Neurodivergent cognition as competitive strategy in the age of average

Generative AI is a regression-to-the-mean engine. It is trained on the distribution of what humans have produced, it predicts the most likely next token, and at scale it pushes output — writing, design, code, strategy, decisions — toward the modal. The more AI mediates production, the more the median output becomes the default output. This is not a bug. It’s the math of a model trained on the distribution. And it creates a second-order consequence that strategy writing has not yet named: as AI homogenizes output, cognitive variance — the capacity to think genuinely off-distribution — becomes the scarcest strategic input an organization can hold.

I see this from two vantages simultaneously. As a builder of credentialing and trust infrastructure, I watch AI flatten the distance between competent and median work, and I think about where defensibility lives when production is commoditized. As a neurodiversity-affirming parent of an autistic child, I watch a cognition that diverges from the modal — pattern-recognition that finds connections the median mind misses, attention that moves nonlinearly, sensory and processing profiles that produce different outputs from the same inputs — and I think about what happens when the economy finally values exactly what it has spent decades pathologizing. Those two lenses converge on a thesis that neither the strategy world nor the neurodiversity world has fully articulated: in the age of average, neurodivergent cognition is not a cost-center accommodation. It is a competitive asset, and the organization that treats it as such captures a reservoir of cognitive variance that homogenizing AI cannot replicate.

The age of average

The claim that AI drives output toward the modal is well-supported and almost trivially true once you trace the mechanism. A language model is a probability distribution over tokens. It predicts the most likely next word given the context. At the level of a sentence, this produces fluency. At the level of a corpus — millions of documents, designs, codebases, strategies produced with AI assistance — it produces convergence. The outputs cluster around the center of the training distribution. The tail thins. The distinct, the contrarian, the off-median, the structurally variant gets scarcer.

This is not about quality in the narrow sense. AI produces competent, readable, functional output. The problem is that it produces the same competent, readable, functional output, with variations that are decorative rather than structural. Ask ten organizations to produce a competitive analysis with AI tooling and you get ten documents that disagree on details but share an underlying shape — the same framing, the same move-to-the-median strategic recommendations, the same hedged conclusions. The variance between them is noise, not signal.

The cost compounds with adoption. The more organizations use AI to mediate production, the more the modal output becomes the default not just within an organization but across the market. The entire output distribution narrows. And the thing that gets scarcest fastest is the genuinely off-distribution judgment — the contrarian read that turns out to be right, the pattern spotted from an angle the median doesn’t occupy, the decision made from a cognitive vantage that the modal consensus can’t reproduce because it doesn’t share the processing architecture that produced it.

Cognitive variance as the scarce input

Here’s the strategic claim: in an economy where AI homogenizes output, the scarcest input is not compute, not data, not even talent in the generic sense. It’s cognitive variance — the capacity to produce judgments and outputs that are genuinely off the modal distribution. And neurodivergent cognition is one of the largest underused reservoirs of it.

This is a reasoning-from-the-construct argument, not an RCT result. Let me be precise about what I’m claiming and what I’m not. Autistic cognition, in many of its profiles, involves pattern-recognition that operates at different granularities and with different salience weights than neurotypical cognition — seeing structure where others see noise, or missing social structure that others prioritize. ADHD cognition involves attention that moves nonlinearly, making connections across domains that linear attention doesn’t traverse. These are not universal claims about all neurodivergent people — the variation within neurodivergent populations is enormous. The claim is directional and population-level: cognitive profiles that diverge from the modal processing architecture produce outputs that diverge from the modal output distribution. In a world where the modal is cheap and abundant, that divergence has strategic value.

Not all variance is valuable. Some variance is noise. Not every off-distribution judgment is correct. The thesis is not “neurodivergent cognition produces better outputs.” It’s “neurodivergent cognition produces different outputs, and when AI makes the modal output abundant and cheap, the different output is the one with marginal strategic value.” The value is in the variance itself — the optionality of holding a cognitive position the market doesn’t share — not in every individual instance of it.

From DEI posture to competitive thesis

The dominant framing of neurodiversity in organizations is a DEI/ESG posture: include neurodivergent people because it’s the right thing to do. That framing is real and defensible. It’s also, from a strategic vantage, an underpricing. It treats neurodivergent cognition as a cost-center accommodation — something the organization carries because inclusion is good, not something it seeks because variance is scarce.

The competitive reframe inverts the question. It doesn’t abandon the moral case; it argues the moral case is weaker than the strategic case, and that leading with the strategic case is both more honest and more durable. The DEI frame asks the organization to carry a cost. The competitive frame identifies an asset. The DEI frame says “include them despite the friction.” The competitive frame says “the friction is a signal that you’re holding cognitive variance, and in the age of average, that variance is the moat.”

This is the gap nobody is naming. Strategy writing treats AI homogenization as a content-quality problem — “AI output is mediocre, we need better prompts, better evals, better human review.” Neurodiversity writing treats workplace inclusion as a moral and accommodation problem — “accommodate neurodivergent workers because it’s right and because the law says so.” Neither connects the two: AI homogenization → cognitive-variance scarcity → neurodivergent cognition as competitive asset. The connection is the thesis, and it’s the thing that makes the argument durable rather than contingent on the moral weather.

Why masking destroys the asset

This is where the strategy argument meets the lived-experience literature, and where the cross-site links do real work. The cognitive variance that’s the strategic asset is exactly what masking suppresses.

As the neuro subdomain’s piece on autistic burnout and masking in knowledge work lays out, masking is the performance of neurotypical cognition — the suppression of one’s actual processing architecture to produce the outputs the modal expects, in the shape the modal expects them. Masking is, literally, the production of modal output from a non-modal mind. It is the thing AI now produces cheaply. An organization that “includes” neurodivergent people but requires them to mask — to communicate only in neurotypical registers, to perform attention in neurotypical patterns, to suppress the cognitive variance that makes them strategically valuable — is paying the inclusion cost and getting modal output and losing the variance. That is the worst of all worlds. You’ve paid for a scarce cognitive input, then engineered the workplace to suppress it into the same output everyone else produces. This is the central institutional failure, and it’s not a DEI failure. It’s a strategy failure. The competitive thesis requires the protocol-design changes the neurodiversity-affirming pieces argue for; you can’t capture neurodivergent cognition as a strategic asset while running a workplace that suppresses it.

Consider the friction that gets neurodivergent cognition pathologized in engineering contexts: terse comments that read as hostile, literal communication that misses social lubrication, nonlinear attention that doesn’t track the meeting agenda. These are the exact behaviors that double empathy problems in your pull requests identifies — not as individual deficits but as protocol mismatches. The autistic engineer’s terse PR comment isn’t a communication failure; it’s a different communication grammar, and the team’s interpretation of it as hostile is equally a mismatch. The piece argues for structured PRs, async-first norms, and explicit communication protocols — not as accommodations, but as better protocols for everyone. Here’s the strategic point: those protocol changes are the mechanism by which the cognitive-variance asset becomes usable. The terse, literal, nonlinear friction is the same friction that, once removed from a social-penalty frame and placed into a protocol-design frame, lets the variance actually contribute. Structured PRs surface the signal. Async norms route around the social-processing tax. Explicit norms replace implicit consensus, which always defaults to the modal. Without these changes, the asset is suppressed into modal output and you’ve paid for inclusion and gotten homogenization. With them, you’ve built the infrastructure that lets off-distribution cognition produce off-distribution judgment — which is the whole point.

The tie-back: variance is the moat the modal can’t copy

This connects directly to the post-content moat. When content is infinite and AI-generated, the moat moves to verifiable doing — work that demonstrates irreducible judgment, not work that demonstrates prompt competence. Cognitive variance is the input that produces the non-modal doing worth verifying. The person who reasons from a different construct doesn’t generate the same project, the same architecture, the same strategic read. That’s the differentiation. And it maps onto the non-consumer strategy: the neurodivergent adult was already the non-consumer — roughly 80% never credentialed into the professional class, filtered out by interviews designed for modal social performance, performance reviews calibrated to neurotypical signaling. The non-consumer frame says: the people the market ignores are where the overlooked position forms. Here, the neurodivergent adult is also the cognitive-variance reservoir that the modal-homogenizing economy systematically undervalues. The same filtering mechanisms that excluded them from credentials excluded the economy from their variance. Building in the gap completes the circuit: the gap market — the space where dismissed, pathologized cognition lives — is where the defensible position forms. The dismissal is the moat, again. The market’s refusal to value the thing is what makes the thing available to the builder who recognizes it.

The honest caveat

Now the honest caveat, because this is where the argument can collapse into a trope I want no part of. Cognitive variance is not universally valuable. Some variance is noise — dysregulation that destroys output, executive dysfunction that can’t be routed around, cognitive profiles that produce friction without signal. Not every neurodivergent cognition is a strategic asset. The claim “neurodivergent = competitive advantage” slides rapidly into the savant/super-powers frame, and that frame is reductive, dehumanizing, and wrong — it flattens real people into a marketable caricature of useful difference while ignoring the daily costs and the majority who don’t fit the savant mold.

The thesis here is directional and population-level: AI homogenization makes cognitive variance more valuable at the margin, and neurodivergent cognition is a significant, underused source of that variance — not that every neurodivergent person is a moat, not that hiring autistic engineers is a strategy. The “AI homogenizes output” claim is well-supported by observable behavior; the “neurodivergent cognition = strategic variance” claim is reasoning-from-the-construct plus lived experience as a builder and a parent watching a kid whose cognition works differently get told repeatedly that different is deficient. It is not an RCT. I’m saying so plainly. The honest version of this argument is probabilistic: if modal output is being commoditized, and variance is the scarce input, and neurodivergent cognition is a large source of variance that the economy suppresses, then the institutions that figure out how to route that variance into usable judgment will have an edge over those that don’t. The edge is at the margin. The margin is where strategy lives.

Coda

In the age of average, the moat moves from the modal — which AI now produces cheaply and at scale — to the variant, which it can’t. Neurodivergent cognition is one of the largest reservoirs of cognitive variance available, and it is the one the economy has spent decades pathologizing, filtering out, and suppressing. The competitive move is to stop treating it as a cost-center accommodation — a line item in a DEI budget, a compliance exercise, a thing you do to avoid liability — and start treating it as the scarce strategic input it has become. But the reframe alone is insufficient. You must also build the workplace protocols that let the variance actually contribute: structured communication, async defaults, explicit norms, the protocol redesigns that convert social friction into productive signal. Because a masked neurodivergent employee is a modal-output employee you paid a premium for. You carried the accommodation cost, suppressed the asset, and got the same output the AI produces for a fraction of the price. That’s not a strategy. That’s a loss.

Reframe neurodiversity from DEI posture to competitive thesis. Build the protocols that make the asset usable. Capture the variance the modal-homogenizing economy can’t.