Users claim Codex harness, not the models, is the bottleneck: zcode 5x faster on same tasks
bdsqlsz · x · 2026-10-09
- Developer @fly3nn reports that slowness attributed to models (GLM, DeepSeek) may stem from the Codex harness itself.
- Running GLM/DeepSeek through a magpie proxy inside Codex was extremely slow; switching to zcode made the same tasks finish in 2 minutes vs Codex's 10 (context compression alone took 3 minutes).
- bdsqlsz amplifies the finding, questioning whether the harness — not the model — is the bottleneck, and criticizing the Codex team.
More from coding & agent
- OpenEnv Arena scales free training jobs 10X, adds metrics and artifacts — ben_burtenshaw · 2026-10-09
- Wes Bos hacks $50 Amazon photo frames into a family dashboard with a coding agent — natesiggard · 2026-10-09
- Real-time fraud detection in Java: Kafka Streams plus vector similarity, explained — blaizedsouza · 2026-10-09
- Systems programming with AI coding agents: how much code should you still write yourself? — blaizedsouza · 2026-10-09
- Simon Willison ships a blog feature by voice, coding with Codex while cooking dinner — athyuttamre · 2026-10-09
- XState mulls optional state-local context for temp, type-safe data in state machines — DavidKPiano · 2026-10-09