krackers

51 ms·

961.
by krackers 2y ago
But wouldn't the same apply to condition variables, not just semaphores? Also I found another reference to that internal post here https://cohost.org/Catfish-Man/post/913314-a-love-letter-to-...
962.
by krackers 2y ago
There's this interesting comment by "John Day" on that page, does anyone have more context/detail?
963.
by krackers 2y ago
What exactly is the logic osx is using here that causes non-aggregate input to always be switched when plugging in external source, but aggregate sources remain sticky?
964.
by krackers 2y ago
It seems every company has decided to crank the ads dial to 11 as they have no other ideas for innovation left.
965.
by krackers 2y ago
I always see SQLite as recommended, but every time I look into it there are some non-obvious subtleties around txn lock, retry behavior, and WAL mode. By default if you don't tweak things right getting frequent SQLITE_BUSY errors seems
966.
by krackers 2y ago
Funny that most companies have no problems with outsourcing roles, but for some reason many expect on-callers to be available outside business hours. They could easily hire someone on the opposite time-zone.
967.
by krackers 2y ago
It'd be absolutely foolish if the agreement wasn't contingent on funding. I assume the reason it's not explicitly stated was some sort of NDA (since IA is also involved in turmoil and Google doesn't want to be part of th
968.
by krackers 2y ago
I think it was Steve Jobs who insisted that that even the inside of their products look as clean as the outside. It unfortunately means it's a pain to disassemble, but they certainly do look pretty.
969.
by krackers 2y ago
> tinygrad has a hardware abstraction layer, a scheduler, and memory management. It's an operating system Doesn't every ML framework have that?
970.
by krackers 2y ago
Seems this was fixed in https://github.com/VadimBoev/FlappyBird/commit/f57a3ec90078b... by forcing the gameloop to be 60fps. It's a simple solution, but does mean that things are effectively capped at 60
971.
by krackers 2y ago
Isn't iTunes already dead and replaced with music?
972.
by krackers 2y ago
If you're writing out protobufs, isn't the recommended approach to use some record-oriented wrapper? Something like TFRecord. Or is parseDelimitedFrom how that's implemented in the first place? But Google must surely have som
973.
by krackers 2y ago
What's the reason why reading 3GB via duckdb is _faster_ than raw HDD speed? Is it compression/caching?
974.
by krackers 2y ago
I sure hope that the continuing sharply elevated rates of excess mortality [1,2] will be objectively and responsibly investigated so we can find the root cause. [1] https://www.bmj.com/company/newsroom/high-excess-
975.
by krackers 2y ago
Compromised account apparently https://news.ycombinator.com/item?id=41613875 This is a lot of effort, what was the end-goal?
976.
by krackers 2y ago
I'm actually surprised they suggest getaddrinfo() directly over the async CF wrapper.
977.
by krackers 2y ago
>so presumably there's some other low-level non-blocking call Correct, CFNetwork is open source so you can check implementation but last I remember it used some variant like `getaddrinfo_async`. But Apple really doesn't want yo
978.
by krackers 2y ago
Ah my bad, great catch! I've updated my comment accordingly.
979.
by krackers 2y ago
I think you're misrepresenting the study. It builds upon previous work that examines the computation power of the transformer architecture from a circuit-complexity perspective. Previous work showed that the class of problems that a &q
980.
Chain of Thought empowers transformers to solve inherently serial problems(arxiv.org)
261 points by krackers 2y ago|184 comments
981.
by krackers 2y ago
There were rumors that 3.5 Sonnet heavily used synthetic data for training, in the same way that OpenAI plans to use o1 to train Orion. Maybe this confirm it?
982.
by krackers 2y ago
If ARC-AGI were a good benchmark for "AGI", then MindsAI should effectively be blowing away current frontier models by order of magnitude. I don't know what MindsAI is, but the post implies they're basically fine-tuning
983.
by krackers 2y ago
>However, it aligns with Apple’s recent policy changes that allow repairs with genuine used parts while maintaining stringent security measures. Doesn't this imply that so long as the phone is unliked from apple id, the parts can be
984.
Scaling LLM Test-Time can be More Effective than Scaling Parameters(arxiv.org)
1 points by krackers 2y ago|0 comments
985.
by krackers 2y ago
Also https://news.ycombinator.com/item?id=41463330
986.
by krackers 2y ago
Given that, why are specs of dust on glasses so noticeable?
987.
by krackers 2y ago
Is this similar to the issue with 2011 MacBook Pro GPU failures? Those also supposedly could be revived by baking in the same way?
988.
by krackers 2y ago
Fyi that 120 nits recommendation is probably because most color grading assumes you work in a dark-ish room. It's probably not a good value for office work.
989.
by krackers 2y ago
>super bright white light Simply going outside should be brighter than your monitor. We're pretty much evolved to hunt (and thus see best) during the daytime. I'm convinced a lot of the reasons people hate light themes is becau
990.
by krackers 2y ago
Isn't this tautologically saying the compiler does sane things if you define "sane" as what the compiler allows? As Linus said standards are written on toilet paper, in the real world you have a need to do signed overflow and

More ›