News for hackers. Stuff that matters.
Re:Hot take: GPT 4.5 is a nothing burger by iainctduncan on Friday February 28, 2025 @07:00PM (#43213623)
Lots of people here hate Marcus, and he can lay it on thick, but I quite like him. He has actually studied human cognition and intelligence outside of programming – something that can't be said about a great many AI pundits.

I'm doing an interdisciplinary PhD in CS and Music right, so both music cognition and AI is in my program. From what I can see (which is not a lot of data, I admit) no one who has actually studied music cognition would think LLMs are going to lead to AGI by scaling. Linguistic thinking is only a small part of human intelligence, and certainly not the most complex. LLMs are so, so far from being able to the thinking that goes in a real-time musical improvisation context it's laughable.

Re:Hot take: GPT 4.5 is a nothing burger by klipt on Friday February 28, 2025 @10:22PM (#43215444)
Well music is arguably its own language. You wouldn't expect a monolingual English speaker to speak fluent Chinese. In the same way you can't expect a musically untrained person (or model) to "speak music".

But if you took an LLM size neural network and trained it on all the music in the world - I dare say you may get some interesting results.

Re:Hot take: GPT 4.5 is a nothing burger by seunosewa on Saturday March 01, 2025 @04:01AM (#43217458)
It's been done. We have AI music websites
Re:Hot take: GPT 4.5 is a nothing burger by perfmode on Saturday March 01, 2025 @08:59AM (#43219293)
such as?
Re:Hot take: GPT 4.5 is a nothing burger by captaindiego on Saturday March 01, 2025 @09:02AM (#43219313)
Give suno a try: https://suno.com/home
Re:Hot take: GPT 4.5 is a nothing burger by williamcotton on Saturday March 01, 2025 @09:13AM (#43219403)
Re:Hot take: GPT 4.5 is a nothing burger by dumpsterdiver on Sunday March 02, 2025 @03:18PM (#43234587)
Re:Hot take: GPT 4.5 is a nothing burger by mountainriver on Friday February 28, 2025 @11:24PM (#43215870)
We aren't very far away from that, if you think we are you should look closer
Re:Hot take: GPT 4.5 is a nothing burger by nyolfen on Friday February 28, 2025 @11:47PM (#43216027)
> LLMs are so, so far from being able to the thinking that goes in a real-time musical improvisation context it's laughable.

have you actually tried any of the commercial AI music generation tools from the last year, eg suno? not an LLM but rather (probably) diffusion, it made my jaw drop the first time i played with it. but it turns out you can also use diffusion for language models https://www.inceptionlabs.ai/

Re:Hot take: GPT 4.5 is a nothing burger by iainctduncan on Friday February 28, 2025 @11:53PM (#43216064)
This is exactly my point. The idea that generating music is the sum total of all the cognition that goes on in real time is totally off base. Can we make crappy simulacrums of recorded music? sure.

Can we make a program that could: work the physical instrument, react in milliseconds to other players, create new work off input it's never heard before, react to things in the room, and do a myriad of other things a performer does on stage in real time? Not even remotely close. THAT is what would be required to call it AGI – thinking (ALL the thinking) on par with a highly trained human. Pretending anything else is AGI is nonsense.

Is it impressive? sure. Frighteningly so even. But it's not AGI and the claims that it is are pure huckersterism. They just round the term down to whatever the hell is convenient for the pitch.

Re:Hot take: GPT 4.5 is a nothing burger by nyolfen on Saturday March 01, 2025 @12:16AM (#43216215)
yes, i absolutely believe we are less than two years from what you describe (aside from physically manipulating an instrument -- but robotics seems to be quickly picking up pace too). what you are imagining is only a difference in speed, not kind. this is the 'god of the gaps' argument, over and over, every time some insurmountable previous benchmark is shattered -- well, it will never be able to do my special thing

> thinking (ALL the thinking) on par with a highly trained human

you are mistaking means for ends

Re:Hot take: GPT 4.5 is a nothing burger by kranke155 on Saturday March 01, 2025 @03:16AM (#43217186)
We are likely more than 10 year away. The algorithms haven’t been found yet I think.
Re:Hot take: GPT 4.5 is a nothing burger by namaria on Saturday March 01, 2025 @03:29AM (#43217263)
There is no god of the gaps argument being made. The argument is pretty clear. LLMs are at a local optimum that is quite far from the global optimum of actual general intelligence that humans have achieved.

Some like Penrose even argue that the global optimum of general intelligence and consciousness is a fully physical process, yes, but that involves uncomputable physics and thus permanently out of reach of whatever computers can do.

Re:Hot take: GPT 4.5 is a nothing burger by akoboldfrying on Saturday March 01, 2025 @04:02AM (#43217467)
>but that involves uncomputable physics and thus permanently out of reach of whatever computers can do.

And yet is somehow within reach of a fertilised human egg.

It's time to either invoke mystical dimensions of reality separating us from those barbarian computers, or admit that one day soon they'll be able to do intelligence too.

Re:Hot take: GPT 4.5 is a nothing burger by namaria on Saturday March 01, 2025 @04:06AM (#43217487)
There's no magical thinking involved in discussing the limits of computability. That is a well researched area that was involved in the invention of digital computers.

Penrose's argument is interesting and I am inclined to agree with it. I might very well be wrong, but I don't think the accusation of magical thinking is warranted.

Re:Hot take: GPT 4.5 is a nothing burger by akoboldfrying on Saturday March 01, 2025 @04:15AM (#43217536)
I'm arguing that, if a fertilised egg is really capable of fundamentally more than computers will ever be, then the only possible explanation is that the egg posseses extraphysical properties not possessed by any computer. (And I'm strongly hinting that this "explanation" should be considered laughable in this day and age.)
Re:Hot take: GPT 4.5 is a nothing burger by namaria on Saturday March 01, 2025 @04:18AM (#43217547)
> the only possible explanation is that the egg posseses extraphysical properties

This is wrong. Computability is by no means the same as physicality. That's the whole point and you're just ignoring it to make some strawman accusation of ridiculousness.

Re:Hot take: GPT 4.5 is a nothing burger by akoboldfrying on Saturday March 01, 2025 @04:21AM (#43217562)
1. What property of a fertilised egg enables it to compute uncomputable things?

2. If that property is a physical property, what prevents simulation of it?

Re:Hot take: GPT 4.5 is a nothing burger by namaria on Saturday March 01, 2025 @04:28AM (#43217599)
> 1. What property of a fertilised egg enables it to compute uncomputable things?

The question is malformed. How can something compute an uncomputable thing?

> 2. If that property is a physical property, what prevents simulation of it?

Is P=NP?

Re:Hot take: GPT 4.5 is a nothing burger by akoboldfrying on Saturday March 01, 2025 @10:33PM (#43226984)
>How can something compute an uncomputable thing?

If humans cannot compute uncomputable things either, on what grounds do you claim we are capable of something that computers are incapable of?

Re:Hot take: GPT 4.5 is a nothing burger by namaria on Sunday March 02, 2025 @02:40AM (#43228269)
Haven't you understood that my argument is precisely that intelligence comprises more than performing computations?

I know you think this is a gotcha moment so I will just sing off on this note. You think physical = computable. I think physical > computable. I understand your argument and disagree with it but you can't seem to understand mine.

Re:Hot take: GPT 4.5 is a nothing burger by akoboldfrying on Sunday March 02, 2025 @03:27AM (#43228503)
It is completely unclear what you think the difference in capability between humans and computers is.

I've tried to follow your reasoning, which AFAICT comes down to a claim that humans possess something connected to incomputability, and computers do not. But now it seems you hold this difference to be irrelevant.

So again: What do you think the difference in capability between humans and computers is?

Re:Hot take: GPT 4.5 is a nothing burger by weiliddat on Saturday March 01, 2025 @04:53AM (#43217737)
A fertilized egg doesn’t automatically become able to compute stuff.

Understimulated or feral children don’t automatically become geniuses when given more information.

It takes social engineering and tons of accumulated knowledge over the lifespan of the maturation of these eggs. The social and informational knowledge are then also informed by these individuals (how to work and cooperate with each other, building and discovering knowledge beyond what a single fertilized egg is able to do).

This isn’t simply within reach of a fertilized egg based on its biological properties.

Re:Hot take: GPT 4.5 is a nothing burger by Xmd5a on Saturday March 01, 2025 @08:37AM (#43219101)
I used to believe this but changed my mind when I learned about that Brazilian orphanage for deaf kids. They were kinda left on their own and in the end developed their own signed language.
Re:Hot take: GPT 4.5 is a nothing burger by weiliddat on Saturday March 01, 2025 @11:30AM (#43220911)
That was cool to read about. I don't think we're disagreeing here. Humans (fertilized eggs) have many needs and interactions that give rise to language itself.

Current LLMs seem to be most similar to the linguistic/auditory portion of our cognitive system, and with thinking/reasoning models some bits of the executive function. But my guess is that if we want to see awe-inspiring stuff come out of them, we need stuff like motivation and emotion, which doesn't seem to be the direction we're heading towards.

Unprofitable, full of problems. Maybe 1 in 100,000 might be an awe-inspiring genius, given the right training, environment, and other intelligences (so you might have to train way more than 100K models).

Re:Hot take: GPT 4.5 is a nothing burger by schwartzworld on Saturday March 01, 2025 @07:53AM (#43218791)
> aside from physically manipulating an instrument

This is the easiest part. Keyboards are only necessary if you have fingers, an AI could very easily send midi notes directly to the instrument.

Re:Hot take: GPT 4.5 is a nothing burger by contrast on Saturday March 01, 2025 @03:29AM (#43217259)
“Pretending anything else is AGI is nonsense”

Is it? I wasn’t aware of “playing a physical instrument on stage with millisecond response times” as a criterion. I’m also confused by the implication that professional composers aren’t using intelligence in their work.

You’re talking about what is sometimes called “superhuman AGI”, human level performance in all things. But AGI includes reaching human levels of performance across a range of cognitive tasks, not ALL cognitive tasks.

If someone claimed they had invented AGI because amongst other things, it could churn out a fresh, original, good composition the day after hearing new input - I think it would be fair to argue that is human level performance in composition.

Defining fresh, good, original is what makes it composition. Not whether it was done in real time; that’s just mechanics.

You can conceivably build something that plays live on stage, responding to other players, creating a “new work”, using super fast detection and probabilistic functions without any intelligence at all.

Re:Hot take: GPT 4.5 is a nothing burger by refactor_master on Saturday March 01, 2025 @04:44AM (#43217680)
https://youtu.be/Gs3ocG5yW88

This has aged well and continues to do so.

Re:Hot take: GPT 4.5 is a nothing burger by borgdefenser on Saturday March 01, 2025 @06:25AM (#43218243)
I have tried them all. They are so unimpressive.

It is not the fault of the model though. MusicLM shows what could be done.

The problem is they just aren't trained on enough interesting music to impress me.

Of course, if you never played music before I am sure it is super cool to produce music.

It would be like if the AI art models had only been trained on a small amount of drawings.

Compared to being trained on every piece of recorded music ever produced?We are just so far from that.

I think to really do something interesting you would have to train your own model.

Re:Hot take: GPT 4.5 is a nothing burger by BeFlatXIII on Saturday March 01, 2025 @12:37PM (#43221541)
They're breathtaking party tricks when you first encounter them, but the similarity in outputs soon becomes apparent. Good luck coercing them to make properly sad or angsty songs.
Re:Hot take: GPT 4.5 is a nothing burger by JohnKemeny on Saturday March 01, 2025 @01:49AM (#43216641)
I, for one, agree with you, and my area is cognitive science and artificial intelligence. It's staggering to see the hive-mind of Sam Altman followers chanting about AGI in a Markov chain.

Yes, we can today make natural-sounding hyper-average text, images, and perhaps even music. But that's not what writing and creating is about.

Re:Hot take: GPT 4.5 is a nothing burger by kranke155 on Saturday March 01, 2025 @03:26AM (#43217250)
As a writer, people thinking about LLms writing a book-length story don’t realise it’s in the wrong problem space.

LLMs can solve equations and write code, like make a React web page.

So they assume “a story is just an equation, we just need to get the LLM to write an equation”. Wrong imo.

Storytelling is above language in the problem space. Storytelling can be done in multiple methods - visual (images), auditory (music), characters (letters and language) and gestural (using your body, such as a clown or a mime, Im thinking of Chaplin or Jacques Tati). That means storytelling exists above language, not as a function of it. If you agree with the Noah Harari vision of things, which I do, then storytelling is actually deeply connected somehow to our level of awareness and it seems to emerge with consciousness.

Which means that thinking that an LLM that can understand story because it understands language is… foolish.

Storytelling lives in a problem space above language. Thinking you can get an LLM to tell a story because you can get it to write a sentence is a misunderstanding of what problem space stories are in.

It’s the same as thinking that if an LLM can write a React web app it can invent brand new and groundbreaking computer science.

It can’t. Not for now anyway.

Re:Hot take: GPT 4.5 is a nothing burger by namaria on Saturday March 01, 2025 @03:31AM (#43217277)
I agree. Language manipulation is a projection of something higher order. LLMs can shuffle the projection but they clearly cannot go over the gap to the higher order thinking.
Re:Hot take: GPT 4.5 is a nothing burger by fooker on Saturday March 01, 2025 @04:06PM (#43223679)
> AGI in a Markov chain

Do you understand why modern LLMs are different from Markov chains?

Re:Hot take: GPT 4.5 is a nothing burger by echoangle on Saturday March 01, 2025 @05:50PM (#43224723)
I actually don’t, can you explain it?

Doesn’t the AI model basically take an input of context tokens and return a list of probabilities for the next token, which is then chosen randomly, weighted by the probability? Isn’t that exactly the definition of a Markov chain?

Re:Hot take: GPT 4.5 is a nothing burger by fooker on Saturday March 01, 2025 @07:08PM (#43225587)
You have almost answered the question.

LLMs basically return a Markov chain every single time. Think of it as a function returning a value vs returning a function.

Now, I'm sure a sufficiently large Markov chain can simulate an LLM but the exponentials involved here would make the number of atoms in the universe a small number. The mechanism that compresses this down into a manageable size is famously 'attention is a you need.!'

Re:Hot take: GPT 4.5 is a nothing burger by JohnKemeny on Sunday March 02, 2025 @06:18AM (#43229314)
> Now, I'm sure a sufficiently large Markov chain can simulate an LLM but the exponentials involved here would make the number of atoms in the universe a small number.

No, LLMs are a Markov chain. Our brain, and other mammalian brains, have feedback, strange loops, that a Markov chain doesn't. In order to reach reasoning, we need to have some loops. In that way, RNNs where much more on the right track towards achieving intelligence than the current architecture.

Re:Hot take: GPT 4.5 is a nothing burger by fooker on Sunday March 02, 2025 @07:34PM (#43236956)
A modern reasoning model is exactly a feedback loop.
Re:Hot take: GPT 4.5 is a nothing burger by seanhunter on Sunday March 02, 2025 @10:53AM (#43231653)
They have the Markov property that the next state depends only on the current state (context window plus the weights of the model) do they not? Any stochastic process which possesses that property is a Markov process/chain I believe.
Re:Hot take: GPT 4.5 is a nothing burger by fooker on Sunday March 02, 2025 @07:35PM (#43236972)
The universe has that property too.

But no, most LLMs have a tweakae 'temperature' parameter that introduces some randomness and sometimes have very interesting results.

Re:Hot take: GPT 4.5 is a nothing burger by seanhunter on Sunday March 02, 2025 @11:36PM (#43238346)
Temperature just changes how an llm samples from the token distribution so I wouldn’t think that constitutes “memory” that would contradict the Markov property
Re:Hot take: GPT 4.5 is a nothing burger by j_bum on Saturday March 01, 2025 @10:41AM (#43220312)
What a wonderful blend of fields for your PhD.

I tend to agree with you, but I would be cautious in stating that an understanding of music is required for AGI. Plenty of humans are tone deaf and can’t understand music, yet are cognitively capable.

I’d love to hear about your research. I have my PhD in neuroscience and a undergrad degree in piano performance. Not quite the same path as you, but similar!

Re:Hot take: GPT 4.5 is a nothing burger by llamajams on Monday March 03, 2025 @09:25AM (#43242035)
Hello, I am interested in your area of research. Would you mind linking to some of your work?