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

Billion-Parameter Theories

worldgov.org

Post reply on HN
Print view View original

61–70 of 92 posts

Re: Billion-Parameter Theories

#61

Post by roughly »

seanlinehan wrote: post #56
roughly wrote: post #51

There's a lot of ink in this spent on how Poverty, Climate Change, Urban Decay, and Financial Markets are Complex Hard Complicated problems. The problem with these is they're also problems where there are actors profiting from the failure to fix the system - the issue isn't that we don't understand the complex nature of the domain, it's that the components of the system actively and agentically resist changes to the…

Reflexivity is nodded to in the definition of complex systems in the piece! I think what you're saying is poverty is actually simple, and the solution is to stop the bad actors causing poverty? But at the same time, you are correctly recognizing that attempts to stop bad actors from causing poverty triggers reflexive responses and cascading repercussions. Which sounds mighty like a complex system?

No, I'm not saying the problem is simple, but I'm saying that in many of these cases a systematic understanding of the problem isn't what we're lacking in pursuit of fixes - the reason the problem seems so intractable is because parts of the system benefit from perpetuating the problem and take agency to ensure the problem does not get fixed.

Poverty is one of these, but I think Climate Change is the most direct - the climate is complex, but climate change is simple: we're releasing too much carbon into the atmosphere, we have been for a century, and we've known that for at least half a century*. The issue isn't that we don't have the capacity to model or understand the problem, the issue is that powerful actors have used the leverage available to them within the system to prevent us from making changes to fix the problem.

And, you're right, that makes the problem difficult, because the system includes those actors resisting changes to the system, but again, it's not difficult because we don't understand it, it's difficult because we're being actively resisted by people who do not want to solve the problem, and that should be acknowledged by people looking to make it an abstract mathematical modeling problem.

* This isn't a conspiracy theory: https://en.wikipedia.org/wiki/ExxonMobil_climate_change_deni...

Re: Billion-Parameter Theories

#62

Post by phyzix5761 »

wavemode wrote: post #58

Not to sound condescending, but this reads like someone fimiliar with LLMs but very unfamiliar with statistics in general. If we could understand economics, or poverty, or any number of other social structures, simply by cramming data into a statistical model with billions of parameters, we would've done that decades ago and these problems would already be understood. In the real world, though, there is a phenomenon…

I could be wrong but I think we crossed the 1 billion parameter threshold in 2019. I'm not sure we had this ability for decades.

Re: Billion-Parameter Theories

#63

Post by clickety_clack »

wavemode wrote: post #58

Not to sound condescending, but this reads like someone fimiliar with LLMs but very unfamiliar with statistics in general. If we could understand economics, or poverty, or any number of other social structures, simply by cramming data into a statistical model with billions of parameters, we would've done that decades ago and these problems would already be understood. In the real world, though, there is a phenomenon…

Really good data only goes back a couple or more decades, so any data you put in your model has only been influenced by the kinds of things we’ve seen in that time. Impact of a hot war between major powers? The gold standard? Stagflation? Invention of the car or train? Transition of major world powers to democracy or communism? All these events left almost no data compared to today, to say nothing of run-of-the-mill changes in styles of monetary policy, economic drivers or shifts in style of government.

Re: Billion-Parameter Theories

#64

Post by snarkconjecture »

wavemode wrote: post #58

Not to sound condescending, but this reads like someone fimiliar with LLMs but very unfamiliar with statistics in general. If we could understand economics, or poverty, or any number of other social structures, simply by cramming data into a statistical model with billions of parameters, we would've done that decades ago and these problems would already be understood. In the real world, though, there is a phenomenon…

Deep neural networks can generalize well even when they're far into the overparametrized regime where classical statistical learning theory predicts overfitting. This is usually called "double descent" and there are many papers on it.

Re: Billion-Parameter Theories

#65

Post by curao_d_espanto »

wavemode wrote: post #58

Not to sound condescending, but this reads like someone fimiliar with LLMs but very unfamiliar with statistics in general. If we could understand economics, or poverty, or any number of other social structures, simply by cramming data into a statistical model with billions of parameters, we would've done that decades ago and these problems would already be understood. In the real world, though, there is a phenomenon…

> The emerging field of mechanistic interpretability suggests otherwise. Researchers are developing tools to understand how neural networks do what they do, from network ablation and selective activation to feature visualization and circuit tracing. These techniques let you study a trained model the way a biologist studies an organism, through careful experimentation and observation.

honestly, when I read that part of the article I imagined that author never studied how computers were made and where the engineering ideas came from, all technology just "popped" and here we are talking about complexity and stuff like the LLM is truly alive

Re: Billion-Parameter Theories

#67

Post by jjk166 »

> And the epistemology shifts in ways that might be uncomfortable. Instead of "I understand the causal mechanism and can predict what happens if I change X," you get something more like "I have a sufficiently rich model that I can simulate what happens if I change X, with probabilistic confidence." The answers are distributions, not deterministic outputs. That's a different kind of knowing.

Being able to simulate something is not a kind of knowing. It is, in fact, the opposite of knowing. If you know how a system behaves, there is no need to simulate it. In particular, if the model you need to simulate it is way more complicated then the phenomenon itself, you really really don't understand it.

I'm reminded of Feynman's observation that to simulate a quantum system, like an atom, with classical methods requires a tremendous number of atoms, and his intuition that there should be a much smaller way to perform such calculations. This is the conceptual underpinning of quantum computation.

A billion parameter neural network may work as a functional tool, but the fact is these supposedly complex problems simply don't have billions of relevant free parameters. You're not going to understand a hurricane by feeding terabytes of data to find the butterfly that flapped its wings in just the wrong way at just the wrong time. Sure extremely small differences in starting conditions can have lead to radically different outcomes, and a butterfly flapping its wings could have influenced a hurricane in some way. But if you understand how hurricanes work, you know that butterfly's influence is just noise - the hurricane started and progresses as it does because of temperature gradients on the ocean. If you found and stopped the butterfly from flapping its wings, the conditions for the hurricane would still exist and something else would set it in motion.

Billion parameter theories work in practice because if you throw everything at the wall, the small amount of stuff that can stick will. Likewise if you throw enough data at a problem, whatever data is actually relevant will be analyzed. This can be useful as a stepping stone to understanding, interrogating the model to reveal which parameters have more relevance and the wights of their interactions. But the idea that because you have a tool that addressed a symptom of your ignorance means you are no longer ignorant is folly.

Re: Billion-Parameter Theories

#68

Post by roughly »

munchler wrote: post #57
newyankee wrote: post #53

Earlier quoted context omitted.

Literally countries with so much surplus land: Canada, Australia etc. have housing crisis where most of the top 10-20% of the population has become speculators in housing and openly NIMBY with no interest in supply side solutions unless forced down.

If there’s surplus land, why build something unwanted in someone’s backyard? I’m a suburban NIMBY homeowner and I feel like you’re actually making my argument without realizing it. I’m all for building new houses on unused land. Can you please just do it without ruining my neighborhood? Build nice new neighborhoods and make them as dense as you’d like, but don’t try to force density on older, established neighborhood…

I'm actually curious - have you spent time in cities like Bern or Bilbao? I think urbanism's been a hard sell in the US because we don't really have a lot of great examples of it - New York's maybe the closest we've got to a European style city, but that's only in certain places and it's still a bit much. I was in Europe last year and I was surprised how calm some of the cities were - green, walkable, a lot of nice cafes and parks, good public transit, and it never really felt overwhelming the way that, say, Chicago or LA does. I grew up in the suburbs, and I felt like some of the smaller European cities delivered the suburban sales pitch better than a lot of places I've been in the US.

(Don't take this as an attack or critique - genuine curiosity.)

Re: Billion-Parameter Theories

#69

Post by BobbyTables2 »

jjk166 wrote: post #67

> And the epistemology shifts in ways that might be uncomfortable. Instead of "I understand the causal mechanism and can predict what happens if I change X," you get something more like "I have a sufficiently rich model that I can simulate what happens if I change X, with probabilistic confidence." The answers are distributions, not deterministic outputs. That's a different kind of knowing. Being able to simulate som…

I think “Hitchhikers’ Guide to the Galaxy” passage talking about the train crashes from a broken clock was extremely prescient.

I feel like enormous models will end up this way…

Re: Billion-Parameter Theories

#70

Post by adamzwasserman »

roughly wrote: post #51

There's a lot of ink in this spent on how Poverty, Climate Change, Urban Decay, and Financial Markets are Complex Hard Complicated problems. The problem with these is they're also problems where there are actors profiting from the failure to fix the system - the issue isn't that we don't understand the complex nature of the domain, it's that the components of the system actively and agentically resist changes to the…

This is the strongest point in the thread. The article treats poverty, climate, and markets as though the obstacle is insufficient model capacity. But these systems contain agents with values and motivations who actively resist interventions. A billion-parameter model of a system whose components are trying to game the model will never be a theory of that system. The agents will simply route around it.

More broadly, the article assumes that scaling model capacity will eventually bridge the gap between prediction and understanding. I have pre-registered experiments on OSF.io that falsify the strong scaling hypothesis for LLMs: past a certain point, additional parameters buy you better interpolation within the training distribution without improving generalization to novel structure. This shouldn't surprise anyone. If the entire body of science has taught us anything at all, it is that regularity is only ever achieved at the price of generality. A model that fits everything predicts nothing.

The author gestures at mechanistic interpretability as the path from oracle to science. But interpretability research keeps finding that what these models learn are statistical regularities in training data, not causal structure. Exactly what you'd expect from a compression algorithm. The conflation of compression with explanation is doing a lot of quiet work in this essay.

Post reply on HN