bugfix-66

10 ms·

31.
Show HN: Makeshift GPU tensor core using 64-bit CPU integer math(bugfix-66.com)
2 points by bugfix-66 4y ago|1 comments
32.
by bugfix-66 4y ago
There are also interesting computational possibilities, like the broadword matrix multiplication algorithm that Knuth presents in The Art of Computer Programming Volume 4A: https://bugfix-66.com/2d447332bf8f8c67348d520c6508b
33.
by bugfix-66 4y ago
Explained by Don Knuth here, if you can't solve it yourself: https://youtu.be/o22BAuQj3ds?t=37m40s
34.
Show HN: Averaging the Corresponding Fields of Two Integers(bugfix-66.com)
1 points by bugfix-66 4y ago|1 comments
35.
by bugfix-66 4y ago
So, specifically what's wrong? It looks no-nonsense and to-the-point. Were you unable to find the information you wanted? Was the site confusing? Or it was just aesthetically unpleasant?
36.
by bugfix-66 4y ago
I aspire to have my site on this list one day. But according to their rules, it would need to get 100 upvotes on Hacker News, and that's not realistic. 5. ... some content from the website must have received at least 100 points on
37.
by bugfix-66 4y ago
The user experience is excellent. Can you point to a specific problem?
38.
by bugfix-66 4y ago
This is an attempt to improve on djb's fantastic libsecded: https://pqsrc.cr.yp.to/libsecded-20220828/README.html We're using a novel Hamming code that doesn't require any interleaving, so the original d
39.
Show HN: Improving on Daniel Bernstein's Libsecded(bugfix-66.com)
1 points by bugfix-66 4y ago|1 comments
40.
by bugfix-66 4y ago
HERE IS THE ANSWER, ACCEPT NO SUBSTITUTES: Take 20 milligrams of propranolol, 1.5 hours (90 minutes) before your first interview. It works for presentations and speeches, too. Propranolol is a beta blocker and you can easily get a prescript
41.
by bugfix-66 4y ago
We are witnessing a monstrous perversion of "fair use" and the greatest theft of intellectual property in human history.
42.
by bugfix-66 4y ago
It's interesting to consider how you might prevent training using a license without being too restrictive. Here is an example of a license that attempts to directly prohibit training. The problem is that you can imagine such software c
43.
by bugfix-66 4y ago
There are people who believe this is real AI, not just aggregation and interpolation. They really believe the software understands code generally.
44.
by bugfix-66 4y ago
From your list, it has solved simple matrix multiplication, LSD radix sort, and pointer padding, all of which appear many, many times in its training set. I'm surprised it can fix the two prediction compressor bugs, even with a hint...
45.
by bugfix-66 4y ago
No, it's just a well-written a piece of software. Thanks :)
46.
by bugfix-66 4y ago
The code you submit is compiled, linked, and executed against a suite of tests. Any functionally correct solution is accepted. If the site rejected your code, it's because your code was wrong (sorry).
47.
by bugfix-66 4y ago
Lots of people here have access to Microsoft Copilot or GPT3. People with access to these models can demonstrate how the system performs on code THAT WASN'T IN THE TRAINING SET by solving a few of these puzzles. The reality is that all
48.
by bugfix-66 4y ago
Please, if you have an "artificial intelligence" that can write and understand code, I'm sure it can fix some tiny bugs in a little code that wasn't in the training set. Why is this so contentious? Surely the "AI&qu
49.
by bugfix-66 4y ago
I think we should question this, and give it a real test. Let's see how Copilot does on code that wasn't in the training set. No problem, right?
50.
by bugfix-66 4y ago
It's possible to prove employment at a company without revealing your identity. But he didn't follow through. That's ok. People are very passionate about "AI" and disillusionment can be painful. Many "AI bros&q
51.
by bugfix-66 4y ago
No, I don't think I will.
52.
by bugfix-66 4y ago
Contact me from your work (non-anonymous) email address, friendo.
53.
by bugfix-66 4y ago
I work at the most important company in the "AI" industry, a company you hear about every day. I write GPU kernels for transformers and convolutions. You probably use my BLAS kernels in your networks. I know EXACTLY what I'm
54.
by bugfix-66 4y ago
My thoughts on the topic of large-scale "AI" plagiarism: https://bugfix-66.com/7a82559a13b39c7fa404320c14f47ce0c304fa... Discussed numerous times here on Hacker News. We are witnessing a monstrous perversion of &q
55.
by bugfix-66 4y ago
Yes, you clearly don't understand it. That's ok, but you probably shouldn't confuse the issue. Basically the attack shows that Gimli's speed and simplicity introduces exploitable flaws and reduces its security: Bernstei
56.
by bugfix-66 4y ago
Published by Tesla? It's unwise to use cryptographic code written by engineers working 72 hour weeks at an "extremely hardcore" company. Here is the Gimli spec: https://gimli.cr.yp.to/spec.html Here is the at
57.
by bugfix-66 4y ago
It's easy to find studies connecting metformin to weight loss and loss of appetite. I'll leave that to you, just use Google and you'll be swamped with evidence. You say you've never heard of metformin affecting food pala
58.
by bugfix-66 4y ago
Here's a question to consider: How much worse would the situation be if they weren't on metformin?
59.
by bugfix-66 4y ago
Yes, I see three responses along these lines, already. Some people don't lose weight. Some people's appetite isn't affected. Some people retain their taste for sweets. And so on. But "misleading and a bit dangerous"
60.
by bugfix-66 4y ago
We're talking about a particular drug here (metformin) not just "anything". I'm saying it's likely your doctor will be willing to prescribe METFORMIN (that drug, in particular) for off-label use if you ask.

More ›