These days, we hear more about the future of neuro-symbolic architectures, mostly as they would mimic the “System 1 + System 2” model of human decisions as defined by Daniel Kahneman in his book “Thinking, Fast and Slow”.

We believe this is the direction that AI should take, and on that path, that LLMs must develop four capabilities to maximize their contribution: the ability to explain, to forget selectively, to learn in real time and to go back in time.

Our last paper of 2024 gives Causality Link’s point of view on these capabilities and how they will bring us closer to Kahneman’s model.

Click above to download the paper.