Neurodiversity at work: the talent case beyond accommodation

In a sprint planning meeting, an engineer keeps her camera on and her face arranged — nodding at the right moments, laughing a beat after everyone else, tracking the social rhythm of the call as carefully as she tracks a failing test. None of that is the work. The work is the architecture problem on the board, which she solved in her head four minutes ago and is now waiting to raise without seeming to interrupt. By the end of the day she is exhausted — not from the engineering, but from the performance of looking like everyone else while doing it.

That performance has a name — masking — and it is one of the quiet, expensive facts of neurodivergent working life. A great deal of what we call “soft skills” in an office is really a set of neurotypical defaults: read the room this way, hold eye contact this long, catch this verbal instruction on the first pass. People whose brains run on different settings can usually meet those defaults — they just pay a tax to do it, and that tax is invisible on every dashboard a company keeps.

For datarekha readers this is not an abstract diversity topic. Technical fields — software, data, research, anything that rewards intense pattern-matching and long stretches of uninterrupted attention — appear to draw a higher-than-average share of neurodivergent people, which means some of your most valuable colleagues are likely running this performance daily. The question is whether we are getting the most out of those minds, or merely the most out of their ability to hide.

What “neurodiversity” actually covers

Neurodiversity is the plain idea that human brains vary — in attention, perception, memory, and information processing — the way human bodies vary, and that much of this variation is natural rather than broken. The term gathers up several recognized profiles: autism, ADHD, dyslexia and dyscalculia, dyspraxia, Tourette’s, and others, often overlapping in one person. The point is not to flatten those differences but to reframe them: less “a deficit to correct,” more “a different operating system with its own strengths and friction points.”

On prevalence, honesty matters more than a tidy figure. You will see neurodiversity described as covering somewhere around 15 to 20 percent of the population, and that is the range I will use — but treat it as an estimate, not a measurement. The number moves a lot with which conditions you include and whether you count formal diagnoses or self-identification; diagnosis is also uneven across gender, age, and country, so many adults are neurodivergent without a label ever attached. The defensible claim is modest: this is common, not rare — a meaningful slice of any large team, and quite possibly a larger slice of a technical one.

The talent case, stated plainly

Move past accommodation for a moment and look at what these minds are good at. Start with pattern recognition. A lot of autistic and ADHD cognition is unusually good at noticing the thing that is slightly off — the inconsistency in a dataset, the edge case the spec forgot, a model’s accuracy “lying” because the classes are imbalanced. In data and quality work that is not a quirk; it is the core skill, the attention that fixates on detail rather than gliding over it.

Then deep focus. The same trait that makes open-plan offices and notification storms genuinely painful for many neurodivergent people is the one that, given the right conditions, produces flow states most workers rarely touch. The ability to disappear into a hard problem for hours — hyperfocus in ADHD, sustained special-interest attention in autism — is precisely the resource complex engineering and research devour. For many, deep work is the native mode, sabotaged mainly by an environment built to interrupt it.

And novel problem-solving. Minds that do not default to the conventional path will sometimes find the unconventional one — the approach nobody proposed because everyone “knew” it was not how things are done. Dyslexic thinkers, for instance, are frequently strong at big-picture reasoning and spatial synthesis even while they wrestle with the mechanics of text. In a field that prizes new approaches over correct execution of old ones, divergence from the default is not a bug to be smoothed out. It is, sometimes, the point.

None of this claims every neurodivergent person is a savant — that stereotype does its own damage. The narrower, sturdier claim is that the talent distribution on such a team differs from the neurotypical one, and a workplace built only for the default leaves a valuable part of it on the floor.

The disclosure dilemma

Here is where the talent case collides with reality. To design work around someone’s strengths and friction points you generally have to know they exist — and most neurodivergent people will not tell you. Borrow a hard number from the broader mental-health data, where the dynamic is identical: across recent workplace research, only about 13 percent of employees said they had told a manager their mental health was suffering, while between 42 and 46 percent feared that disclosing a struggle would damage their career. When that many people are convinced honesty will cost them a promotion or the benefit of the doubt in a review, the overwhelming majority stay silent and cope in private.

The neurodivergent version of this is masking, and it follows the same logic. If telling people you are autistic or have ADHD risks being read as “high-maintenance,” “not a culture fit,” or “not leadership material,” the rational move is to hide it — script the small talk, suppress the stim, and never request the thing that would help. The cost lands twice: the individual burns scarce energy on camouflage instead of the job, and the organization pays for talent it accidentally configured itself to suppress, never learning why its sharpest detail-finder quietly burned out and left. This is why “accommodations” is the wrong word to lead with. It implies a known need and a willing discloser, but the data says disclosure is the exception and fear is the norm — so a policy that only kicks in once someone formally identifies themselves will, by construction, miss most of the people it was meant for.

Psychological safety is the precondition

If silence is the default, the first real intervention is not a process but a climate — one where saying “I work better this way” does not feel like handing your manager a reason to doubt you. Psychological safety is the shared sense that you can speak up, admit a limit, or ask for what you need without being punished for it, and it sits underneath every other item on this list: none of the design changes below matter if people are too afraid to use them. You can publish a flexible-hours policy and a quiet-room booking system, but if the lived culture rewards the person who powers through the open-plan chaos and side-eyes the one who asks for a door, your policy is decoration.

You can read a team’s safety in small tells: whether “I didn’t follow that, can you write it down” gets treated as reasonable or as slow, whether the weird-sounding approach gets heard out or talked over. You cannot order people to disclose, and you should not try — you can only make the environment safe enough that fewer feel they have to hide, then make sure the work itself does not require hiding in the first place.

Design for everyone, not case by case

Which brings us to the move that actually changes things: stop treating this as a series of individual exceptions and start treating it as design. The disability world calls this universal design — building for the full range of users up front, so far fewer people ever need a special arrangement. A curb cut was meant for wheelchairs; it turned out to help anyone with a stroller or a bad knee. Good neuroinclusive design works the same way: aimed at neurodivergent friction, useful to nearly everyone.

Take the four most useful changes in turn. Clear written communication is the big one: a one-paragraph brief stating the goal, the constraints, and the definition of done removes the guesswork that verbal-only instructions force on anyone who processes language differently — and it also rescues the remote worker in another timezone and the new hire who never absorbed the unwritten context. Writing decisions down is good engineering hygiene that happens to help neurodivergent people most.

Protected focus time is the second. A culture that treats instant availability as the measure of a good colleague taxes exactly the people whose best work requires uninterrupted depth. Hours where messages can wait, batched rather than scattered meetings, and a “do not disturb” block treated as legitimate rather than antisocial lift the whole team’s deep work — and relieve the sharper pain that constant context-switching inflicts on many neurodivergent minds.

Third, vary the interview. The standard hiring loop is largely a test of fluent real-time social performance under pressure, which can screen out a brilliant engineer for being bad at the one thing the job will rarely require. A choice of formats, questions shared in advance, a take-home instead of pure live whiteboarding, judging the work rather than the eye contact: all of it widens the funnel without lowering the bar. You still measure whether someone can do the job — just no longer whether they can mask while doing it.

Fourth, give people environmental control — quiet space, the option to dim the sensory load, permission to put headphones on and disappear into the work. Notice the pattern across all four: not one requires anyone to announce a diagnosis. The benefit arrives without the disclosure, which is exactly how it reaches the masking majority the accommodation model never reaches.

The honest version of the takeaway

It would be neat to end by promising neuroinclusion is pure upside, but pretending so only sets up the backlash. Real differences sometimes create real coordination problems, and communication styles can clash in ways that take patience to bridge. Inclusion is work, not a slogan.

Still, the version of the argument I trust is narrow and hard to wave away. A large, talented share of your colleagues — the rough 15-to-20-percent estimate, very likely more on a technical team — are wired in ways your workplace was not designed for, and most will never tell you, because the same fear that holds disclosure down to roughly 13 percent keeps them masking instead. You can leave that talent to spend its energy on camouflage, or you can change the work itself — write things down, protect focus, vary the interview, lower the sensory tax — in ways that cost little, require no confession, and make the job better for everyone else too. The accommodation framing asks who needs a special exception. The better question is why the default was built so narrowly, and what it quietly costs you to keep it that way.