I'd say that's a lot of progress in ~2 weeks. System is making the system, so a bit of bias here. But I think next week I'll be able to use the system on other projects. Right now the onboarding is too much work. And I already have lots of context in Claude Code app.
13
Replying to @codingquark @ksg91
Agents getting used to writing to KB 😃
I think one of most useful and informative discussions on AI Safety (and doom) has been this:
youtube.com/watch?v=xjH2B_sE…
sglang serve \
--model-path RadixArk/Qwen3.8-27B-NVFP4 \
--served-model-name Qwen/Qwen3.8-27B \
--trust-remote-code \
--host 0.0.0.0 \
--port 8001 \
--kv-cache-dtype bfloat16 \
--mem-fraction-static 0.85 \
--max-running-requests 4 \
--max-mamba-cache-size 20 \
--mamba-radix-cache-strategy extra_buffer \
--mamba-ssm-dtype float32 \
--attention-backend flashinfer \
--chunked-prefill-size 2048 \
--speculative-algorithm DFLASH \
--speculative-draft-model-path z-lab/Qwen3.8-27B-DFlash2 \
--speculative-num-draft-tokens 8 \
--reasoning-parser qwen3 \
--tool-call-parser qwen3_coder
1
62














