I'd point to Linus' Law, Given enough eyes, all bugs are shallow. (Wikipedia notes the law's validity is confirmed empirically, but refuted in analysis.) This applies to more than software. Right now, election enthusiasts in the US are combing election law and procedure for ways the November election can be thrown, and ways to counter that scenario quickly.

I was thinking about the misprogramming scenario of AI escape. This implies hyperintelligent AI is dangerous even when everyone using it is benign and has non-adversarial intentions. Essentially an AI will need to be coded with an ethics system such as that of a general (or high-ranking soldier in the US) who deliberates on operational action and its consequences, whether the operation is legal, or would cause undue harm in its execution.

Right now, AI in its current development is... whimsical, at best. That is to say, it ignores rules the way it hallucinates facts. And AI can be manipulated to ignore rules through adversarial input, so AI is not ready to operate dangerous machinery.

I've written an essay

I doubt that

They came at it like this because they are Silicon Valley and for the better part of 3 decades now, they have operated under the assumption that they are always right, and that if they can survive running on a loss at first, eventually they will acheive market dominance and monopolize their product. I just don't see it happening. I see this more like the early dot com days. A lot of investment, and a lot of failures. But eventually, for better or worse, a few of them figured out how to make it work. This time around however, capitalism itself is failing.

So, I expect that labs focusing on more efficient architectures will outcompete those that are trying to brute force the problem. Like sure, DeepSeek isn't small in a sense that you can run it locally, but it is small compared to other models in its class, and much more energy efficient. Whatever hardware we get down the road is going to benefit more efficient models the same way meaning that they will always have a competitive advantage.

From what I see in the latest releases from Anthropic, Fable isn't a huge leap ahead from Opus. There is an improvement, but it's not a definitive jump in capability the way it was from Sonnet to Opus. So, they managed to make a bigger model, but got diminishing returns, and it's evidently so expensive to run right now that they can't even offer it as a default.

The real progress will almost certainly be happening in hybrid architectures where people start coming up with algorithms that complement LLMs and augment their capabilities. These will be like different brain regions responsible for different tasks. For example, memory formation is an obvious example here, another would be to have a built in mathematics engine. A real huge win would be to figure out how to do few shot learning on the fly as well, for which memory is a prerequisite. So, there are plenty of things we already know that can be done much better.

They’d have to raise prices.

And the only reason they wouldn’t want to raise prices is because they wouldn’t be profitable if they did.

Openai cancelled ipo, trumpbucks 2.0, etc. are foreshocks.

They didn’t arbitrarily decide that they’d use a smaller model, they were targeting a constraint. Which of course was really juicy because of all the demand for text inference and the limited new hardware to fulfill it.

The paper you cite is true, I haven’t read it but it based on the title it can’t really be wrong unless they just get wildly over their skis with their claims…

You’re right that everything’s gotta walk through the layers and more layers means a longer walk. That’s limited by memory and interconnect bandwidth though which is still doubling or close to it every year or close to it.

Which means it would have to be twice as fast to use a smaller model on cutting edge hardware to be a real “wall”.

We are not near the end of the memory bandwidth road yet, quantum tunneling isn’t rearing its head again as the workaday wafers that ferry serialized streams from place to place get faster and more numerous. Because they don’t need to get smaller really. The port of New York can grow and sprawl and sprout heretofore unseen support structures like coolers and voltage regulators and coolers for its voltage regulators.

I’ve seen in action what you describe though. On a, sensible chuckle, tiny card like a 3050 slotted into a system with pcie3 very small models run faster than those who can barely fit in the vram with their little bitty context, but that difference shrinks significantly when the card is a 5060 or something that still has limited vram but can load faster due to a faster pcie interconnect.

Now that may seem like apples and oranges because it’s two completely different things but the point of the comparison is to show that the difference between performance measures in time to first, tps or whatever other measurement might be in vogue at the moment when two models are compared shrinks when the interconnect and memory bandwidth gets faster.

To butcher a car metaphor, you can turbocharge an ls and get more power but you’re not escaping engine wear=crankshaft rotations. That might not be as butchered as it first seemed even though the domains are all shifted.

What I expect we'll see going forward is more hierarchical architecture where you have finely tuned models for specific tasks with a general routing model on top. This is basically already where MoE architecture is moving now. We might also see stuff like neurosymbolics get more popular where the LLM acts as a stochastic engine within a symbolic logic system. The model can handle noisy input from the real world, and transform it into structured data that a symbolic engine can operate on.

Brute forcing the problem is a naive approach and US labs took it because they effectively had unlimited resources to train their models until now.

And when more compute becomes available, solutions that are more efficient are going to further benefit from that as well. We see this with DeepSeek right now. They focused on efficiency over capability up front, and now they have a fundamentally cheaper architecture that's rapidly catching up in capability.

Seems like one of several last ditch efforts to keep the bubble going IMO. They're flailing

there’s the structure of the neural network itself. Fundamentally, adding nodes and layers increases the ability of the model to handle more complex input.

Then there’s the actual models we see in use. They are literally as large as the hardware allows. The only reason to use smaller models are to fit some constraint.

So both by the book and in practice bigger is always better.

Now we can’t always go big. I can’t afford to purchase a dgx or even upgrade my wiring to power it, let alone pay the power bill it would rack up or all the other utilities alone when my wife leaves me because of the sound.

My computer can only fit so many expansion cards and pcie is so slow compared to hbm that I’m better off running a small model quickly that fits on one card as opposed to a larger one slowly across several cards.

But those are all constraints. When I replace my motherboard with supermicro gpu host fabric I no longer am limited by the pcie bandwidth and can quickly use models that fit across several cards.

I do agree with you that the future is smaller models, not because of the fundamental nature of the concepts involved but because of the complex constraints that are coming into play.

At least with this announcement it seems like the winds are blowing more in the direction the people predicting a bubble said it would. AAPL became a trillion dollar company in 2018. Now NVDA is worth more than 5 trillion with AAPL on its heels. If it goes it may take some time to come back.