Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
the8472
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
50 ms
·
901.
▲
by
the8472
4y ago
Everyone who doesn't tightly control the systems they deploy to. io_uring is a new and developing interface so to support extant systems running libraries need to have fallback codepaths. Or more likely: they already have existing code
902.
▲
by
the8472
4y ago
Maybe bubbling the air through water or a mist/air counterflow like scrubbers in coal plants? Throw in some UVC light source to keep germs from growing in the water. Are there UVC LEDs yet? There also are electrostatic filters but they
903.
▲
by
the8472
4y ago
> It can mean reading meeelions of tiny files and seeking back and forth randomly. Tiny files are fine if you fetch them from NVMe and keep the queues fed, i.e. you need to issue multiple concurrent reads, not do it sequentially. Not bei
904.
▲
by
the8472
4y ago
seccomp filters generally are whitelists. If they haven't whitelisted io_uring_setup it's a non-issue. They only suffer the performance consequences of userspace having to use the synchronous syscalls. also, iouring doesn't a
905.
▲
by
the8472
4y ago
Threads are much cheaper to spawn compared to processes because threads share memory while forked processes are copy-on-write which requires updating the page tables for the parent and copying the kernel bookkeeping entries over to the chil
906.
▲
by
the8472
4y ago
For example rust doctests consist of compiling each little doc snippet and executing it ensure that the asserts are true. That's two processes (rustc + the compiled piece of code) per test. This is an order of magnitude slower than mul
907.
▲
by
the8472
4y ago
Even a tidally locked body can still liberate (as the moon does) so the terminator would at least move around over some fraction of the surface. Maybe together with a dense atmosphere that could be enough to redistribute the heat and keep t
908.
▲
by
the8472
4y ago
Some additional linux-specifics: https://lwn.net/Articles/855049/ https://lwn.net/Articles/754980/
909.
▲
by
the8472
4y ago
My understanding is they load in weights occasionally into sram and then pump in training data on the sides of the die and have multiple cores operate on a wavefront of data. So the cores don't compete for host memory bandwidth because
910.
▲
by
the8472
4y ago
Would anonymous sources be more credible? Because they could always be anonymized.
911.
▲
by
the8472
4y ago
The term AGI means "artificial general intelligence", it does not mean human-like intelligence. So it being incomprehensible to us does not preclude it from being an AGI.
912.
▲
by
the8472
4y ago
kTLS has been added to linux too including offload. It also has p2p-dma, so in principle you can shovel the file directly from NVMe to the NIC and have the NIC encrypt it, so it'll never touch the CPU or main memory. But that only work
913.
▲
by
the8472
4y ago
They'd be solving other low level problems instead I assume. Systems that are unintentional blackboxes instead of intentional ones.
914.
▲
by
the8472
4y ago
63 for me, and most of those are libraries which should already be in memory. Which is nanoseconds away from the CPU while servers are dozens to hundreds milliseconds away.
915.
▲
by
the8472
4y ago
I think the argument is that in a better world where everything came with open specs we wouldn't need reverse engineers and those smart people could spend their time on something else. It's an argument pointing out where we live a
916.
▲
by
the8472
4y ago
And replacing the water in your body with an aprotic solvent such as DMSO.
917.
▲
by
the8472
4y ago
From central european here. The initial interview usually has been done by a HR person + someone from the department that is hiring. If that went well there was an on-site tour of the department in question, talking with engineers, looking
918.
▲
by
the8472
4y ago
W7X is a bit like the LHC. It already was operational in 2015 and got incremental updates since then.
919.
▲
by
the8472
4y ago
Is commercial ICF realistic though? Each shot needs a carefully prepared fuel pellet. To get commercial power they'd have to fire a shot per second or so. That seems like a really expensive manufacturing operation to keep it going.
920.
▲
by
the8472
4y ago
There's a middle ground. You can permit novel substances but if it is discovered to be harmful then the burden of proof is inverted for the family of related chemicals. Like not only narcotics but also chemical analogues (even those th
921.
▲
by
the8472
4y ago
Not quite, the mandatory insurance is capped at 2.5G€. The taxpayer would have to cover the rest. Compared with the fukushima cleanup costs that's quite insufficient.
922.
▲
by
the8472
4y ago
She isn't responsible for energy, infrastructure or finance, which should be responsible departments regarding the power plants. > On February 8, 2022, it was announced that she was to be appointed to the Federal Republic of Germany
923.
▲
by
the8472
4y ago
Could be the classic nagle / delayed ack interaction. Or shoddy wifi.
924.
▲
by
the8472
4y ago
io_uring already has IOSQE_IO_DRAIN which sounds like it's currently implemented as stalling the IO pipeline, but maybe it could be translated to hardware barriers instead in some circumstances (e.g. when the surrounding IO is all O_DI
925.
▲
by
the8472
4y ago
That's neat and quite a unixy solution but still fairly limiting. Are there any plans for a hybrid solution where most of podman runs as non-root via userns but invokes a suid helper to setup the network?
926.
▲
by
the8472
4y ago
> - no global gc pauses, low-latency ZGC has sub-millisecond pauses https://malloc.se/blog/zgc-jdk16 And afaik azul's C4 collector has no global pauses, only per-thread pauses (which are also short)
927.
▲
by
the8472
4y ago
Unless you have solid-state circuit breakers. https://youtu.be/szhDolRV2S0?t=2268
928.
▲
by
the8472
4y ago
or became more concerned with AI alignment than thinking about potential upsides
929.
▲
by
the8472
4y ago
re the chinese room, you might want to consider the computational and memory complexity of a lookup table. https://www.scottaaronson.com/papers/philos.pdf page 14
930.
▲
by
the8472
4y ago
This is an old idea known as hashcash. https://en.wikipedia.org/wiki/Hashcash Newer variations (such as argon2) are tunable so you can include memory footprint and cpu-parallelism. There also are time-lock puzzles or v
More ›