Anthropic says the key to agent memory is knowing what to forget, not how much to store Anthropic found that selective note-taking outperforms exhaustive memory archives for AI agents, with its newer Claude Fable 5 model getting four times as far as the older Sonnet 3.7 in a Pokémon Red test by keeping just a handful of strategy files instead of more than 400. Anthropic also described a "Dreaming" process that retroactively corrects faulty memories. Anthropic studied how Claude retains knowledge across sessions. Selective notes beat exhaustive archives by a wide margin. In a Pokémon Red test, the newer Claude Fable 5 got four times as far as the older Sonnet 3.7 by keeping just a handful of strategy files instead of more than 400. A "Dreaming" process retroactively corrects faulty memories. The article Anthropic says the key to agent memory is knowing what to forget, not how much to store https://the-decoder.com/anthropic-says-the-key-to-agent-memory-is-knowing-what-to-forget-not-how-much-to-store/ appeared first on The Decoder https://the-decoder.com .