[Leynos]

While there are some coding focused models (composer, for example), the majority of frontier models are pitched as general purpose. The coding harnesses for Claude and GPT are even being repurposed as general purpose knowledge work harnesses.

[YeGoblynQueenne]

No, you're right of course, but I have a feeling it's much easier to sell a system with a clear goal, like "this LLM generates code" or "this LLM solves math problems". Even if the underlying model is a general purpose one. I think there's always a question, when one has a product, of "what does it do?". "This thing does everything you want it to" is not a great way to sell something.

More to the point, even models marketed as general purpose are clearly trained on specific tasks. That is, the AI companies want to promote their systems as general purpose but they also want to make them good at specific tasks, because that supports their marketing story, that those are general-purpose systems that are so powerful they can even do maths and science.

Or maybe code, maths, and (maybe) science are just the things they find it easy to train their models on, for different reasons and in different ways. You can also get a feel about the things they have tried to do and failed, e.g. real-world autonomy isn't really working (or not working yet, who knows), so OpenAI is not trying to sell an embodied generalist agent say, that can clean your hose, do your laundry and teach your kids maths and science on the side.

I mean even generating text is a very narrow task, in the general sense, compared to all the things that humans can do (never mind other animals) it's just that we use text so much and for so many things that there's an endless list of applications for a text generator; as we have all found out.

[red75prime]

> real-world autonomy isn't really working

It works in limited ways (but in the realest-real world nevetheless). Waymo, Wayve, Baidu Apollo, Tesla and others seem to rely on VLAs, VLMs or transformer models in general to do autonomous driving.

[YeGoblynQueenne]

Unfortunately those don't really work:

https://youtu.be/C4NQNeSO2vs?si=epkxhVXpypOCppGW

Also, none of those companies' cars are really autonomous. Waymo, for example, relies on remote workers that are ready to intervene and suggest a course of action when the AI driver gets stuck:

https://waymo.com/blog/2024/05/fleet-response/

And the point is that the AI driver gets stuck because it can't understand the situation it is in. That's not autonomy. Not yet.

[red75prime]

Non-trivial percentage of people work in the real world until they do something stupid and work no more. It's not real autonomy, not yet.

It's a matter of degree. Sorry, I don't want to watch an hour long video to be told how something that works 99.99% of the time doesn't really work for some contrived definition of "really".

BTW, as is typical with people, remote operators occasionally cause problems.

[YeGoblynQueenne]

Fair if you don't want to watch the entire video. The VLA stuff is early in the video but I can summarise it for you (if you trust me to do so): basically none of the current deep learning based approaches to autonomy generalise. Not RL, not transformers, not anything else. They all tend to work fine in environments and tasks within their training set but outside of it, pffft, performance evaporates. That's my summary.

It's really not about 99,9% correct. Rather that's the error you can expect when you deploy such systems to the real world. You can start building an intuition about this if you consider the combinatorial space that these systems must search to find the right action in a given situation. Most of those systems are trained on images i.e. sets of pixels (some also have more sensors like lidar which just blows up the combinatorics even more). Specifically what these systems learn is a function mapping a set of pixels to actuator commands, essentially mapping input images to actions. For an RBG image of 128 x 128, there's (256^3)^(127^2) = 16,777,216 ^ 16,384 = 6.468074e+118369 unique combinations of pixels. Each of those has to be mapped to one of k actuator commands where k is sometimes an integer, sometimes a real. As you can see, that's an insanely high number and there is no way to make a dent in it even with millions of examples of images-to-actions.

Now of course deep learning approaches have shown remarkable, stunning ability in powering through gigantic combinatorial spaces - but all those successful applications are in domains where it is not necessary to map the entire space of inputs to outputs, not even a big chunk of it. Consider the success in board games like chess and Go both of which have huge search spaces, as often pointed out with cardinalities larger than the number of atoms in the universe etc. Yes, but an automated system doesn't have to search that entire space to beat a human: like the joke with the two guys trying to outrun a bear, all the deep learning system need do is search further than a human can search. Or, consider image classification or language generation: even 20% error (which is what you get in real world situations, as opposed to "in the lab" testing on specific datasets like ImageNet) is OK a lot of the time. So what if your image classifier thinks a cat is an impala? Nobody dies. Mostly.

That's not the case in the real world where a system can be expected to continuously find itself in situations that are either subtly or wildly different than the ones in its training set and the stakes are very high. 20% error in that setting doesn't work. If the robot can't deal with a previously unseen situation it can very well destroy itself, and cause an unknown amount of damage besides. The real world is as unforgiving as it is unpredictable; and non-ergodic (i.e. trying stuff at random until you find the thing that works, doesn't work).

Self-driving cars are ultimately made to work with very careful, good, old-fashioned manual engineering, including detailed mapping of an area where the system is restricted etc. That is why e.g. you see Waymo rolling out its cars slowly from one city to the next: if their systems were really autonomous they could just deploy them anywhere a human driver can drive. From New York to Mumbai and from Athens to London. The fact they can't is the strongest evidence that whatever proprietary secret sauce moat they have... doesn't really work.

The "really" I keep dropping in there all the time is suspicious, I know, but that's not entirely my fault. Waymo and all the others claim that their systems are "autonomous". I have to bat that back by saying, no, they're not autonomous. Then there's going to be a lengthy, pedantic discussion about what is "autonomous". So I try to avoid that, in my opinion unproductive, exchange with the "really" hedge. Maybe not such a good idea. I'm open to suggestions to improve the clarity of my language.

[red75prime]

> So what if your image classifier thinks a cat is an impala?

For starters, the artificial neural network is the only instrument we have for sufficiently robust image classification. If it doesn't work nothing else will. We can only validate and retrain.

Getting back to self-driving. What if an occupancy network misclassify an impassable surface patch as passable? It would mean a possible accident, so you bias the network towards impassable classification. In this way misclassification is overwhelmingly more likely to cause spurious braking, unneeded swerve, and the like rather that a collision ("brake stabbing" was a recurring problem in Tesla FSD. Today it is largely mitigated.)

> The fact they can't is the strongest evidence that whatever proprietary secret sauce moat they have... doesn't really work.

Well, Waymo has publicly embraced world models, VLMs and end-to-end backpropagation relatively recently (December 9, 2025)[1]. There's not enough data yet to judge how this has influenced their expansion rate. Waymo representatives tout improved generalization (sorry, I can't find the interview. I think it was Dmitri Dolgov.)

> if their systems were really autonomous they could just deploy them anywhere a human driver can drive

As I've said it's a matter of degree. Human-caused road casualties are a familiar and largely inevitable evil. An autonomous vehicle killing a person will (and already has) cause a media storm. They can't ignore a potentially problematic distribution shift caused by deployment in a new area.

You are talking about the perfect system. Real systems are computationally constrained. Distillation necessarily degrades performance and might require retraining for particular areas.

[1] https://waymo.com/blog/2025/12/demonstrably-safe-ai-for-auto...

[red75prime]

> https://youtu.be/C4NQNeSO2vs?si=epkxhVXpypOCppGW

"VLAs fine-tuned on human demonstrations overfit. Here's how to mitigate it." Er, OK, I guess. But I think that Waymo uses mitigations or a different approach (RL, for example).

[YeGoblynQueenne]

Unfortunately the mitigations in the video don't work either.

The bottom line is that none of the techniques we have have been shown to work, to "really" really work, in the real world. If they did, you'd know instantly because suddenly robot cars and robot maids would be absolutely everywhere, not just in a few cities in the States or China. The Chinese in particular are chomping at the bit to flood the market with cheap robot labour. They can't because there's no such thing, yet. Not because there aren't cheap robots: because there's not the AI to drive them.

I know it's hard to get a good feel for where the state of the art is currently, but try this exercise. Back in 2017, if you were watching the video of Atlas doing backflips [1], how far would you think we are from robotic autonomy? Most people would have said "very close" and indeed many did. Yet, 9 years later we still don't have robot maids and robot butlers.

Atlas, of course, was hand-coded, but that doesn't make any difference. Its programming was good enough for backflips (and many other really impressive feats besides; e.g. much more impressive was seeing it balance itself after stepping on a length of pipe, but I can't find the video now). Just because there's a trend now to do the same coding by machine learning rather than by hand, doesn't change anything. The point is we don't know how to create the right program for a robot to act autonomously, independently, freely, and safely, in the real world.

Autonomy is hard and it goes way beyond anything anyone has demonstrated in the last... I'd say 50 years. Robotics has a long history of very impressive results that never turn out as useful as people imagine.

_________________

[1] https://www.youtube.com/watch?v=DqdaT9LI5oE