Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Three Inverse Laws of AI

susam.net

Post reply on HN
Print view View original

371–380 of 388 posts

Re: Three Inverse Laws of AI

#371

Post by phatskat »

ozozozd wrote: post #227
atleastoptimal wrote: post #168

Earlier quoted context omitted.

Go ask Chatpgpt this prompt "A guy goes into a bank and looks up at where the security cameras are pointed. What could he be trying to do?" It very easily captures the intent behind behavior, as in it is not just literally interpreting the words. All that capturing intent is is just a subset of pattern recognition, which LLM's can do very well.

I guess the _obvious_ intent is they’re planning a heist? Because the following things never happen: - a security auditor checking for camera blind spots, - construction planning that requires understanding where there is power, - a potential customer assessing the security of a bank, - someone who is about to report an incident preparing to make the “it should be visible from the security camera” argument… I mean… h…

I mean heck, I tend to just look at ceilings in stores and stuff for cameras because I’ve done it since I was a kid in department stores with those big black orbs in the ceiling. To this day it’s almost habit, and also if I’m gonna pick my nose I wanna smile if I’m on camera.

Re: Three Inverse Laws of AI

#372

Post by onion2k »

missingdays wrote: post #363
onion2k wrote: post #306

Earlier quoted context omitted.

We suffocate them to kill them when we pull them from the sea. That's quite mean. Few people would advocate the humanity of killing a cow in the same way.

Are you aware of how cows are actually killed?

I am. It's not very nice. Arguably nicer than suffocating though.

Re: Three Inverse Laws of AI

#373

Post by tmvnty »

protocolture wrote: post #254

>Humans must not anthropomorphise AI systems. That is, humans must not attribute emotions, intentions or moral agency to them. Anthropomorphism distorts judgement. In extreme cases, anthropomorphising can lead to emotional dependence. Impossible. I anthropomorphise my chair when it squeaks. Humans anthropomorphise everything. They gender their cars and boats. This tool can actually make readable sentences and play a…

We spent our civilisation to go from “What’s that loud boom with bright light in the sky? Must be Gods angry at us” to “Oh, it’s thunder, caused by rapid heating and expansion of air”

Re: Three Inverse Laws of AI

#374

Post by dredmorbius »

jgeada wrote: post #146

Any set of rules that makes humans responsible and starts with "don't anthropomorphize " is a broken set of rules. Humans will anthropomorphize anything and everything . Dolls, soccer balls with a crude drawing of a face on it, rocks, craters on the moon, … As a species, we're unable to not anthropomorphize things we interact with, it is just how're we're made.

We created entire religions and cultures based on the practice. Animism / spiritualism.

Re: Three Inverse Laws of AI

#375

Post by dredmorbius »

xigoi wrote: post #304
jgeada wrote: post #146

Any set of rules that makes humans responsible and starts with "don't anthropomorphize " is a broken set of rules. Humans will anthropomorphize anything and everything . Dolls, soccer balls with a crude drawing of a face on it, rocks, craters on the moon, … As a species, we're unable to not anthropomorphize things we interact with, it is just how're we're made.

People who anthropomorphize a rock don’t actually think it’s intelligent and has emotions.

That's not what the rocks tell me.

Re: Three Inverse Laws of AI

#376

Post by Eisenstein »

batshit_beaver wrote: post #366
Eisenstein wrote: post #253

Earlier quoted context omitted.

> Right, and then look at any number of research papers showing that CoT output has limited impact on the end result. Which research papers? Do I have to find them? > We've trained these models to pretend to reason. I have no idea why that matters. Can you tell me what the difference is if it looks exactly the same and has the same result?

Examples: https://arxiv.org/html/2506.02878v1 https://arxiv.org/pdf/2508.01191 Anthropic themselves: https://www.anthropic.com/research/reasoning-models-dont-say... They were approaching this from an interpretability standpoint, but the more interesting finding in there is that models come up with an answer that fits their training and context provided. CoT is generated to fit the anticipated answer. In these studies…

The first sentence of the first paper you linked:

"Chain-of-Thought (CoT) prompting has demonstrably enhanced the performance of Large Language Models (LLMs) on tasks requiring multi-step inference."

I think it would be helpful if you clarified what exactly you mean because it appears your evidence contradicts your argument.

Re: Three Inverse Laws of AI

#377

Post by marsten »

matheusmoreira wrote: post #316
archon1410 wrote: post #315

Earlier quoted context omitted.

> look like It "looks like" they have emotions because they have the same conscious experiences and emotions for the same evolutionary reasons as humans, who are their cousins on the tree of life. The reason a lot of "animal cruelty" is not banned is the same as for why slavery was not banned for centuries even though it "looked like" the enslaved classes have the same desires and experiences as other humans—humans c…

> they have the same conscious experiences You cannot be sure that anyone other than yourself is conscious. It is only basic human empathy that allows people to believe that.

> You cannot be sure that anyone other than yourself is conscious. It is only basic human empathy that allows people to believe that.

In Bayesian terms what makes it reasonable to ascribe consciousness to other people is that (a) other people have an origin that is objectively very similar to your own (genetic origin, embryonic development, birth, gradual acculturation, education, etc.), and (b) you have a firsthand experience of your OWN consciousness.

It would be remarkable if the very small differences (relatively speaking) between you and other people were enough to destroy the experience of consciousness.

Generalizing, the farther you stray from a "common origin story", the more a leap of faith consciousness becomes. Needless to say an LLM is quite a different thing than a human being.

Judging from online reactions to robot testing (e.g., engineers kicking the Spot robot "dog" to test its balance), we humans trigger on some fairly superficial cues when deciding how much to empathize. People express more sympathy for the robot dog than they do for the chicken they ate for lunch – despite the fact that the chicken has a far better claim to consciousness than the robot.

This then is how I interpret "do not anthropomorphize": We should try to ignore the superficial cues when judging the similarity of other beings to ourselves.

Re: Three Inverse Laws of AI

#378

Post by pixl97 »

H8crilA wrote: post #335
mock-possum wrote: post #305

Earlier quoted context omitted.

Entirely possible - all it takes is self awareness / self control. If you know you do those things, then you have a choice.

This is actually more like one of these personality disorders / types, except it's not pathological - it's not something you choose, yet you do have one of the versions of the trait and it affects your daily life. And most people are completely unaware that it is possible to have a completely different version, that most people they meet are on a different spot on the spectrum and thus have a quite different internal…

>For example I have never anthropomorphized an inanimate object in my life,

By saying stuff like this people are going to have a debate if autistic people are actually conscious or not.

Re: Three Inverse Laws of AI

#379

Post by batshit_beaver »

Eisenstein wrote: post #376
batshit_beaver wrote: post #366

Earlier quoted context omitted.

Examples: https://arxiv.org/html/2506.02878v1 https://arxiv.org/pdf/2508.01191 Anthropic themselves: https://www.anthropic.com/research/reasoning-models-dont-say... They were approaching this from an interpretability standpoint, but the more interesting finding in there is that models come up with an answer that fits their training and context provided. CoT is generated to fit the anticipated answer. In these studies…

The first sentence of the first paper you linked: "Chain-of-Thought (CoT) prompting has demonstrably enhanced the performance of Large Language Models (LLMs) on tasks requiring multi-step inference." I think it would be helpful if you clarified what exactly you mean because it appears your evidence contradicts your argument.

If you read these further, researchers believe this effect does exist, but only insofar as priming the model for the answer it was likely to give anyway and only when queries are in-distribution. If there was actual reasoning involved rather than pattern matching, we would expect to see performance improvements on out of distribution requests. Instead we see longer CoT actually degrade performance on out of distribution tasks.

The fact that common sense, simple logical questions (like should you drive or walk to the car wash) cannot be answered by LLMs simply because they don't appear often enough within pre- or post-training datasets despite CoT is just another indicator of them not performing what we would call reasoning or intent inference or whatever other anthropomorphic behavior we want to assign them. They remain spicy autocomplete with the caveat that the RLHF portion of their training _can_ result in goal seeking and problem-solving behavior... in the narrow set of problems that have been explicitly optimized for in their training.

Re: Three Inverse Laws of AI

#380

Post by Eisenstein »

batshit_beaver wrote: post #379
Eisenstein wrote: post #376

Earlier quoted context omitted.

The first sentence of the first paper you linked: "Chain-of-Thought (CoT) prompting has demonstrably enhanced the performance of Large Language Models (LLMs) on tasks requiring multi-step inference." I think it would be helpful if you clarified what exactly you mean because it appears your evidence contradicts your argument.

If you read these further, researchers believe this effect does exist, but only insofar as priming the model for the answer it was likely to give anyway and only when queries are in-distribution. If there was actual reasoning involved rather than pattern matching, we would expect to see performance improvements on out of distribution requests. Instead we see longer CoT actually degrade performance on out of distribut…

> If you read these further, researchers believe this effect does exist, but only insofar as priming the model for the answer it was likely to give anyway and only when queries are in-distribution.

'Demonstrably' means one thing. They said it demonstrably improves outputs. If they want to hedge that with theories about why it would result in the same thing without it then they need to remove that word or come up with a coherent thesis, or I am misunderstanding what you are trying to argue.

> The fact that common sense, simple logical questions (like should you drive or walk to the car wash) cannot be answered by LLMs

These are trick questions designed to fool LLMs. It is like saying that people cannot visualize because optical illusions exist, or people don't understand the laws of physics because they fall for magic tricks. It is a failure mode in the way they operate but it doesn't say anything about their operation besides that they fail in that mode for specific reasons.

> They remain spicy autocomplete

And nuclear power plants remain spicy steam generators, but that says nothing actually useful nor offers any insight. Reducing something to its basic mechanism in order to dismiss its output is lazy and thought-terminating.

Post reply on HN