GrayShade

On systemd, you can use ~/.config/environment.d to set it. Don't rely on your shell.

Fedora is patched.

I get 20 t/s on the UD-Q6_K_XL quant, Radeon 6800 XT.

It only supports CUDA on Linux.

You can, after you unlock it.

We've had shitty bloated websites before LLMs were a thing.

It's plain FUD. systemd always had fields for the full name, email address and location. They were optional, just like the date of birth. Bad systemd!

The operating system does not need to know your full name, email and location in order to manage your hardware and software, yet systemd has had optional fields for those for years and nobody complained. They added an extra optional field for the date of birth.

Some of this has been fueled by a misinformation campaign that has targeted the systemd project and Taylor specifically, resulting in Taylor being doxxed and receiving death threats.

I see.

It's easier to torrent stuff than to get 4K in Netflix on Linux.

1.82.7 doesn't have litellm_init.pth in the archive. You can download them from pypi to check.

EDIT: no, it's compromised, see proxy/proxy_server.py.

It's gone: https://github.com/bjarneo/cliamp/commit/085734a86343a80176d...

Don't forget location.

It doesn't, and it's optional.

Looks like it's a bit slower today. Running llama.cpp b8192 Vulkan.

$ ./llama-cli unsloth_Qwen3.5-35B-A3B-GGUF_Qwen3.5-35B-A3B-UD-Q4_K_XL.gguf -c 65536 -p "Hello"

[snip 73 lines]

[ Prompt: 86,6 t/s | Generation: 34,8 t/s ]

$ ./llama-cli unsloth_Qwen3.5-35B-A3B-GGUF_Qwen3.5-35B-A3B-UD-Q4_K_XL.gguf -c 262144 -p "Hello"

[snip 128 lines]

[ Prompt: 78,3 t/s | Generation: 30,9 t/s ]

I suspect the ROCm build will be faster, but it doesn't work out of the box for me.

That's not a shell, it's a Python interpreter compiled to WASM and running in the browser.

This feels a bit pessimistic. Qwen 3.5 35B-A3B runs at 38 t/s tg with llama.cpp (mmap enabled) on my Radeon 6800 XT.

And footnote 3 is unreferenced.

It did, there's two incompatible approaches, zram and zswap.

The Signal device linking feature is just as fast. It's partly a trick -- it will look for QR codes even outside the central area, so under good conditions it can get a read before you even get a rough orientation.

Maybe that's the only API-visible change, saying nothing about the actual capabilities of the model?

This patch uses a DNS extension to report the running version and check for updates. The messages are encrypted in flight.

The feature is gated at build-time, and enabled in the packages.sury.org repo and in the ondrej PPA.

The indexer was probably clangd, not Code itself.

It used to be like that, but they've changed it to a POST request a while ago.

Hey! Sorry, I didn't get the chance to test it yet (like I promised when you launched), but can you say more about the rewrite? The title made me think you're porting it from Rust to another language :-).

I've written my own Git clients and have built a web server around Git repositories. I don't want to lose the hack-ability of Git.

And they will keep working because the repository format isn't affected by the language git is written in.

It doesn't say that:

In addition to redesigning and testing the Noctua fan grill, we also evaluated various other scenarios. These included replacing the NF-A12x25 with its G2 variant and incorporating an additional 8cm fan for exhaust purposes.

If you're wondering why the differences in the image pairs under the sliders are so subtle, try loading the page in a Chromium-based browser.

Not exactly the same thing, but see https://en.wikipedia.org/wiki/GNU_IceCat#Additional_security....

Who protects you when the packagers decide to trust a shady CA (adding it to the root store) because it's used by the distro's infra?

To be clear, the EU is also pushing for encryption and hardware backdoors.