Y
HN Search
Hacker News Search
new
|
comments
|
top
|
jobs
_flux
searching PlanetScale…
1.
▲
2.
▲
3.
▲
4.
▲
5.
▲
6.
▲
30 ms
·
541.
▲
by
_flux
2y ago
So apparently a 20 TB HDD can be purchased at around $250, plus perhaps $30 per hdd slot with a computer (I just recall a number of this range from calculating years ago how much it costs to have storage at home), making it 2 million dollar
542.
▲
by
_flux
2y ago
Actually I wasn't that interested in throughput, only the latency in terms of instructions executed since sending until it is received, though indeed the throughput is also superior with tokio. For most applications this difference doe
543.
▲
by
_flux
2y ago
In addition, they presumably don't want to enter a race to the bottom by undercutting others, as competing by other means (sound quality? selection? brand image?) is less risky.
544.
▲
by
_flux
2y ago
It doesn't seem my perf provides metric for L1 cache evictions (per perf list). Here's the results for 100000 rounds for taskset 1 perf record -F10000 -e branch-misses -e cache-misses -e cache-references target/release/R
545.
▲
by
_flux
2y ago
Hmm so it reloads godot-rust/gdext while the Godot environment is running (recentish new feature), but can it also reload while the game itself is running? I should really give it a shot..
546.
▲
by
_flux
2y ago
> If you ever pull up a debugger and step through an async Rust/tokio codebase, you'll get a good sense for what the overhead here we're talking about is. So I didn't quite do that, but the overhead was interesting to
547.
▲
by
_flux
3y ago
I wonder if you could teach millions of terminal sessions to an LLM and then have mosh run the inference algorithm to hide the latency (and eventually correct when it gets the real data)..
548.
▲
by
_flux
3y ago
And huggingface is hosting (randomly assuming 8-64 GB per model) 5..40 PB of models for free? That's generous of them. Or can the models share data? Ollama seems to have some ability to do that.
549.
▲
by
_flux
3y ago
Unstable rustc can compile stable code. For code you'd write with Dafny I highly doubt it would be an issue. Perhaps you have a proof of the opposite? I've only ever needed nightly when doing something a bit off the beaten path, l
550.
▲
by
_flux
3y ago
Surely generating Rust is not a moving target, if you choose to target stable?
551.
▲
by
_flux
3y ago
I imagine a lot of companies would like to integrate their apps to cars. Think the police, ambulances, taxis/Uber, etc. Currently they just solve the problem by mounting more devices to the car. And even if you want to use only audio a
552.
▲
by
_flux
3y ago
I mean you need to have global high scores, right.. In my one test-run it didn't remember my previous score in the start screen as the "best score". My favorite typing speed game remains typespeed, which can be found from e.g
553.
▲
by
_flux
3y ago
There is a point here: if your implementation doesn't match the specification, then that can be the case. And your example presumes the specification assumed unbounded numbers, while an implementation did not have them. Well, in that c
554.
▲
by
_flux
3y ago
I guess I need to disagree :). Why isn't fuzzing testing? Ultimately it ends up running the code and seeing if it works—for some definition of "working", that is, it doesn't crash—even if it may pick the ways it chooses
555.
▲
by
_flux
3y ago
True, I should have remembered to express it as Non-exhaustive testing cannot.. . I also consider formal methods not testing. But model checking (e.g. with tlc) is something in between, and sort of takes advantage of the "This usually
556.
▲
by
_flux
3y ago
I suppose it would be nice to require at least 100% line coverage for tested encryption-related functionality, when tested by a fuzzer. Deviations from this could be easily detected in CI. Testing cannot be used to prove that a flaw doesn&#
557.
▲
by
_flux
3y ago
ReVanced soon?
558.
▲
by
_flux
3y ago
And how much do you need to invest to get that energy also during the nights, in a green-energy-compliant manner? Arguably it seems it's a good idea to run nuclear power, if it's not more costly than solar, provides 24/7 gree
559.
▲
by
_flux
3y ago
Is it really the case that running an existing nuclear powerplant is less cost-effective than the alternatives, so much so that closing plants is a good idea from green energy point of view?
560.
▲
by
_flux
3y ago
"crashing" could also mean a panic! caused by the runtime ensuring you have at most one mutable reference to an object.
561.
▲
by
_flux
3y ago
Indeed in OOP (or in any kind of indirect call mechanisms, including ones using function pointers in C, right?) you would not find which particular method is being called, but if you are using a nominative type system with OOP, you would at
562.
▲
by
_flux
3y ago
Wouldn't it make sense to release it on Android as well, given it should probably be quite easy with Godot, or is that market just too saturated and attract cloners? Though you'd need to think about how to do the controls on touch
563.
▲
by
_flux
3y ago
Nice in principle, but doesn't it make it quite difficult to analyze such code? E.g. finding places that use a particular function of a protocol. I recall Pyright for Python is not able to find such call sites. But perhaps there are be
564.
▲
by
_flux
3y ago
Indeed it's not the solution to LLM hallucinations—as far as I know nobody knows a solution to it. But it is the solution to needing to run the model again and checking the format of the output to ensure that it conforms to you expecta
565.
▲
by
_flux
3y ago
llama.cpp has the feature to enforce a certain structure from the LLM (but you still need to have the LLM be able to produce that structure and it's beneficial to prompt it towards the exact result). So this particular story doesn'
566.
▲
by
_flux
3y ago
Well, that's the wrong impression. Here's some Arch documentation about it, basically you just create a btrfs on top of multiple devices and it works (to some extent with raid5/raid6 as well): https://wiki.archlinu
567.
▲
by
_flux
3y ago
So this is a good use case for it: you will most likely get immediate feedback if it's wrong, and a bit delayed feedback if it achieved the wrong thing; and you can prod it to try harder. LLMs are best used when you can easily verify t
568.
▲
by
_flux
3y ago
..or merely tested and supported on? Or that it just can't work with anything else?
569.
▲
by
_flux
3y ago
In ext[234]fs the reserved blocks is something else though: they are reserved to a specific user, by default root. So if normal users fill out the filesystem, the root user still has some space to write. Sort of a simple quota system. I bel
570.
▲
by
_flux
3y ago
Extending words in one's public vocabulary to include things one doesn't like is a form of misleading, and thus lying, and thus violence.
More ›