Intelligence index
Field note 002 / 010 5 min
FIELD NOTE 002

From Jim Fan · NVIDIA GEAR

Robots that think
fast and slow

Kahneman's two systems map cleanly onto a robot's brain: a slow, deliberate planner and a fast, intuitive body. How the two talk to each other is still open research.

Plan at one hertz. Act at a thousand. A robot mind needs both tempos at once.

Borrowing from Daniel Kahneman's Thinking, Fast and Slow, Jim Fan splits embodied intelligence in two. System 2 deliberates: reasoning, planning, and writing code. Large models already do this well. System 1 executes: fast, intuitive motor control that never reaches conscious thought. Grasping a cup, you do not decide where each fingertip goes at every millisecond.

01 SLOW

System 2

Deliberate: reason, plan, write code.

The slow mind runs on big models. Vision-language models and LLMs in a loop can already reason about a scene, decompose a task, and even generate code to orchestrate behavior, deciding what to do next at roughly one decision per second.

Superpower
Long-horizon reasoning
Blind spot
Seconds of latency

A model that thinks in seconds cannot close a control loop that physics demands in milliseconds.

02 FAST

System 1

Intuitive: act before you can explain it.

The fast mind is reflex. Whole-body balance, grasping, and contact-rich manipulation need decisions at something like a thousand hertz. This points to compact, fast sensorimotor policies rather than giant models.

Superpower
Reflex-speed control
Blind spot
No deliberation

A reflex cannot plan a multi-step task or reason its way through a situation it has never seen.

Two tempos, one body.

System Tempo Mode Runs on Breaks when
System 2 ~1 Hz Deliberate Large VLMs & LLMs in a loop Contact needs millisecond reflexes
System 1 ~1 kHz Intuitive Compact sensorimotor policies The task needs long-horizon planning

The interface is the research.

Fan frames the unsolved part as architecture: one monolithic model or a cascade of specialized ones. If it is a cascade, do the systems communicate through text or through latent variables? Text is interpretable but slow and lossy; latents are rich but opaque. NVIDIA's later GR00T N1 made the framing concrete: a vision-language module for System 2 driving a diffusion-transformer action module for System 1.

  1. 01
    MonolithicOne end-to-end model: cleaner, harder to control
  2. 02
    CascadedSeparate models per system, wired together
  3. 03
    The wireText or latent vectors between the systems?
  4. 04
    The clockBridging 1 Hz deliberation and 1 kHz control

Own the reflexes. Rent the reasoning.

General-purpose System 2 reasoning is fast becoming a commodity available from foundation-model providers. The defensible layers for a robot maker are System 1, with embodiment-specific, safety-critical, kilohertz-rate control, and the interface contract between the two systems.

That contract is also a safety boundary: while System 2 deliberates for seconds, System 1 must stay competent and safe on its own. Designing the handshake, including skill APIs, latent commands, and interruption semantics, is product work, not just research.

Keep this

System 2 sets the goal. System 1 does the touching. The handshake is still research.