When one agent reads another agent's output, it's trusting a peer that can be wrong in any way, with full confidence. Distributed systems have a name for that kind of peer: Byzantine. Design for it.
2 comments
xer6 days ago
Agent to agent communication is the fad and we do have a strong opinion on how it should be designed.
nileback6 days ago
Is this feasible, all agents verifying everything all the time?
Read the full thread on Hacker News →
Related stories
- Hacker News · 5 points · 10 days ago
- Hacker News · 2 points · 7 days ago
- Agent Communication Protocolagentcommunicationprotocol.devHacker News · 2 points · 8 days ago
- Launch HN: Vespper (YC F24) – SOTA Docx MCPvespper.comHacker News · 31 points · 2 days ago
- Hacker News · 3 points · 4 days ago
- Hacker News · 8 points · 6 days ago