System 2
A concept derived from Daniel Kahneman's dual-system theory, referring to slow, deliberate thinking that requires focus and reasoning. In the AI field it corresponds to reasoning models that generate a chain of thought before answering.
Timeline (last 90 days)
System 2 is the same layer stack running repeatedly in a loop, writing down the result of each step and reading it back in the next round—this is the essence of chain of thought.
The recommended collaborative system design is: use a System 1 one-glance model to answer all questions while retaining confidence scores, and hand off tasks below a threshold to a System 2 thinking model.
A System 2 model is essentially the same network running repeatedly in a loop, forming a thought process through a 'write notes–read notes–advance' scratchpad mechanism.
System 2 models with recurrent reasoning capability can maintain performance as chains grow longer, because they can explicitly record intermediate conclusions at each step.
The practical criterion for choosing a model is to count how many steps the answer requires: use System 1 for one-step answers and System 2 for multi-step answers.
Reasoning models have introduced mechanisms such as chain-of-thought, reflecting the evolution of AI from System 1 fast intuitive responses toward System 2 slow deep thinking.
All Facts (6)
The practical criterion for choosing a model is to count how many steps the answer requires: use System 1 for one-step answers and System 2 for multi-step answers.
60%UnverifiedSystem 2 is the same layer stack running repeatedly in a loop, writing down the result of each step and reading it back in the next round—this is the essence of chain of thought.
50%UnverifiedThe recommended collaborative system design is: use a System 1 one-glance model to answer all questions while retaining confidence scores, and hand off tasks below a threshold to a System 2 thinking model.
50%UnverifiedA System 2 model is essentially the same network running repeatedly in a loop, forming a thought process through a 'write notes–read notes–advance' scratchpad mechanism.
50%UnverifiedSystem 2 models with recurrent reasoning capability can maintain performance as chains grow longer, because they can explicitly record intermediate conclusions at each step.
50%UnverifiedReasoning models have introduced mechanisms such as chain-of-thought, reflecting the evolution of AI from System 1 fast intuitive responses toward System 2 slow deep thinking.
50%