Skip to content

Blog

import { Card, CardGrid } from ‘@astrojs/starlight/components’;

An architecture defense — why we chose stateless workers over stateful agents, and what we learned.
2026-04-27
An honest benchmark. Which model writes better Norwegian? We tested 12 tasks across both models.
Coming soon

More posts are in the works. Topics we’re exploring:

  • How we keep research swarms under $0.10
  • What happened when we ran 100 parallel swarms
  • Norwegian language and LLMs: capabilities and gaps
  • The debate topology: when 8 agents argue about ethics
  • Building your first swarm in 10 minutes

Want to write a guest post? Email us.