it is worth pointing out that tink has binding for a bunch languages (C++, objective-c, rust, python, go and java) and has support for a bunch key management systems (GCP, AWS and Hashicorp)
The biggest difference for you is probably that WebGPU has compute shaders and WebGL does not, they are pretty nice if you want to implement more complex algorithms.
Both versions are quantized and should use the same amount of RAM, the difference with QAT is the quantization happens during training time and it should result in slightly better (closer to the bf16 weights) output
The WiFi Alliance has a habit of always have a marketing name and a different name in the spec, you'll a lot more references to it in places like WPA supplicant if you search for Neighbor Awareness Networking (NAN). Also here is the link to the spec https://www.wi-fi.org/system/files/Wi-Fi%20Aware%20Specifica...
It is almost never I good idea to assert in a library, unless the error is truly unrecoverable. I think returning an error code\throwing an exception would be very reasonable and a much better API than failing silently though.
If you're going to do this you should also check for k=rsa since other key types will be different lengths. But I'd really recommend something like in the other comment where you base64 decode and parse it with something like openssl.
rr is only works on Linux and the release of Windows TTD was after this blog post was published. Also the huge slowdown from time travel debuggers can sometimes make tricky bugs like this much harder to reproduce.
It looks with what is currently implemented you'd have to drop the tables from the destination if you want repeated copies, probably not quite what you want. Close enough to your use case that it still might be worth testing it out though.
Standardized terminology is an awesome step. I'd love to see some of standardized file format for setting up the right keys on different devices. In the past I'd had annoying issues getting all the messages to decrypt on multiple devices, especially if I wasn't using the same client every device. Honestly though I suspect I was doing something wrong.
It is mentioned at the top of the readme, but the actual code changes appear to be on the branch v3.1.x-windows (also a few other branches with -windows in the name). Also the triton seem willing to collaborate as long as the patches sent are reasonably minimal and high quality https://github.com/triton-lang/triton/pull/4045#issuecomment...
Oh yeah that can cause lots of weird problems. I've run into that sort of issue a lot when cross-compiling, cause often then you might have a system copy of a library and a different version for the target, that can be a real pain.
You can, but only private forks of a private repo are allowed. Private forks of public repos are not allowed by design (modulo some weird bugs that were discussed on a past post).
The codereview of the WIP patches are a nice read. They do a good job of showing how quirky Cosmopolitan Libc is. To be clear I think it is awesome, just don't expect porting to it to be pain-free.
They still have the bug bounty for the Pixel devices. I don't know they ever had one for the open source Android OS (AOSP), but a bug in AOSP would be likely to effect a Pixel device. The bug bounty that is getting removed is one that google offered for certain very popular apps in the play store. I also see that have bug bounty's for some of the main Google android apps.
People have tested the 96GB DRR5 SODIMM kit and reported it to work with framework AMD laptop (framework lets you pick none for RAM, so you won't be paying RAM you won't use). Also as far as I know no one makes DDR5 SODIMM's with >48GB per stick.
I don't it is true that Windows tries to prevent\warn about the loading of a downloaded dll file and even if it did that wouldn't be a security boundary which is really what want to be relying on. https://www.microsoft.com/en-us/msrc/windows-security-servic...