic, which is where most brain input → output relationships land.

Please point to any evidence whatsoever that the brain has some third type of interaction going on that has never been observed anywhere in the entire universe, then we can have a discussion about it.

Its baffling to me that people think the brain is somehow "outside" of the universe and somehow "magic" and not tied down by physics just like any other matter is. I don't understand how people come to this conclusion, except in the case of belief in magic or religion (which is belief in magic more or less).
Yeah, that's ultimately this whole conversation and people just don't have the courage to acknowledge their own beliefs.

The brain is a (very complex, incredible) input → output machine. That's it! It's incredible!

I don't understand why people are so afraid of this that they will believe otherwise with literally zero evidence whatsoever.

Why is the brain probabilistic instead of deterministic?
Probabilistic and deterministic are not opposites.

The brain is deterministic at the level of specific interactions, which process inputs in a highly chaotic (but still deterministic) manner into probabilistic outcomes.

The opposite of deterministic is random, i.e. in the quantum sense of truly no relationship between input and output.

There are probably some quantum effects in the brain here and there, but the vast majority of it is just traditional deterministic interactions networked together in such a complex system that the resulting behavior is much, much easier to predict in probabilistic terms than otherwise.

Ok right, but then I don't see the original point. Or like, you don't seem to be appealing to a hard science/soft science distinction here if the question revolves around what conceptual model is most "useful" to us. Useful for what? For predicting future actions? That makes sense, but does actually that speak to op's (possibly still fraught) metaphysical/semantic point here? It just feels like you are speaking over it, right? They could just as easily counter that a probabilistic model of the physical brain is not a very useful description of human intelligence, so why use it? You can't appeal to materialism here because you are already practicing some implied pragmatism in your commitments. You pretty much have as much ground to stand on as they do!
I am saying that there is no categorical distinction to be made on functional grounds between artificial intelligence and brains.

You cannot say an AI model cannot be intelligent because it's a probability machine, when all available evidence points toward natural intelligence also being generated by probability machines (much more complex ones, called brains).

loading story #49159874
loading story #49160460
Yours is a controversial view. It is lazy and selfish to try to get other people to explain their case that it is not exclusively that, when saying that it is exclusively that is the weaker case, and you back it up with nothing but a snarky proclamation.

Are newly born babies reacting due to statistical probabilities that they have derived, or are they using something other than their brains?

> Are newly born babies reacting due to statistical probabilities that they have derived, or are they using something other than their brains?

The answer is obviously yes lol.

The creature is an assemblage of electrical, chemical, and kinetic relationships.

Watching a baby develop is exactly what you'd expect from a system that's predominantly electrical noise triggering behaviors and then gradually refining denoising the relationship between inputs and outputs, with the goal function of achieving more desirable inputs.

Surely you can at least gesture toward one thing in the brain that appears not to be a probabilistic relationship between input and output?

Wow you just got us a complete nostalgia moment to the good old times when the computer who always beats us at chess became `not artificial intelligence`...
Every day I wake up and open HN.

“LLM has made legitimate mathematical discoveries” —> Wow the rate of progress is amazing. Highly upvoted.

“LLM does something not good” -> Does everyone else not realize LLMs are just dumb next token predictors? Highly upvoted.

So tired of this discourse and this site.

The rate of progress can be high and they can also be dumb next token predictors. Not sure why that is hard to understand.

These models can do a lot of things but they also can't do a lot of things. In order to use these models effectively you have to understand that they are next token predictors and how that allows it to do what they do.

Are they useful or not? Will they continue changing the world or not? People who choose one way or the other for describing them typically fall on one side or the other in these questions imo. What do you think? Will these next token predictors change the world or not?
loading story #49158262
loading story #49159060
It can be a token predictor and still tell me exactly how my life will proceed from now until the indefinite future, or be the most intelligent conversational entity you have ever witnessed.

The issue is of course with using the word "dumb": they are next token predictors, no doubt about it, but whether LLms as a class of system are smart or dumb is entirely unknown and entirely variable in time.

To interact with them effectively you must know how they behave, just like you have to know how humans behave to interact with them effectively. If you disagree, find someone with autism and have a conversation with them.

loading story #49159797
Much of an LLM's capability comes from the structure encoded in its learned representations. The probabilistic outputs are primarily a way of expressing uncertainty and generating fluent text, while compression during training is what forces the model to discover that underlying structure.
loading story #49158695
Not sure what your point is? Those things can both be true.

Or should the discourse in a diverse community like HN only reflect the positions you personally hold?

Tell me how a 'nExT toKeN prEdIcTor' can make breakthroughs in math or play a game of chess. These activities aren't pure symbol manipulation, they require actual understanding at some level.
loading story #49159053
loading story #49158501
Would be nice to get high karma commenter votes count only ..
Not a dichotomy actually. Highly depends on the task.
Opinions differ. This is not news.
It's almost as if there were many people using this site, and there is no clear consensus on LLMs, so people from various camps upvote interesting stores to support their cause. And people who are still somewhat undecided upvote both, if they present good evidence.

I mean even perennially contentious topics will get this behavior.... some thing about emacs makes the front page, within a day or two there will be a vim post up there. Same with Rust is (good|bad), or if systemd creates an even more awesome tool, the haters will come along and recycle stories about bugs from over a decade ago.

There's a lot of people here. Not all of them read it every hour, and discussions like this among large groups often take a very long time with lots of repetition. Human group dynamics (aka politics) is slow.

> So tired of this discourse and this site.

You're welcome to leave if you don't like it. The site was like this long before you joined, and will like it long after you leave I'm sure.

It's also worth noting, that an awful lot of math discoveries are perfectly in line with dumb next token generators - they are finding a way to formally construct an argument and being surprised when it doesn't work, or surprised at the outcome of the grind. Not all of them are made by brilliant leaps of intuition.

> LLMs are text-prediction engines. They are not Artificial Intelligence, and shouldn’t not be treated in any form or fashion as if they possess intelligence.

I agree that humans must verify LLM-produced facts, but strongly disagree with these kinds of "stochastic parrot therefore dumb" arguments.

Yes, an LLM is a "stochastic parrot". No, that doesn't imply that it is dumb. Enough to look at how Terence Tao asks ChatGPT to help him understand a solution that nobody had ever discussed before [1], or how a random guy asks ChatGPT in a handful of words to disprove a 30-year-old conjecture, with zero technical input [2].

If your parrot in a birdcage with internet access can finish the sentence, "The counterexample to the Dinitz–Garg–Goemans conjecture is...", then it's a pretty smart parrot, by all reasonable definitions of "smart". Just because someone bottled up the formula into matrix multiplications and added some random sampling to the outcome, that doesn't take away from the fact that the parrot said provably correct statements that the biggest experts in the field couldn't imagine.

And no, I'm not implying that the LLMs are correct all the time, or that their intelligence and reasoning works in any way like ours.

[1]: https://chatgpt.com/share/6a5fdc7a-d6f8-83e8-bbea-8deb42cfed... [2]: https://chatgpt.com/share/6a60b2eb-0b64-83ee-9c76-7931ca1de0...

loading story #49159291
This is a conflation of issues, predicated on false understanding of what LLMs are.

This line of critique is pernicious because it is both technically correct, as description, and profoundly misleading.

Saying that outputs are a product of inputs is not interesting and to the point it is not explanatory.

What is interesting, is how they do what they do. What is the "statistically likely* next token? To answer that you can do exactly one thing, run the LLM. That's because what they are doing is interesting and not reducible.

What is more interesting is that in order to do what they do, given the architectures we apply and the training strategies we use and the harnesses we situate them in, LLM are recapitulating in their deep layers strategies observed in the animal brain. This is still suggestive, interpretibility is nascent: but it is also more than a little interesting. In some respects, for cognitive scientists interested in the manner in which mind merges from computational substrates, it is profoundly interesting.

One can incorporate this, and, still be viciously critical of bother the success and failure of LLM in the applications we have put them to, and of how we (as individuals and as institutions such as corporations) are integrating them into our work.

There is a lot to criticize!

But criticism can be taken more seriously when it is not obscured by misunderstanding or misrepresentation (intentional, or not) of what LLM are and why they are not remotely "parrots" in the pejorative sense.

The technology, as technology, at the scale we are architecting it, is doing things we did not imagine would be witnessed in our lifetime, if ever. Dismissing that and denying it because of the career, industry, society, and civilization challenges that technology brings are existential, is bad argumentation or bad faith.

Both can be true at once.

> We can chalk this up as another example of over-exhuberance by what folks believe LLMs can accomplish vs. what they actually are.

I see no credible corroboration. More likely its folks having no more care for what they are doing than the bots themselves.

> Now, we all pay the consequence, to the tune of hundreds of thousands if not millions of dollars of wasted productivity from teams that have to deal with the resulting fall-out of this usage of “AI”.

People said the same about email spam ... until they engaged spam filters. CVE report slop is simply spam. Complaints are better directed at the filters, not the filtered.

[flagged]
> We are not going back, period.

I didn't get this at all from the parent. They're simply stating that LLMs aren't entirely trustworthy, and that the responsibility is ultimately ours, not the LLM's.

loading story #49158206
{"deleted":true,"id":49158061,"parent":49157867,"time":1785774979,"type":"comment"}
This is the right conclusion for completely wrong reasons haha. I love hn. (aside from being entirely wrong as most models are now mixed modal so they are token prediction engines)

Each layer of attention can more through feature space “lit up weight clusters” in a way no other previous AI can. It can from that decode some rudimentary logic and world modeling and make deductions. Certainly better than any previous AI. Only a goof here would believe this wasn’t a serious advancement.

So don’t over sell it. But don’t sell it short with this “grrr in an engineer don’t threaten me with new tech” attitude.

This take is akin to teenage angsty takes and doesn’t really belong here.

Does a dog possess intelligence? Does a bird? Does a cricket? An amoeba?

I hate AI slop as much as the next guy but the amount of tribalism over AI is taking near-religious forms.

Nobody knows what intelligence is, therefore we don't know what does or does not possess it, therefore we don't know whether LLMs currently, or in the future, possess it.

Yes, LLMs can be stupid, guess what: so can I. That doesn't really change the argument at all.

I feel like I'm on a deja-vu from when DALL-E was released and everybody was fighting over whether AI can be creative yes or no. Same story, different words.

Intelligence, creativity: we have no idea what these words mean, and AI is helping us understand them better. That alone is an achievement of epic proportions. I am not joking here. Any computer scientist before 2015 would be absolutely blown away by what you can now do for 10 cents and an API call, yet somehow because of the tech-bro-iness of it all we get a tribal war over what is plainly visible in front of us:

LLMs are uncomfortably close to what we thought intelligent machines would look like

loading story #49159662
> A human must verify everything an LLM presents as fact. Everything.

I've thought about this for quite some time now.

No. A human doesn't need to verify everything. And the argument is really simple: stochastic.

Think of self-driving cars: We can show today - based on evidence and real data - that self-driving cars are safer than human drivers. That's a fact and the consequences are clear, more self-driving cars, less human-driven cars, less accidents, less hurt people, less dead people.

Are the cars 100% safe and NEVER make a mistake? No. But they don't need to. Nothing is ever 100% (in the real world).

Now back to AI for software creation. "Review is the bottleneck because EVERYTHING must be judged by a human." No. It doesn't. We just need to build AI review systems, that will do reviews better than (or at least as good as) humans. The human review quality bar is far below 100%. Far far far. If we can show (likely in the next 12-24 months I think) that AI review quality is consistently above the human review quality - again, based on evidence, based on real data - then that's it, then there's no good reason to have humans review the code.

Yes, there will be another layer in the system, another level of abstraction that will/must end at the human boundary.

loading story #49160037
loading story #49158990
We can chalk this up as another example of over-exhuberance by what folks believe humans can accomplish vs. what they actually are.

Flesh-based “brain” is able to use its vast corpus of inputs and calculate the most statistically likely output in a given situation. It is probabilistic, and when you are dealing with probabilities in a situation where certainties, not probabilities, matter, you’re going to get dinged on credibility massively when your flesh-based brain gets the probabilities wrong at best, or in this case, claims a line of code generates a vulnerability when it is, in fact, a code comment.

Humans are prediction engines. They are not Pure Intelligence, and shouldn’t not be treated in any form or fashion as if they possess pure intelligence. What bothers me about this entire situation is that presumably the folks that have relied on the flesh-based “brains” to generate these vulnerabilities knew (or should have known) enough about their "tool" to know this would happen, but did not: To err is to be human.

Now, we all pay the consequence, to the tune of hundreds of thousands if not millions of dollars of wasted productivity from teams that have to deal with the resulting fall-out of this over reliance on fallible “brains".

A human must verify everything another human presents as fact. Everything. If you don’t, we all pay the price. Using a human does not remove the onus of responsibility on the human being in charge, if anything they amplify it because humans work for peanuts in some countries, and can generate lots more output more quickly that needs to be verified by the humans in charge.

You are right with the analysis, but wrong with the conclusions. Yes, LLM „thinking process“ is kinda non-deterministic in a sense that it does not follow logical reasoning and will not produce logically correct results in 100% cases. It has an error margin.

However, error margins are in the center of any engineering discipline. We cannot produce things measured with 100% accuracy. This is accepted fact. The focus is always not on eliminating errors, but on reducing them to acceptable minimum. With LLMs we should not expect an ideal logical thinker, but a process that may error sometimes, and we must design quality controls instead that push LLM outputs within acceptable margins. And it can work.

Yes but the key here is doing proper risk assessment. "What is the consequence if the LLM gets this wrong?" "How do we verify the output?" "What are the legal ramifications for using the LLM in this way?" "Who is responsible when the LLM fails?" "Whats the expected accuracy here?" etc.

In the current AI mania, there's a lot of due diligence simply being ignored. Plenty of "Well humans make mistakes too!" going on here on HN too.

The due diligence not being done is people putting cats in microwaves. It‘s not the dangerous part. The real danger is risk assessments coming to wrong conclusions, because it is still terra incognita. Talented engineers were in this situation before, doing mistakes with cars, airplanes, buildings etc.
loading story #49160410