dsp_person

I was running it with a VFIO passthrough firewire 800 card to run my RME fireface audio interface, mainly to control the graphical mixer.

It's amusing thinking if I rigged up an electron app for a mixer it would take up more memory than the entire disk+ram.

We need a steve.ai

Check out SmolXP [1]. The one I setup had a 340MB qcow2 image, took about ~10 sec to boot to desktop with qemu, and used only 58MB ram idling on desktop.

https://github.com/AFellowSpeedrunner/SmolXP

Funny cause I was just thinking about the tradeoff of "internal wasm app" vs "internal native app".

The former has convenient distribution, but worse performance and other limitations.

The latter can be tricky to keep updated, ensure the environment is the same for everyone and/or cross-platform differences, etc., but significantly better/faster.

But both binaries about the same size. Assuming using something like sokol or SDL3.

So is this aimed at small models only? Is there any advantages to these models compared to what I can run locally on a 16GB VRAM GPU?

Would be nice for something at the level of like Claude 3.5

2. Sometimes watching Claude do inline edits I cringe watching it delete almost the entire file line by line. I think when early in exploring a problem it's best to torch the previous version and fully re-write than to assume incremental.

> For a given dimension d, Klartag can pack d times the number of spheres that most previous results could manage. That is, in 100-dimensional space, his method packs roughly 100 times as many spheres; in a million-dimensional space, it packs roughly 1 million times as many.

Those numbers sound wild. For various comms systems does this mean several orders of magnitude bandwidth improvement or power reduction?

Anyone know of any mobile apps that have done this and bundled their own fork of syncthing under the hood for syncing?

Just a hair under google, it crossed earlier this week I think. This list is fun to look at https://8marketcap.com

Curious about this process. Can anyone submit patches to bcachefs and Kent is just the only one doing it? Is there a community with multiple contributors hacking on the features, or just Kent? If not, what could he do to grow this? And how does a single person receiving patreon donations affect the ability of a project like this to get passed bus factor of 1?

Looking forward to the tmux-c re-rewrite next

    if ((Π⁻ < Π) && (Π < Π⁺)) {

I spent way too long trying to figure this out as C code instead of

    if ((Π⁻ < Π) && (Π < Π⁺)) {

I spent some time hacking on things with dodo [1] but have fallen back to thunderbird for the time being.

I'd settle for something markdown-like rather than full-blown html. Basic headings, lists, and inline images are all I want.

[1] https://github.com/akissinger/dodo

smart tvs too busy adding spyware and AI bloat

maybe streamio + unified remote?

from phoronix:

    Noted changes with XLibre 25.0 include:
    a) Xnamespace extension: a novel approach for isolating clients from different security domains (eg. containers) into separate X11 namespaces, where they can't hurt each other

that seems significant, I wonder why xorg hasn't implemented something like it so far.

would you be happy running a few years old linux kernel missing security patches?

> In this paper, we argue for a triangle come back

Go team triangles!

he explains

~40y ago invented data warehouse, where an ETL process overnight would collect data from smaller dbs into a central db (the data warehouse)

~15y ago, data lake (i.e. hadoop) emerged to address scaling and other things. Same idea but ELT instead of ETL: less focus on schema, collect the data into S3 and transform it later

ah maybe there's a fully rootless option recently too

https://github.com/systemd/systemd/issues/30239 https://github.com/systemd/systemd/pull/26826/files

    systemd-nspawn may be invoked with or without privileges. The full functionality is currently only available when invoked with privileges. When invoked without privileges, various limitations apply, including, but not limited to

Ah maybe a significant advantage of incus is it doesn't require sudo. I remember that also being required for nspawn.

ooh i didn't know about vmspawn. maybe this can replace some of where I use incus

> you can always create zip bombs that are links on a web page that don't show up for humans (white text on white background with no highlight on hover/click anchors)

RIP screen reader users?

Also frustrating that UBO Lite just changed from "permissionless" to requiring broad see everything permissions.

Too bad xz/lzma isn't supported in these formats. I often get pretty big improvements in compression ratio. It's slower, but it can be parallelized too.

aren't abstract sockets un-jailable unless using network namespaces?

or in the other direction, to truly prevent e.g. xorg socket from being accessed by a bubblejailed application, it should exclude --share-net, regardless if you bind the actual path to the socket (since abstract permeates beyond that)

When I did my first github "release" I was shocked that it involved picking files to upload...

I've found useful doing an implementation with void* and having thin macros to do casting with typeof() so the usage is type checked.

Hmm can you explain? Would that be like line 1 starts out purple, transitions to light purple at the end of the line, and 2nd line starts light purple, transitions to pink, 3rd line starts pink and transitions to purple?

Better comparison would be fresh code in Rust vs fresh code in C. Re-writing in both languages following best practices testing and fuzzing. A big difference vs comparing to legacy C code that is trying to be maintained by throwing 0.25 google engineers at it to fix an ongoing stream of fuzzing issues.

In my e-reader I changed the font to Roboto Mono and completely forgot it was monospace. It's not like reading feels any less efficient. Idk might even feel more efficient or comfy somehow.