programjames

18 ms·

361.
by programjames 2y ago
So, capitalism is built on the idea of mutually beneficial trades. It's supposed to be a cooperative game, and in cooperative games you should always provide as much information to your counterpart as possible. Capitalism breaks down w
362.
by programjames 2y ago
AoPS was a good resource when I was a kid: https://artofproblemsolving.com/community
363.
by programjames 2y ago
Yes, essentially proofs. I didn't use that word because people usually only write proofs for each other, so they end up being much more formal. When it's just yourself, chicken scratch in the margins is good enough. How you learn
364.
by programjames 2y ago
Probably figuring out math paths. I've always been bad at remembering formulas or theorems, but that's because I remember how to get to that formula or prove the theorem instead. E.g. in grade school, I never memorized the sum of
365.
by programjames 2y ago
As far as I can tell, flow-based models are bound by the exact same requirements as energy based models (flow = diffusion/normalizing flow/flow-matching models). But they're absolutely right about VAEs. Those are a memetic vi
366.
by programjames 2y ago
They're good for reinforcement learning. E.g. Cicero uses piKL which samples according to p ∝ anchor_policy * exp(utility / temperature) The utility is exactly the same as "energy". The article ignores entropy, but you c
367.
by programjames 2y ago
Where would you use 3D Gaussian splatting? Static environments for video games?
368.
by programjames 2y ago
Yes, pop in a normalizing flow that removes tone and recreates it with a small audio sample as context. https://arxiv.org/abs/2312.01479
369.
by programjames 2y ago
I think the problem is worse than just algorithms. If something is around long enough, it'll naturally evolve an implicit coordination, as those who don't coordinate die out.
370.
by programjames 2y ago
This is only irrational for the landlords, not the software suggesting prices to them. If you churn 10% of your units, but prices go up by 20%, then you're making more money as a rental management company.
371.
by programjames 2y ago
> permanent, immanant I'm out of my depth here, but I'm imagining something like a path integral[0] between n particles. For example, with two fermions, A and B, you'd get a fraction f of A moving to B's location, and
372.
by programjames 2y ago
Something I've heard physicists say but doesn't make sense to me is, "gravity is a just a distortion in space-time, it's not like the other forces." Can someone with more knowledge than me explain this? As far as I
373.
by programjames 2y ago
Ah, that would do it. I've just been reading some of his posts and watched his podcast with Chris Williamson[1], and he seems like a rather smart, earnest guy trying to do his best to help out the country. He did mention the whole bure
374.
by programjames 2y ago
I'm not British. Is Cummins not terribly liked in the UK?
375.
by programjames 2y ago
I found the last half of the essay more interesting so I linked and titled that. Sorry for the confusion!
376.
Judging Books by Their Covers [pdf](bradthiessen.com)
2 points by programjames 2y ago|0 comments
377.
How to help parents give their children better education? (2023)(dominiccummings.substack.com)
3 points by programjames 2y ago|7 comments
378.
by programjames 2y ago
> I think we should do our best to ensure that we the dominating factor that determines outcome is due to what someone can actually control. I think this is where I'm coming from as well. When I got to university, I met tons of peop
379.
by programjames 2y ago
As I said, > I know many people who did take... In fact, the vast majority of my friends did, so my mental model is more useful to me than one that apportions a larger cut to the rest of the population. I also find it egregious that thir
380.
by programjames 2y ago
I know many people who did take multivariate calculus, group/ring theory, and thermodynamics in high school, and think this should be the norm. I believe I consider "high school" math what most people consider "undergrad
381.
by programjames 2y ago
These remind me of interaction combinators [1], which are being used in the Bend programming language [1]. I think it'd be good for the standard to also be a valid interaction net. [1]: https://core.ac.uk/download/
382.
by programjames 2y ago
It's not that these AI companies need to raise $100 billion, it's that they can. They could go slower, working on custom hardware or better architectures to get 1000x cheaper training, but their models would only be as good as t
383.
by programjames 2y ago
> they prefer take home assignments Proof or it didn't happen. I have never met--or even heard of until now--a developer that preferred this.
384.
by programjames 2y ago
The first thing that pops up is > We use cookies to ensure that we give you the best experience on our website. > Accept | Decline Do "non-essential" cookies actually make my experience any better? I always decline and haven
385.
by programjames 2y ago
> but not that many bald heroes - you've got The Rock, Jason Statham, who else? Can't forget Aang!
386.
by programjames 2y ago
That doesn't implement a trainable network on hardware, it's just creating a "reservoir" of associations between the inputs.
387.
by programjames 2y ago
> These methods are typically slow because the number of gradient updates scales linearly with the number of learnable parameters in the network, posing a significant challenge for scaling up. This is a pretty big problem, though if you
388.
by programjames 2y ago
You can regularize the networks to make them transfer easier. I can't remember the abstract's title off the top of my head though.
389.
by programjames 2y ago
Eh, imaginary numbers were invented to be useful. Cardano needed them as intermediary steps when plugging in his cubic formula.
390.
by programjames 2y ago
No, we measure angles in radians so that d^4/dx^4 (sin x) = sin x.

More ›