Compact memory to 50%

Resolved 💬 2 comments Opened Mar 22, 2025 by d46 Closed Jun 10, 2025

We get the most benefits when memory and context understanding are between 50% and 90%. When we compact, we literally lose everything, and for the AI, the compacted data makes no sense.

View original on GitHub ↗

This issue has 2 comments on GitHub. Read the full discussion on GitHub ↗