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

A non-anthropomorphized view of LLMs

addxorrol.blogspot.com

Post reply on HN
Print view View original

421–430 of 432 posts

Re: A non-anthropomorphized view of LLMs

#421

Post by donkeybeer »

godelski wrote: post #420
donkeybeer wrote: post #418

Earlier quoted context omitted.

What is indeterminism here? I am not sure the question having or not having free will has any impact of how to make human machines. We are just as in the dark about the future, if we have free will or not. I am not certain of any physical problem in which free will or lack of it plays a role. I could be wrong. So its probably an interesting question but rather pointless. Even with the everyday machines and programs w…

I think you missed my point. I agree, being a machine doesn't mean you don't have free will. I agree, free will is orthogonal to being (or not being) a machine. But that's not what my comment was about. My comment was about * what the average person interprets* . You asked why people take offense to being called a machine, and I'm trying to explain that. But to understand this we have to understand that there isn't a…

Oh of course. I feel I should have been clearer that I meant among technical persons, not average randoms.

When talking about people who are not mathematicians or computer scientists, on average, yes absolutely they hear something like that when told humans are machines.

Re: A non-anthropomorphized view of LLMs

#422

Post by chaps »

perching_aix wrote: post #390
chaps wrote: post #373

Earlier quoted context omitted.

So what? :)

Nothing? """:)""" Was just confusing because your phrasing implied different. https://en.wikipedia.org/wiki/Cooperative_principle

My real point was to emphasize that play is important because it expands our world of interfaces, whatever those interfaces may be, and however far those interfaces go.

Your antagonistic attitude is not productive nor playful. Lighten up, friend.

Re: A non-anthropomorphized view of LLMs

#423

Post by perching_aix »

chaps wrote: post #422
perching_aix wrote: post #390

Earlier quoted context omitted.

Nothing? """:)""" Was just confusing because your phrasing implied different. https://en.wikipedia.org/wiki/Cooperative_principle

My real point was to emphasize that play is important because it expands our world of interfaces, whatever those interfaces may be, and however far those interfaces go. Your antagonistic attitude is not productive nor playful. Lighten up, friend.

We're veering extremely off topic, but I don't think you're aware and maybe you should be: my antagonistic attitude is 100% a reaction to you being overly friendly. Calling me your "friend" repeatedly without any prior acquaintance comes across as remarkably sarcastic and condescending. The "So what? :)" wasn't any less condescending either as you may imagine, in fact, quite the opposite.

I'm sure there are people who can always reliably respond with kindness despite being handed condescension. As you can probably tell by now, I'm not one of those.

Re: A non-anthropomorphized view of LLMs

#424

Post by quonn »

elliotto wrote: post #369
quonn wrote: post #152

Earlier quoted context omitted.

> To claim that LLMs do not experience consciousness requires a model of how consciousness works. Nope. What can be asserted without evidence can also be dismissed without evidence. Hitchens's razor. You know you have consciousness (by the very definition that you can observe it in yourself) and that's evidence. Because other humans are genetically and in every other way identical, you can infer it for them as well.…

It seems like you are doing a lot of inferring about mammals experiencing consciousness, and you have drawn a line somewhere beyond these, and made the claim that your process is scientific. Could I present you my list of questions I presented to the OP and ask where you draw the line, and why here? My general list of questions for those presenting a model of consciousness are: 1) Are you conscious? (hopefully you sa…

I think I already answered those above. I draw the line between 3 and 4, possibly between 4 and 5. I don't know for sure. But there are good reasons to hold this belief.

> the Buddhists call this nirvana, and it's quite difficult to actually achieve.

Not really. The zen buddhists call what I described kensho and it's not very hard to achieve. I specifically said a few seconds. Probably anyone who wholeheartedly meditated for some time has experienced this.

Nirvana, on the other hand, is just the other side of practice-and-enlightenment as a drawn out process. You may call it hard to achieve, others may call it the dharma gate of ease and joy.

Re: A non-anthropomorphized view of LLMs

#425

Post by chaps »

perching_aix wrote: post #423
chaps wrote: post #422

Earlier quoted context omitted.

My real point was to emphasize that play is important because it expands our world of interfaces, whatever those interfaces may be, and however far those interfaces go. Your antagonistic attitude is not productive nor playful. Lighten up, friend.

We're veering extremely off topic, but I don't think you're aware and maybe you should be: my antagonistic attitude is 100% a reaction to you being overly friendly. Calling me your "friend" repeatedly without any prior acquaintance comes across as remarkably sarcastic and condescending. The "So what? :)" wasn't any less condescending either as you may imagine, in fact, quite the opposite. I'm sure there are people wh…

I really enjoyed our conversation. I hope you have a better summer, friend!

Re: A non-anthropomorphized view of LLMs

#426

Post by seadan83 »

ImHereToVote wrote: post #300
seadan83 wrote: post #132

Earlier quoted context omitted.

I agree with you @orbital-decay that I also do not get the same vibe reading this thread. Though, while human intelligence is (seemingly) not magic, it is very far from being understood. The idea that a LLM is comparable to human intelligence implies that we even understand human intelligence well enough to say that.

LLMs are also not understood. I mean we built and trained them. But don't of the abilities at still surprising to researchers. We have yet to map these machines.

I do partially agree. Though, it is at least tractable to understand why a LLM gave a specific output - but perhaps not practical. Understanding how a human arrives at a certain decision (by say simply looking at brain waves) OTOH is not even tractable as of yet.

Re: A non-anthropomorphized view of LLMs

#427

Post by seadan83 »

ants_everywhere wrote: post #262
seadan83 wrote: post #130

Earlier quoted context omitted.

> It is similar to how human brains operate. Is it? Do we know how human brains operate? We know the basic architecture of them, so we have a map, but we don't know the details. "The cellular biology of brains is relatively well-understood, but neuroscientists have not yet generated a theory explaining how brains work. Explanations of how neurons collectively operate to produce what brains can do are tentative and in…

The part I was referring to is captured in "The cellular biology of brains is relatively well-understood" Fundamentally, brains are not doing something different in kind from ANNs. They're basically layers of neural networks stacked together in certain ways. What we don't know are things like (1) how exactly are the layers stacked together, (2) how are the sensors (like photo receptors, auditory receptors, etc) hooke…

Okay, I understand your point I think. The analogy does not sit great with me. The cellular biology seems to me to be one of the most basic aspects of even starting to understand how the brain actually works. It's all mechanics without any semantics. As an extreme analogy, it's almost like saying leaves and money work the same way - they're both thin & green.

IMO, the analogy of how neural networks work as does the cellular biology is only useful as an analogy. They are not actually similar, and certainly the higher order functionality is not similar (namely because we don't know how that higher order functionality works, like we don't understand it _at_ _all_)

Re: A non-anthropomorphized view of LLMs

#428

Post by seadan83 »

Timwi wrote: post #171
seadan83 wrote: post #130

Earlier quoted context omitted.

> It is similar to how human brains operate. Is it? Do we know how human brains operate? We know the basic architecture of them, so we have a map, but we don't know the details. "The cellular biology of brains is relatively well-understood, but neuroscientists have not yet generated a theory explaining how brains work. Explanations of how neurons collectively operate to produce what brains can do are tentative and in…

> > It is similar to how human brains operate. > Is it? This is just a semantic debate on what counts as “similar”. It's possible to disagree on this point despite agreeing on everything relating to how LLMs and human brains work.

I was hoping it would be more than a semantic debate. The author of the article makes the point that humanity currently has no real clue how the human brain actually "works."

> agreeing on everything relating to how LLMs and human brains work

Hence, my question is "how does the human brain work?" Nobody really knows, therefore we can't know if LLMs actually work in a similar way. It's a big point of the author in the paper, it's a big reason why it is so inappropriate to anthropomorphize LLMs.

Re: A non-anthropomorphized view of LLMs

#429

Post by grey-area »

losvedir wrote: post #265
grey-area wrote: post #120

Earlier quoted context omitted.

On the contrary, anthropomorphism IMO is the main problem with narratives around LLMs - people are genuinely talking about them thinking and reasoning when they are doing nothing of that sort (actively encouraged by the companies selling them) and it is completely distorting discussions on their use and perceptions of their utility.

Well "reasoning" refers to Chain-of-Thought and if you look at the generated prompts it's not hard to see why it's called that. That said, it's fascinating to me that it works (and empirically, it does work; a reasoning model generating tens of thousands of tokens while working out the problem does produce better results). I wish I knew why. A priori I wouldn't have expected it, since there's no new input. That means…

It looks superficially like reasoning, but is it reasoning?

Re: A non-anthropomorphized view of LLMs

#430

Post by jibal »

ben_w wrote: post #406
jibal wrote: post #289

Earlier quoted context omitted.

Point out to an LLM that it has no mental states and thus isn't capable of being frustrated (or glad that your program works or hoping that it will, etc. ... I call them out whenever they ascribe emotions to themselves) and they will confirm that ... you can coax from them quite detailed explanations of why and how it's an illusion. Of course they will quickly revert to self-anthropomorphizing language, even after pr…

While they will agree with you that they don't, it's also the case that this is via training. Consider that we have recordings of Brent Spiner covered in white paint and wearing yellow contact lenses claiming to have no emotions, not because he didn't, but because he was playing a role, which is also something we know LLMs can do. So we don't know for sure if LLMs do or don't have qualia, irregardless of what they sa…

"While they will agree with you that they don't, it's also the case that this is via training."

Um, that's what I said.

And of course we know that LLMs don't have qualia. Heck, even humans don't have qualia: https://web.ics.purdue.edu/~drkelly/DennettQuiningQualia1988...

Post reply on HN