On: Cache-to-Cache: Direct Semantic Communication Between LLMs (2025)

[ayewo]

Sounds similar to Ramp's Latent Briefing for multi-agent coordination.

https://x.com/RampLabs/status/2042672773747589588