Qwen3.8-27B spends 21 days and 180 subagents failing to beat llama.cpp
udmrzn · x · 2026-10-29
- A stress-test experiment had Qwen3.8-27B spend 21 days trying to build a CUDA inference engine that could outperform llama.cpp on a single RTX 3090. It didn't succeed — but the data is revealing.
- Scale: 21 days, 180 subagents, 230M tokens processed, 1.7B cache-read tokens, 699 context compactions (83 hours), and only 12 human interventions.
- DeepSeek Harness coordinated subagents, task handoffs and context limits; the model ran via vLLM with HyperQwen.
- Key takeaway: the gap is between an agent that can sustain a weeks-long software project and one that can actually hit a performance target.
More from coding & agent
- Video2Game: H3 Max Generates Gameplay Video, Claude Turns It Into a Playable three.js Game — OdinLovis · 2026-10-29
- After 3 Years of Heavy AI Prompting, Domain Vocabulary Beats Prompt Tricks — Dimillian · 2026-10-29
- Agents push compute back to CPUs: $221B server CPU market seen by 2030 — BenBajarin · 2026-10-29
- Cursor removes its beloved Cmd+K command, frustrating power users — jbhuang0604 · 2026-10-29
- Designing a skill like an app: a developer's iteration diary for interface-bts — round · 2026-10-29
- One prompt one-shots a walkable 3D ship where AI agents visibly work around the clock — InternationalGap3698 · 2026-10-29