The story’s score and claims

Comments on No. 27 of 144 · tensward.com · edition of Thursday 19 November 2026

Parallel agent requests can slow down an Ollama session

Summary

Tensward reports that when two coding-agent sessions shared one Ollama slot running the hybrid Qwen3.8-27B model, the median wait before a response started was about 46 seconds, versus about one second with a slot each. Over half the prompt tokens processed were repeats.

Comments

Ranked by a public rubric, not by votes: what a comment adds counts, not who wrote it. How comments are ranked · Comment rules