Brett

What are you using to run the model? Llama.cpp will automatically split the model between your system ram and graphics card’s vram.

Qwen 3.6 is a mixture of experts model with only 3B parameters active at a time. Even without quantization your card could easily run that.

Nice callback

Is that quantized? 4 bit Qwen 3.6 can get 22tps on a 1060.

Saved. Awesome stuff 🥕

This is a fantastic post. I am very happy to have seen it tonight.

I agree with everything said here and appreciate your nuanced take, but:

Capitalism’s obsession with wealth as the primary means of accumulating power means the old notion of noblesse oblige, wherein the social contract could be invoked against the elite, is greatly weakened. Instead, our social contract is directly with the state, creating a ‘middleman’ situation wherein we struggle to make demands directly against capitalist elites, who are not reliant on our direct cooperation the way feudal and clientistic societies are. The capitalists do not need goodwill - the state they rely on needs goodwill. The capitalists need only cold, hard capital.

Is this not at least partially counter-balanced by labour unions?

Canada will add bike lines Not in Ontario while Doug Ford is in power (at least 4 more years).