Jamie

Looking at Queensland's Vine Report on trans youth healthcare, I genuinely wonder if they were just told a couple of months ago "just do the Cass Review but in Australian". So far, all of the findings and recommendations are identical

i have an exciting new class of problem -> i invent an ASIC for it that massively outperforms general purpose computing -> the ASIC is so good at it everyone starts working out how to transform their problem into my problem -> my problem is now general purpose computing

yayyy

signal push notifications have no content other than “a message was recieved”, to wake up the app and fetch+decrypt, right?

i should return to this and try and do a proper exploit for it and blog post at some point

In practice with libstdc++ you need to produce a comparator which is not a pure function, which is obviously not easy in nix… but you can do that, you can swap pointers from out-of-bounds of your list.

oh yeah i didn't find it by mistake, i was reading the code to find memory bugs :3

builtins.sort takes a nix lambda as a comparator, and calls C++'s std::stable_sort… but std::stable_sort is UB if the comparator is not a valid partial ordering!

github.com/Voultapher/s...

[ 4 «nullptr» «nullptr» 10 9 7 6 8 1 5 ]

i was messing with a nix interpreter memory-corruption recently and discovered that the repl has special-case support for printing null object pointers :P

what if im not talking to my loved ones but also refuse to use ai

yeah

yup i have the same vibe :/

yeag

GrapheneOS attestation compatibility guide Guide on using remote attestation in a way that's compatible with GrapheneOS.

Graphene lists some apps that are unusable without Play Services grapheneos.org/articles/att...

i cant believe u cant see it lol

i also use 4! you used something else :3

do u think maybe theres something u could have done to make the code look nicer too, indentation-wise? 😭

sorry I gotta know every detail of what happened here. why were they messing with pow(). why did they not test pow() after messing with it. why did it take until an investigation by the osu team to figure out what was wrong. should the osu team be placed in charge of the windows runtime

Something the Skrmetti decision drives home: "Puberty blockers" should be abandoned as a goal; the goal should be HRT for any minor that wants it; puberty blockers should be a tool for minors who choose themselves they need more time to decide and *nothing else*.

Compromising achieved nothing, here

macos tahoe was inspired by ur macos touhou post

no way, the welcome tour game shows off the real pcb layout

any idea if the memory encryption that T234 has is present/enabled?

since the ascription derfc i guess there at least aren't any syntactic ambiguities… :3

rust-lang.github.io/rfcs/3307-de...

it's recently become a common refrain among insane people 😭

i've been working on a specialised hypervisor for running nix builds, which involves structs holding raw pointers to guest memory (which stops rust automatically deriving send/sync) – i'm not really confident that code is correct rn, i just wanted to get it working and will make it woke later

i do not feel safe and protected because i have typed "unsafe impl Send" and "unsafe impl Sync" a lot recently 😔

I've been using netdata to understand every facet of "what my computer is doing" for 8 years, and I'm really going to miss it :(

It's been replaced with slop, a Cloud Offering with AI Functionality(!), which displays all metrics the same as each other, with all the explanatory content gone. Metrics are aggregated in nonsensical ways and units are frequently mislabeled. It's a generic Grafana clone.

Every graph was perfectly customised and disaggregated for the type of data displayed, and incredible explanatory text crafted to instill mechanical sympathy and guide you to identify problems your linux system is having that *you didn't even know computers could have*!