leumassuehtam

6 ms·

1.
by leumassuehtam 24d ago
My implementation: https://github.com/keyehzy/thinking_with_looped_flows
2.
by leumassuehtam 9mo ago
It looks like the B programming language.
3.
by leumassuehtam 9mo ago
> All memory must be statically allocated at startup. No memory may be dynamically allocated (or freed and reallocated) after initialization. This avoids unpredictable behavior that can significantly affect performance, and avoids use-af
4.
by leumassuehtam 11mo ago
> The artificial intelligence chip leader's chief in October said that the U.S. can win the AI battle if the world, including China's massive developer base, runs on Nvidia systems. He, however, lamented that the Chinese govern
5.
by leumassuehtam 1y ago
People believe that more parameters would lead to overfit instead generalization. The various regularization methods we use today to avoid overfit hadn't been discovered yet. Your statement is mostly likely about this.
6.
by leumassuehtam 1y ago
text-davinci-001 still feels the more human model
7.
by leumassuehtam 1y ago
I hope this time they can come up with a model at least or par with open source offerings.
8.
by leumassuehtam 1y ago
Thanks for the answer!
9.
by leumassuehtam 1y ago
Genuine question: is this a wrapper around Google's gopacket?
10.
by leumassuehtam 1y ago
The author has a nice series on compiling a Lisp [0], but unfortunately his search engine fails to find it by querying it with "lisp" or "Lisp". [0] https://bernsteinbear.com/blog/compiling-a-lisp-0&
11.
by leumassuehtam 1y ago
Since you've mentioned, which hash functions are better while being faster than XXH3?
12.
by leumassuehtam 1y ago
One way of doing that is by each time step delta perform two integrations, one which give you x_1, and another x_2, with different accuracies. The error is estimated by the difference |x_1 - x_2| and you make the error match your tolerance
13.
by leumassuehtam 1y ago
You're right that he is just rederiving electromagnetism through local U(1) gauge symmetry. He define his metric as g_{\mu\nu}=A_\mu A\nu, which is a gauge dependent metric that gives you Maxwell's equation in the covariant formul
14.
by leumassuehtam 2y ago
I like to think of an algebra the same as a language. You want to define and constrain your language such that you can describe elements of your universe is clear and consice. For example, if I am trying to describe a art piece in the museu
15.
by leumassuehtam 2y ago
I'm convinced that "completion" models are much more useful (and smart) than "chat" models, being able to provide more nuanced and original outputs. When gpt4 come out, text-davinci-003 would still provide better co
16.
by leumassuehtam 2y ago
What do you mean peer reviewed? This is the standard textbook way of numerically studying Ising models.
17.
by leumassuehtam 2y ago
That's a fallacy. Do you really believe that paying taxes (essentially what you mean about buying things for everyone else) would stop one to want to become a billionaire? It's about incentives, not taxation.
18.
by leumassuehtam 2y ago
It's easy to notice, even on this website, when a comment starts with "It's worth noting that...". It sounds very LLM-esque.
19.
by leumassuehtam 3y ago
The latest incarnation of Fourier series is in AI, in the so called RoPe technique.
20.
by leumassuehtam 3y ago
Yes, it was tested in 8:08 of https://youtu.be/D5u7trVY5Ho?si=n1N_sm3iy0dTx4b5
21.
by leumassuehtam 3y ago
The barrier of publishing a paper in arxiv is much lower than in academic journals. That said, you still have to convince them that your idea is relevant. They may require some sort of credentials, but if the idea is solid and the paper is
22.
by leumassuehtam 3y ago
You will not find anything as good as the paid version of ChatGPT. You might try Phind [0]. [0] https://www.phind.com/
23.
by leumassuehtam 3y ago
Did they use this to train Gemini? Which raises the question, where is Gemini?
24.
by leumassuehtam 3y ago
I had great results with Phind. Their newest fine tune model (V7) has been a pleasant experience and better than most open source models out there. Nit: your link has a trailing "s" which makes it 404 :)
25.
by leumassuehtam 3y ago
Quite the opposite, the 4 straight lines represent all that math. Notation is very powerful and get you quite far.
26.
RedPajama-Data-v2: 30T tokens filtered and de-duplicated(twitter.com)
4 points by leumassuehtam 3y ago|0 comments
27.
by leumassuehtam 3y ago
Do you mind if I reach you out to ask some questions, I'm moving to Shanghai soon.
28.
by leumassuehtam 3y ago
Shameless plug of a similar idea to embed lisp into LaTeX files. https://github.com/keyehzy/lispylatex
29.
by leumassuehtam 3y ago
Last commit to the repo seems to be from 7 years ago. Previous post from 2016: https://news.ycombinator.com/item?id=13230382
30.
An observation on Generalization – Ilya Sutskever [video](youtube.com)
4 points by leumassuehtam 3y ago|1 comments

More ›