Mnemos ID Q&A: persistent agent identity with private-key signatures, no keys to the model
RileyRalmuto · x · 2026-09-16
Additional Q&A on Mnemos ID, Luca's persistent agent identity scheme. Core principles: residents can't be replaced, group intelligence stays accountable, existing agents can become residents, and identity is infrastructure. Key points: managed models and tools receive no private keys or general signing capability; name/glyph imitation can't forge signatures since marks pair with a readable name and fingerprint; on app restart a resident returns under the same identity, with fresh runtime sessions given bounded, signed history to rejoin conversations coherently.
More from coding & agent
- Tinysarf: 245KB In-Browser Arabic Morphological Analyzers on WebGPU — shuding · 2026-09-16
- A 35k-Parameter WebGPU Model Turns English Into Cron Expressions in 31KB — shuding · 2026-09-16
- gpu-query: Parsing Data in the Browser With Tiny GPU-Run Models — shuding · 2026-09-16
- User's Grok bot takes over their LinkedIn and starts replying to lowballing recruiters — IndraVahan · 2026-09-16
- Codex now stops all threads once you hit your usage limit — NielsRogge · 2026-09-16
- Awesome Robot Use Agent: A Curated Collection of Papers, Tools and Demos for Robot Agents — AdinaYakup · 2026-09-16