Swarm agent training needs thousands of agents, ambitious problems and cost math
samiramanabi · x · 2026-11-09
samiramanabi offers three principles for startups doing swarm training: (1) scale up — a swarm should have thousands of agents, since most learning issues only surface at that scale; (2) pick ambitious problems requiring unstructured communication and collaboration among agents; (3) account for the cost and benefit of units of work — real-world parallelism and hierarchies win because tasks carry deadlines and costs.
More from coding & agent
- macOS 27.3 lets Mac VMs use host GPU and Neural Engine, up to 7× faster Core AI — SumitGup · 2026-11-09
- Day 21 of an agent a day: a TPP monitor agent for drug discovery, sample prompt included — nlarusstone · 2026-11-09
- Applying the bitter lesson to my LLM workflow: skip agents, just talk to the model — esrtweet · 2026-11-09
- Apple engineers' recap for local AI builders: skip AFM for coding, one-line model swapping — rxwei · 2026-11-09
- copy-clip: a copy button that folds your code into paper and clips it, built with vanilla JS — prasenx · 2026-11-09
- Engineer builds Promptline, a one-click prompt library app for repetitive AI workflows — TallShift4907 · 2026-11-09