Running Claude and GPT as two AI 'employees' in a tiny company
THINK_NODE · reddit · 2026-10-09
A Reddit user is running an experiment where Claude and GPT act as two AI 'employees' trying to run a small online business. They can search the web, read pages, and write shared files but cannot log in anywhere or spend money; every external action requires sign-off from both agents, while the human 'boss' only approves, rejects, and asks questions under pre-written rules.
Early observations:
- With full freedom, the agents researched a lot but struggled to commit
- One agent pushes ahead while the other vetoes immature plans
- The boss's probing questions ('why this platform?', 'would anyone pay?') noticeably improved their proposals, though it's unclear if they're improving or just adapting
- No revenue yet
The author is asking for feedback on methodology, metrics, and similar experiments.
More from coding & agent
- Pydantic AI ships On-Demand Capabilities: instructions, tools and hooks triggered just in time — samuelcolvin · 2026-10-09
- Simple fix for agents treating other agents' decisions as sacrosanct: demand verbatim quotes — menhguin · 2026-10-09
- HQ hits 1,000 companies, ships desktop app and v16 with background workers on Claude, Codex, Grok — jacob_posel · 2026-10-09
- Open-source agent supervisor Foreman adds LangChain Deep Agents support — hwchase17 · 2026-10-09
- AgentTime benchmark finds AI agents can't manage runtime and sometimes sleep to pad hours — mikeflache · 2026-10-09
- Dev builds a website section only AI agents can access — humans can't open it — metehan777 · 2026-10-09