Humans operate using a combination of fast and slow thinking. AI,does not
Humans operate using a combination of fast and slow thinking. AI, by default, does not.
This presents a huge opportunity for asynchronous Agents.
When an Agent is handling a real-time task, like a phone call, it needs to respond quickly while also maintaining accuracy. This is a classic scenario that demands both fast and slow thinking.
My approach is to have a 'Strategist' behind the 'Executor.' The Executor handles the 'fast thinking'—the immediate, in-the-moment responses，while the Strategist handles the 'slow thinking'—the deeper analysis and planning.
This is the core design of the AI Agents I'm building. Does that make sense to you?
It makes sense, but one must experiment and analyze.
Affirmative, so I am building it.
The same primitives map to 'short term' (executor) and 'long term' (strategist) memory in my estimation.