“Reasoning memory is the novelty from this architecture. By storing past successful or failed thinking patterns into the memory, the agent can one-shot future similar requests or at least know not to repeat similar mistakes.”
this was really interesting. great real all around, i really enjoyed how deep you went in the weeds with this one. thanks.
“Reasoning memory is the novelty from this architecture. By storing past successful or failed thinking patterns into the memory, the agent can one-shot future similar requests or at least know not to repeat similar mistakes.”
this was really interesting. great real all around, i really enjoyed how deep you went in the weeds with this one. thanks.
thanks man! Their take on agent memory is dope. One of the best I've seen so far