@Letta_AI
Building machines that learn.
AI
Letta says it evaluates agent memory through downstream task performance, while existing benchmarks focus on recall.
AI
