Nvidia put a strix halo in a laptop. Or a dgx spark basically. When that bitch is $1200 I'd go for it. But that will be 5 years latter when I want something bigger and faster.
alphakamp
null
npub1e3fl...3mjr
null
Vsauce here, here is the sound rabbits being killed. The FBI played it to the branch davidians. Isn't that neat.
sloptimus prime vs the decepticons
View quoted note →
I need more compute
They can't have GPU, we can't have model😕🙃
View quoted note →
🤔 what China doing
Really wish I hadn't been up till 3 am last night. Oof
chatgpt agent has a guardrail for agents reading an agent log and will refuse to read thinking logs, because it is private.I switched over to kimi, it did it immediately
Mood while waiting on agents
View quoted note →
YOOOO I went from qwen3.8 27b on llamacpp at ~50t/s with 128k context to qwen3.8 flash-next on strata at ~115t/s with 256k conext LETS GOO!!