Not really, llama.cpp can download for quite some time, not as elegant as ollama but:

    llama-server --model-url "https://huggingface.co/bartowski/DeepSeek-R1-Distill-Qwen-32B-GGUF/resolve/main/DeepSeek-R1-Distill-Qwen-32B-IQ4_XS.gguf"

Will get you up and running in one single command.

And now you need a server per model? Ollama loads models on-demand, and terminates them after idle, all accessible over the same HTTP API.

Which is not the same model, it's not R1 it's R1-Distill-Qwen-1.5B....

A distinction they make clear and write extensively about on the model page, yes?

wheres that made clear in "ollama run deepseek-r1” the command to download/run the model?

Which you have to go to the model page to find.

Why would they want to sell ?

And what are they going to sell? The weights and the model architecture are already open source. I doubt the datasets of DeepSeek are better than OpenAI's

plus, if the US were to decide to ban DeepSeek (the company) wouldn't non-chinese companies be able to pick up the models and run them at a relatively low expense?

I just set up systemd-timers to nix-gc/docker-prune daily.

Still a bit of an annoyance, but one I don't notice once it's set up.

Great idea. I don't know why I didn't think of that.

I have my seedbox behind hideme vpn. They seem not that good privacy wise, but for torrenting that's not the main concern IMO.

Port forwarding works and you can get a raw wireguard config to dockerize it.

For general use in e.g. open hotspots I still use mullvad/mozilla vpn as I trust them more. (And can pay cashfor mullvad)

They dropped 2 nuclear bombs on, mostly on innocent civilians, seems like pretty good revenge to me.

And even before that, burned > 500 square miles of Japanese urban areas to ash.

Granted, by the last weeks of the war, the USAAF was dropping leaflets warning they'd be back with incendiaries in a few days, a masterstroke in psychological warfare.

They hiked prices massively so I wanted to transfer away, it was a massive shitshow.

Auth-codes given on the website were expired and they took 2 weeks to give me the correct ones near the end of the registry period.

Support was extremely unresponsive. As this this was a side project I couldn't spent time on every day my domain went into quarantine for a short time. They answered 2 days before the end of the rental period, when requesting the auth codes ~2.5 weeks before.

Will never use them again after this experience.

Porkbun is my new home for most stuff and domains.lt for .lt which porkbun doesn't offer yet sadly.

Wow, I never knew that to be the case! How would I find a registrar that supports .at LTD? Cloudflare, AWS, Google—neither supports it.

From looking through https://www.nic.at/en/my-at-domain/at-partnerfinder the only one that I knew and did not hear shit about seems to be OVH.

Others might be good, but no idea who exactly, many small unknown companies in the list that could be either great or shitty.

Thanks for Porkbun suggestion, I'll keep that in mind; it doesn't support .at but I'm now tempted to move my other domains there. Gandi used to be good, it's a shame what it's become.

I'm using dynadot for my .at domains

Dank

The last time i tried to get pi running with a distro that wasn't raspbian it was huge PITA, with having to use custom kernel patches and weird proprietary firmware and annoying boot process, am I misremebering or did this change ?

On any x86 UEFI system I tried i never had any issues at all, but Pi's I remember as a huge PITA...

Or is this just a general ARM issue ?

> Or is this just a general ARM issue ?

I understand that this is the case. IBM did the PC world a solid by not encumbering the original H/W. That provided a basis for a platform that could be copied and provide a "standard" that others could leverage, including copying and reverse engineering the BIOS that launched the OS. That provided the foundation of the PC explosion. They tried to reverse their "mistake" with the Microchannel Architecture and that saw limited success (AKA failed.)

IBM was also part of a consortium that tried to develop a common platform for the Power architecture. I'm not sure if any of that survives in their current offerings but it did not take anything away from the Intel (Windows) market and except for Apple, the Power PC was nowhere to be seen.

ARM processors have proliferated in the hand held (phone, tablet) market where volumes are sufficient for the manufacturers to provide their own solution to bootstrapping the OS. There are some efforts to agree on a common boot architecture (Yocto, UEFI and probably others) and I'm not familiar with them, but I hope one succeeds so we can easily load a variety of OSs on our devices.

Incidentally... The GPU in the Raspberry Pi performs the initial stages of the boot process. Weird. I think that Pi 4/5 generations are moving away from that but I'm not familiar with the detains. As near as I can tell, they've preserved the S/W structure required for the GPU to perform this function.

IMHO it's just that it's a beam VM language, which is a fatter runtime/ecosystem than is really needed to achieve the goal stated above can bring it's own bag of problems (but also it's own superpowers).

Also to be productive you have to utilize the rest of the erlang ecosystem, so at least some superficial knowledge in elixir & erlang is helpful for for some use-cases.

Syntactically I actually don't think it's that for off, but I dunno what GP was thinking, maybe that it leans more into functional patterns & sugar for those whereas rust/go can also be used in a procedural style. (Though at least personally I am using way more functional patterns in rust than I expected)

Sure, there's "native" UI with stuff like slint(QT-like) or iced(system76 is actually building the COSMIC linux desktop environment with this).

And you can also get an electron-like stack going, that is actually much less bloated than "normal" electron, by using tauri-webview, which uses the OS-provided webview and combining it with one of the many cool rust WASM-based reactive web-ui frameworks, like leptos or dioxus. This gets you compiled sizes of ~10s of MB compared to electrons 100s of MBs.

There's also bindings to a lot of traditional ui libs, like GTK, QT & Tk.

I'm currently going for the 2nd option (with leptos for the web part) as I'm used to the web-stack and am very productive with this approach, but native UI also seems very tempting to dig into further.

Some related links:

- https://www.arewewebyet.org/topics/frameworks/

- https://areweguiyet.com/#ecosystem

I can personally vouch for Slint. The DSL is very pleasant to work with, it's still evolving (and the devs are very active and responsive), tooling is also getting better and better each release and it supports translations and accessibility already (through AccessKit). I've been (veeeery) slowly working on a Matrix client with it and I've quite enjoyed it so far.

I'm also developing a chat client (tho for LLMs) using QML. Here's a little video showcasing it: https://custom-downloads.s3.us-west-2.amazonaws.com/chat_cli...

Still, very much in its infancy.

The backend is currently C++ but I'm considering using other languages since there are so many bindings (Mojo is on my radar).

Looking really good so far, keep it up! Is the source available somewhere I could take a peek? o.o

Thanks! Unfortunately not, it's been too hard for me to monetize FOSS projects, so it's going to be closed source.

Very understandable, it is indeed quite hard :(

Best of luck! I'll keep an eye on it ;)

Thank you! I'll keep en eye on Eigen too! I'm curious how it will work out with Slint.

Hey that is really cool thank you. I have Sublime as my IDE, and it looks like I can use that as well, and using a webview might leverage my understanding of webtech.