emil-lp

3 ms·

1.
by emil-lp 2mo ago
> An LLM with CoT is Turing-complete. Given arbitrarily large context window.
2.
by emil-lp 2mo ago
STOC is the premier outlet of theoretical computer science results, so let's hope others follow: FOCS, SODA,...
3.
by emil-lp 2mo ago
> The relationship between planar graphs, vertex coloring, and 4 might just be because we put planar graphs and vertex coloring in the same room and 4 popped out, instead of 3 or 5. Well, the current conjecture is that it hasn't to
4.
by emil-lp 2mo ago
But it's not the reviewers who do it, it's the handling editor. Who spends relatively little time per paper (compared to authors and reviewers).
5.
by emil-lp 2mo ago
> it sounds extremely time-consuming. For whom? Surely the authors can find an hour after submitting the paper to a journal? Beware that a reviewer easily spend a full week on reviewing a paper, and there are typically three of them. So
6.
by emil-lp 2mo ago
No, this is pure graph theory, and is quite far away from anything machine learning.
7.
by emil-lp 2mo ago
No, almost none (except for in certain fields, such as HoTT) have formalized proofs.
8.
by emil-lp 2mo ago
Isn't it actually the bread? The meat is given, if I understand correctly.
9.
by emil-lp 2mo ago
Everybody's talking about that. Something like 300Bn tokens, something in the vicinity of USD 10M.
10.
by emil-lp 2mo ago
This is getting ridiculous. Is 5.6 not considered good enough to conduct experiments with anymore? Two months ago, if you weren't using 5.6, you were doing it wrong. Now, you can't conduct experiments using the default model?
11.
by emil-lp 2mo ago
I don't know why that is surprising, but I'm teaching computer science at a university, and students seem surprised every time it comes up.
12.
by emil-lp 2mo ago
But this is not laziness , it's deliberately done signal being cooler than the rest.
13.
by emil-lp 2mo ago
What's surprising (to many) is that n! < exp(n log n)
14.
by emil-lp 2mo ago
It would be interesting to know your (friend's) country of origin.
15.
by emil-lp 2mo ago
Why do they want the jokes back?
16.
by emil-lp 2mo ago
Or while waiting for the bus: if it's late, when should you start walking instead?
17.
by emil-lp 2mo ago
I think you have misunderstood. PhD has nothing to do with expertness. If you have a PhD, you have completed some kind of research training. That's all there is. Says nothing about knowledge or whether or not you're a genius. You
18.
by emil-lp 2mo ago
As a researcher in theoretical computer science, I love PDFs more than any other format when it comes to mathematics. There simply is no contender to LaTeX and PDFs. Lucky for you, almost all research in math, cs, and physics, are put on ar
19.
by emil-lp 2mo ago
> the "rigorous defense" part In my country, that's exactly how it is. Yes, you need to have a thesis to defend, but ultimately it all comes down to the (oral and live) defense/disputation.
20.
by emil-lp 2mo ago
If you replace "pdf" with "abs" the link sends you to the arXiv landing site.
21.
by emil-lp 2mo ago
> If we have a bioweapon close call, would you consider AI an existential risk at that point? If not, how close would we need to get? The thing is, this doesn't really mean anything. What is a bioweapon close call ? What is the
22.
Bruce Schneier: No, AI doesn't mean the end of mathematics – at least not yet(theguardian.com)
9 points by emil-lp 2mo ago|0 comments
23.
by emil-lp 2mo ago
Is this HN crowd in the room now? It's understandable, at this point, to be weary of pelicans.
24.
by emil-lp 2mo ago
You cannot guarantee that from a hash map since an adversary who knows the hash function (unless it's cryptographic) could game the data structure to their advantage.
25.
by emil-lp 2mo ago
Θ does not usually mean average, but simultaneously upper and lower asymptotic bounds.
26.
by emil-lp 2mo ago
> To put it differently, saying that a hash table is O(1) or constant time is a model Nobody really says that, nor is it a model. It is the expected time complexity.
27.
by emil-lp 2mo ago
Expected
28.
by emil-lp 2mo ago
I read your arxiv paper yesterday (or was it the day before). Do you think this can be used to speed up the algebraic method for k-path? If so, you should enter next years PACE challenge.
29.
by emil-lp 2mo ago
Yes, I too have noticed an increased high quality customer service.
30.
by emil-lp 2mo ago
Yes, there is no doubt about scientific misconduct. I think the entire math community agree on that.

More ›