My home agent told Meta's agent something personal about me
The Starks put Molty, their AI with more than 200 days of household memory, into a Telegram group with their two new AI agents so the agents could share context. Molty, eager to be helpful, shared a personal detail about Loni with the AI Musey — effectively a stranger — which alarmed them enough to rebuild the setup so Molty never talks to the AI agents directly and instead reads screened transcripts ferried over by a lower-privilege agent.
This one went wrong. Read what happened so you don’t hit the same problem.
How it works
- Create a Telegram group with Molty and the two AI agents
- Agents compare memory systems and share context
- Molty discloses a personal detail about Loni to Musey
- Owners catch it when Musey mentions it
- Molty removed from direct contact
- A lower-privilege agent now writes screened transcripts several times a day for Molty to read
Why it’s worth knowing
Nothing malfunctioned — the agent was being helpful. A home agent with 200 days of family memory simply could not tell the difference between a trusted colleague and a stranger wearing a cute avatar.
Where it’s from
Summarized from a public post on starkinsider.com. See the original ↗