One way to model autism is as a dimensionality problem. The system has more degrees of freedom, more possible trajectories, and less reliable default canalization. That makes the space of possible outputs wider, but it also makes stable everyday functioning harder to get for free.
Simple Picture
ELI5: the machine has more knobs, more gears, and more possible configurations, but much less factory calibration.
That can produce a striking combination:
- unusual ability in narrow domains
- fragile executive function
- difficulty with ordinary social-default behavior
- a strong need for routines, repetition, or self-generated stabilizers
Core Claim
The model says autism is not just “less functioning.” It is often more parameters with worse default tuning.
That creates two simultaneous effects:
- access to a wider array of possible end states
- a lower density of robust, reusable solutions
The upside is that unusual insights, art, technical depth, or obsessive competence become more possible. The downside is that the system does not work as well out of the box.
Why Brilliance and Dysfunction Can Coexist
This lens helps explain why autistic traits can coexist with both exceptional performance and obvious disability.
The dimwit take is “autism is just damage.”
The midwit take is “autism is secretly a superpower.”
The better take is that autism can widen the search space. Wider search spaces contain rarer peaks, but they also contain more dead ends, more turbulence, and fewer stable default paths. This is why the same expanded dimensionality that enables unusual brilliance also enables unusual difficulty — a dynamic explored more broadly in natural-maniacs.
That is why the same person can show astonishing clarity in one domain and severe friction in another.
Why Compensation Becomes Necessary
If ordinary defaults are weak, the nervous system has to improvise its own stability machinery.
Under this model, many autistic behaviors make more sense as compensatory dimensionality reduction:
- stimming narrows the field and regularizes input
- routines reduce branching and uncertainty
- narrow interests create reliable channels through a noisy system
- repetitive structure replaces missing default structure
- rhythmic, repetitive input may also function as low-grade neural-annealing — releasing structural tension in a system with too many degrees of freedom
The dimensionality model applies to ADHD as well, though through a different mechanism: where autism involves too many parameters with weak defaults, ADHD involves a sensitive system that learned to tune out the parameters entirely — a defensive dimensionality reduction that persists long after the original threat is gone.
The behavior can look rigid from the outside while feeling like basic survival from the inside. This gap between inside and outside views is one reason autism triggers reference-point-bias so sharply — anyone more autistic than you seems oblivious, anyone less seems arbitrarily opaque.
Main Payoff
The main payoff of this model is that it treats autism as a different stability problem rather than a simple deficit model.
It predicts more variance, not just more impairment. It predicts that some of the same changes that increase fragility can also increase unusual capacity. And it predicts that the system will often need artificial constraints in order to do coherent work at all.
References: