so not linux based ?

Funnily for me it's like "oh no shit ... oh well whatever"

i'll never understand how discord became the standard for gamers. Back in the day we used to use Teamspeak + IRC and that was great. Decentralised, low latency, etc. I still use teamspeak occasionally. Still works great. The discord app was just a mess from the very beginning. Resource hogging electron nightmare. Still everybody thought it was great. And now all sorts of communites, even open source and maker communities are on there. Makes 0 sense to me.

Ah i see then we were talking about different things. When i think of communism, i think of Karl Marx's definition. A stateless, moneyless, classless society. According to him, socialism is the transition phase between capitalism and socialism and if socialism is implemented correctly, the concepts of class, state and money just vanish over time. As far as i understand it, anarchists and communists completely agree about the end goal there.

That's why i purposely talked about socialism, not communism. Those former projects were socialist projects as well (or at least tried to be that, however you wanna see that).

I think this newer definition of communism vs. socialism largely stems from red scare anti communist propaganda and politicians with progressive agendas that try to distance themselves from those failed socialist experiments due to said propaganda. After all proudly proclaiming yourself to be a communist completely disqualifies you from any kind of political career. I would assume the label "anarchist" would have a similar affect 😅 . Even saying you are a "socialist" is still difficult. That's why the term "democratic socialist" emerged. I think that term is completely redundant. Socialism cannot be socialism if it is not democratic.

About communism 2.0 and protections against bad leaders. Supposedly that is what china is doing with their concept of promotions in the party purely based on how well they do on different objective metrics. If that is actually what is happening there, i don't know. At least in theory that seems better than being able to get into a position you are wildly unqualified for purely based on populism.

i’ll never understand this take. Socialism doesn’t work because of corruption and greediness so the best thing we can do is capitalism, which literally is based on corruption and greediness.

What about laws against corruption and greediness on top of a democratic system designed around production for human needs instead of the profit motive?

I mean on the one hand you have a system that breaks and ruins lives for most people if people work against it and corrupt it and on the other hand you have a system that benefits only a few and fucks over everyone else when it works as intended.

Saying socialism doesn’t work because of a few failed attempts is like the coyote dismissing every single of his ideas immediately after the first try and directly going to the next one except in this case the “solution” is to keep the system that has been shown for hundreds of years to work so badly that it is literally driving humanity towards extinction.

I meam you are not wrong in principle. What you mean with “capitalism with a shit ton of socialism” IS socialism. there is no capitalism with socialism because those two contradict each other. It is all about who controls the means of production. It is either a few rich assholes control production or the production is democratically controlled by society as a whole. The latter is socialism.

If you want markets and commodities etc on top of that, thats cool, can still be pure socialism though.

edit: outed myself as german “usw” == “and so on” 🤦

and of course 100% agree regarding fuck borders fuck states fuck fascism … that is communism btw 😁

fair point. I think however that the trajectory the USA is on would not have changed significantly even with the Democrats in power and the world would have maybe avoided at least a few tragedies. Not the big one in Palestine, obviously. Everyone seems to be on board with that one. But i don't know, to be honest with you. I just hope that that whole shit is over sooner rather than later ...

dump is always worse. It is chaotic, self destructive and pulls everything and everyone further down more quickly.

you know there are weight classes right?

the 27b model fits completely on the gpu so ram and cpu makes no difference.

For the deepseek model you need a fast ssd because no matter how big your ram is that model won’t fit there since it is almost 300gb.

I’ve got a ryzen9 5900x and 40gb ram. I don’t know how much of an impact the cpu actually has because i think most of it is evaluated on the gpu but could be that the router that selects the experts is executed on the cpu.

in any case the main impact will probably be vram size, pci-e throughput and ssd read speed (and maybe cpu speed).

qwen3.8-ridge-gguf runs great in ollama on the 9070xt in 4bit kv cache quant at 35t/s. You can also run deepseek 4.1 flash in strata if you have time. both models are very capable.

I have nothing really to compare it to in terms of the big ones because i refuse to give anthropic/openai any money. But i have used this qwen model in opencode to port an old CUDA/Optix based project to HIP and that worked fairly well. Not an overly complex refactor to be honest because 99% is just replacing the cuda prefix everywhere with hip, but it did fix the CMakeLists.txt for me and ported the optix prime acceleration structure to hipRT as well. I also let it vibecode a react based portfolio page because i have absolutely 0 knowledge of web development and refuse to learn it because i think it is insane and that turned out perfectly fine as well.

So it is absolutely usable for real world projects and the performance is quite OK as well with ~35tok/s. If i up the context size to the full 256k window, the performance drops to 5 tok/s, which is ok, if i let it do a boring refactor over night but at that point it isn't faster than running DeepSeekV4.1-Flash-1017 through Strata and that one is much closer to the current SotA.

Definitely exciting times. We are close to being able to run very capable LLMs locally that are more than good enough for development. I am personally not a fan of 100% vibe coded projects anyway, because i like to architect things myself, so letting it go in smaller steps and/or help be with the boring stuff is absolutely fine. Unless the project is some personal stuff that just needs to solve one simple problem then vibe coding is fine imho.

I've been running the Ridge-GGUF version on my RX 9070 XT with ~60-80k contextsize with q4 kv cache quantization. Everything fits completely in 16GB VRAM