Our nightly cloud agent got expensive.
So we redesigned it with DeepSeek V4 Flash + Qwen 3.8 27B on AMD Ryzen AI Max+. It is a hybrid agent stack built to use fewer cloud tokens. See how:
Watch anon! this is what 2x dgx sparks look like under load serving qwen 3.8 flash next, official fp8
Tensor parallel across both boxes over the connextx cable. and the hermes agent telegram stream sitting on top of it has not dropped a single message. two of these machine on
DeepSeek V4.1 Flash vs Qwen 3.8 Flash: The Cabin Comes Alive ☕❄️
Two cocoa machines. Two snow globes. Both models now have working outputs after their own repairs. Same briefs, camera paths and timing. Full recovery comparison, with earlier failures preserved. Rendered
Qwen 3.8 Max Full COURSE 1 HOUR (Build & Automate Anything)
You have reached the end of the archive
All of qwen38