Interesting approach. Curious how you think about the tradeoff between external memory systems vs. models that manage their own context natively. It feels like the industry is split on which layer should own this.
Thanks. We think that while short term memory can be handled by LLMs directly and can be commoditised, it would be harder to do that natively for long term memory. We write about it in our article: https://medium.com/versanova/understanding-memory-in-ai-agen...