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

AGI is an engineering problem, not a model training problem

vincirufus.com

Post reply on HN
Print view View original

371–380 of 442 posts

Re: AGI is an engineering problem, not a model training problem

#371

Post by danenania »

taormina wrote: post #370
danenania wrote: post #193

Earlier quoted context omitted.

I see a lot of people saying things like this, and I’m not really sure which planet you all are living on. I use LLMs nearly every day, and they clearly keep getting better.

Grok hasn't gotten better. OpenAI hasn't gotten better. Claude Code with Opus and Sonnet I swear are getting actively worse. Maybe you only use them for toy projects, but attempting to get them to do real work in my real codebase is an exercise in frustration. Yes, I've done meaningful prompting work, and I've set up all the CLAUDE.md files, and then it proceeds to completely ignores everything I said, all of the con…

It can definitely be difficult and frustrating to try to use LLMs in a large codebase—no disagreement there. You have to be very selective about the tasks you give them and how they are framed. And yeah, you often need to throw away what they produced when they go in the wrong direction.

None of that means they’re getting worse though. They’re getting better; they’re just not as good as you want them to be.

Re: AGI is an engineering problem, not a model training problem

#372

Post by danenania »

Mikhail_Edoshin wrote: post #281
danenania wrote: post #230

Earlier quoted context omitted.

It’s fine to think that—many clearly do. But it would be more honest and productive imo if people would just say outright when they don’t think AGI is possible (or that AI can never be “real intelligence”) for religious reasons, rather than pretending there’s a rational basis.

When you try to solve a problem the goal or the reason to reject the current solution are often vague and hard to put in words. Irrational. For example, for many years the fifth postulate of Euclid was a source of mathematical discontent because of a vague feeling that it was way too complex compared to the other four. Such irrationality is a necessary step in human thought.

Yes, that’s fair. I’m not saying there’s no value to irrational hunches (or emotions, or spirituality). Just that you should be transparent when that’s the basis for your beliefs.

Re: AGI is an engineering problem, not a model training problem

#373

Post by HarHarVeryFunny »

x1f604 wrote: post #360
HarHarVeryFunny wrote: post #358

Earlier quoted context omitted.

I think the gist of TFA is just that we need a new architecture/etc not scaling. I suppose one can argue about whether designing a new AGI-capable architecture and learning algorithm(s) is a matter of engineering (applying what we already know) or research, but I think saying we need new scientific discoveries is going to far. Neural nets seems to be the right technology, and we've now amassed a ton of knowledge and…

> Demis Hassabis has recently given an estimate of human-level AGI in 5 years He said 50% chance of AGI in 5 years.

Yes, and I wonder how many "5 year" project estimates, even for well understood engineering endeavors, end up being accurate to within 50% (obviously overruns are more common then the opposite)?

I took his "50% 5-year" estimate as essentially a project estimate for something semi-concrete they are working towards. That sort of timeline and confidence level doesn't seem to allow for a whole lot of unknowns and open-ended research problems to be solved, but OTOH who knows if he is giving his true opinion, or where those numbers came from.

Re: AGI is an engineering problem, not a model training problem

#374

Post by JPLeRouzic »

galangalalgol wrote: post #173
Uehreka wrote: post #116

Earlier quoted context omitted.

Nah. The real philosophical headache is that we still haven’t solved the hard problem of consciousness, and we’re disappointed because we hoped in our hearts (if not out loud) that building AI would give us some shred of insight into the rich and mysterious experience of life we somehow incontrovertibly perceive but can’t explain. Instead we got a machine that can outwardly present as human, can do tasks we had thoug…

I think Metzinger nailed it, we aren't conscious at all. We confuse the map for the territory in thinking the model we build to predict our other models is us. We are a collection of models a few of which create the illusion of consciousness. Someone is going to connect a handful of already existing models in a way that gives an AI the same illusion sooner rather than later. That will be an interesting day.

That's what I am thinking too. Thanks for expressing it more clearly and concisely than I can.

Re: AGI is an engineering problem, not a model training problem

#375

Post by galangalalgol »

9dev wrote: post #356
prmph wrote: post #332

Earlier quoted context omitted.

Yes, but that is not an illusion. There's a reason I am perceiving something this was vs that other way. Perception is the most fundamental reality there is.

And yet that perception is completely flawed! The narrative part of your brain will twist your recollection of the past so it fits with your beliefs and makes you feel good. Your senses make stuff up all the time, and apply all sorts of corrections you’re not aware of. By blinking rapidly, you can slow down your subjective experience of time. There is no such thing as objective truth, at least not accessible to human…

When I used the word illusion, I meant the illusion of a self, at least a singular cohesive one as you are pointing out. It is an illusion with both utility and costs. Most animals don't seem to have meta cognitive processes that would give rise to such an illusion, and the ones that do are all social. Some of them have remarkably few synapses. Corvids for instance, we are rapidly approaching models the size of their brains and our models have no need for anything but linguistic processing, the visual and tactile processing burdens are quite large. An LLM is not like the models Corvids use, but given the flexibility to change it's own weights permanently, plasticity could have it adapt to unintended purposes, like someone with brain damage learning to use a different section of their brain to perform a task it wasn't structured for (though less efficiently).

Re: AGI is an engineering problem, not a model training problem

#376

Post by danenania »

crystal_revenge wrote: post #261
danenania wrote: post #206

Earlier quoted context omitted.

The idea that biological intelligence is impossible to replicate by other means would seem to imply that there’s something magical about biology.

I'm nowhere implying that it's impossible to replicate, just that LLMs have almost nothing to do with replicating intelligence. They aren't doing any of the things even simple life forms are doing.

They lack many abilities of simple life forms, but they can also do things like complex abstract reasoning, which only humans and LLMs can do.

Re: AGI is an engineering problem, not a model training problem

#377

Post by prmph »

9dev wrote: post #356
prmph wrote: post #332

Earlier quoted context omitted.

Yes, but that is not an illusion. There's a reason I am perceiving something this was vs that other way. Perception is the most fundamental reality there is.

And yet that perception is completely flawed! The narrative part of your brain will twist your recollection of the past so it fits with your beliefs and makes you feel good. Your senses make stuff up all the time, and apply all sorts of corrections you’re not aware of. By blinking rapidly, you can slow down your subjective experience of time. There is no such thing as objective truth, at least not accessible to human…

> The narrative part of your brain will twist your recollection of the past so it fits with your beliefs and makes you feel good.

But that's what I mean. Even if we accept that the brain has "twisted" something, that twisting is the reality. In order words, it is TRUE that my brain has twisted something into something else (and not another thing) for me to experience.

Re: AGI is an engineering problem, not a model training problem

#378

Post by andy99 »

galangalalgol wrote: post #173
Uehreka wrote: post #116

Earlier quoted context omitted.

Nah. The real philosophical headache is that we still haven’t solved the hard problem of consciousness, and we’re disappointed because we hoped in our hearts (if not out loud) that building AI would give us some shred of insight into the rich and mysterious experience of life we somehow incontrovertibly perceive but can’t explain. Instead we got a machine that can outwardly present as human, can do tasks we had thoug…

I think Metzinger nailed it, we aren't conscious at all. We confuse the map for the territory in thinking the model we build to predict our other models is us. We are a collection of models a few of which create the illusion of consciousness. Someone is going to connect a handful of already existing models in a way that gives an AI the same illusion sooner rather than later. That will be an interesting day.

  illusion
For who's benefit?

Re: AGI is an engineering problem, not a model training problem

#379

Post by prmph »

galangalalgol wrote: post #354
prmph wrote: post #331

Earlier quoted context omitted.

Let's be careful of creating different classes of consciousness, and declaring people to be on lower rungs of it. Sure, some aspects of consciousness might differ a bit for different people, but so long as you have never had another's conscious experience, I'd be wary of making confident pronouncements of what exactly they do or do not experience.

You can take their word for it, but yes, that is unreliable. I don't typically have an internal narrative, it takes effort. I sometimes have internal dialogue to think through an issue by taking various sides of it. Usually it is quiet in there. Or there is music playing. This is the most replies I have ever received. I think I touched a nerve by suggesting to people they do not exist.

I get you somewhat, but remember, you do not have another consciousness to compare with your own; it could be that what others call an internal narrative is exactly what you are experiencing; it just that they choose to describe it differently from you

Re: AGI is an engineering problem, not a model training problem

#380

Post by vjerancrnjak »

Uehreka wrote: post #116
hnuser123456 wrote: post #47

Earlier quoted context omitted.

Even more fundamental than science, there is missing philosophy, both in us regarding these systems, and in the systems themselves. An AGI implemented by an LLM needs to, at the minimum, be able to self-learn by updating its weights, self-finetune, otherwise it quickly hits a wall between its baked-in weights and finite context window. What is the optimal "attention" mechanism for choosing what to self-finetune with,…

Nah. The real philosophical headache is that we still haven’t solved the hard problem of consciousness, and we’re disappointed because we hoped in our hearts (if not out loud) that building AI would give us some shred of insight into the rich and mysterious experience of life we somehow incontrovertibly perceive but can’t explain. Instead we got a machine that can outwardly present as human, can do tasks we had thoug…

This is also true when conversing with other humans.

You can talk about your own spark of life, your own center of experience and you'll never get a glimpse of what it is for me.

At a certain level, thing you're looking at is a biological machine that can be described with constituents so it's completely valid you assume you're the center of experience and I'm merely empty, robotic, dead.

We might build systems that will talk about their own point of view, yet we will know we had no ability to materialize that space into bits or atoms or physics or universe. So from our perspective, this machine is not alive, it's just getting inputs and producing outputs, yet it might very well be that the robot will act from the immaterial space into which all of its stimuli appear.

Post reply on HN