Vehicles
In 1984 the neuroscientist Valentino Braitenberg published a slim book of thought experiments, Vehicles: Experiments in Synthetic Psychology. Each chapter builds a small machine: a body, two wheels, a sensor or two, and wires running between them. Then he sets it loose and watches.
What happens next is the point. Wire a vehicle one way and it shies from the light. Cross the wires and it charges straight at it. We call the first one afraid and the second one aggressive, and we can't help it. Nothing inside has changed but the routing. The fear and the aggression are ours, supplied the moment motion starts to look like purpose.
That is Braitenberg's quiet provocation: behavior that reads as mind arrives long before anything like a mind is there. Forty years on, surrounded by systems that talk back, it is one of the most useful ideas to carry around.
These essays rebuild the vehicles one at a time, as working models you can poke at. Move the light, cross a wire, add a threshold, and watch a temperament appear. The machines grow richer as the series goes on. The question stays the same: how little does it take?
Why Braitenberg Still Matters
Simple sensorimotor loops still explain more than we admit.
A short opening essay on why Vehicles still matters for AI, robotics, and software.
Open entry →Getting Around
A single sensor and a single motor are enough to trigger stories.
Restlessness, temperature fields, and the speed with which observers hallucinate intention.
Open entry →Fear and Aggression
Cross one wire and approach becomes avoidance.
How same-side and crossed excitation create radically different behavior.
Open entry →Love
Inhibitory wiring makes machines linger, orbit, and appear selective.
Tiny changes in coupling can produce behavior that looks social.
Open entry →Values and Special Tastes
Preference lives in the transfer curve.
How tuned nonlinear responses begin to look like taste and aversion.
Open entry →Logic
Thresholds are where mechanism starts to look rule-like.
Simple gates, compound triggers, and the illusion of deliberation.
Open entry →Selection, the Impersonal Engineer
Competence can accumulate without a planner.
Selection pressure as a search process for viable mechanisms.
Open entry →Concepts
Categories can emerge from use before they exist as symbols.
Concept-like behavior from recurrent structure and selective response.
Open entry →Space, Things, and Movements
A world model can be implicit in the loop, not explicit in memory.
How sensorimotor coupling starts to look like spatial understanding.
Open entry →Shapes
Form can be inferred from how interaction unfolds over time.
Shape perception as accumulated relation, not static inspection.
Open entry →Getting Ideas
Novelty often looks like recombination before it looks like insight.
Internal activation, recombination, and the mechanics of idea-like behavior.
Open entry →Rules and Regularities
Pattern sensitivity is a mechanism before it is a theorem.
What it means for a simple system to become responsive to regularity.
Open entry →Trains of Thought
Persistence and chaining make sequences feel cognitive.
How recurring transitions and retained activation begin to look like thought.
Open entry →Foresight
Prediction can appear long before explicit planning does.
Anticipation, delayed effects, and mechanism that feels prospective.
Open entry →Egotism and Optimism
Bias is one of the fastest ways to make a system feel human.
Asymmetric evaluation, self-reference, and observer-friendly distortion.
Open entry →