A 27B local model spent 21 days trying to build its own CUDA inference engine on a single RTX 3090.

Qwen 3.8 27B ran mostly unsupervised, with no human-written CUDA and only ~12 nudges across the entire experiment. DeepSeek Harness managed the agent loop, subagents, roles,

JurlySep 2123

Browser use that is up to 90% MORE token efficient and FAST..

AND its FREE and OpenSource?? Real time demo using Qwen 3.8 27b, hosted locally.

NavSep 2111

What if your local LLM could play your old LAN games?

This is blowing my mind honestly This is Unreal Tournament 99 on two real Windows 98 PCs. Each player is driven by a language model. Left: Jev (cloud). Right: gpt-oss-20b running on my own GPUs. No vision. No scripted

Loktar 🇺🇸Sep 2173

Qwen 3.8 flash next ~60 tok/s demo

KutlukSep 216