I think differently, and I think AI might help

I recently read a Fortune piece about Palantir CEO Alex Karp arguing that neurodivergent people could be well placed for the AI era. Karp has dyslexia himself, and his view is that as AI takes over more of the routine work, people who come at problems from a different angle will matter more.

I agree with him, and I should say up front that it’s a bit personal for me. I’ve never been diagnosed with anything, but I have a fair number of the signs of ADD, and a lot of this article made me nod along.

My mind tends to jump between ideas, sometimes making connections that seem obvious to me but are harder to explain to someone else. I can spend ages thinking about a problem from different angles and then struggle to turn it all into a proper document. Meetings can be difficult to follow, especially when my thoughts have already moved on to something else.

The way we measure ability at work has always been pretty narrow, and AI is starting to show it.

Workplaces have long rewarded people who follow the process, communicate the expected way, juggle lots of tasks and fit in with the environment. Those are fine skills. They’re just not the same ones that produce good ideas or crack hard problems. Plenty of people are great at complex technical work and still bomb a standard interview. Others spot patterns nobody else sees but are wiped out after a day of meetings. And some come up with a completely different solution and struggle to explain it the way people expect.

We tend to notice the difficulties first, and that’s where we’ve been getting it wrong.

The useful question is often the odd one

Software makes this easy to see. Say a team is working on a slow application. Everyone is digging into database queries and caching, trying to squeeze more out of the architecture. Then somebody asks why we’re loading this data at all. Maybe the answer isn’t a faster version of the work. Maybe the app shouldn’t be doing half of it.

Debugging is similar. One developer checks each component in order until they hit the bug. Another notices that two unrelated-looking errors seem connected and goes off chasing that instead. Both approaches work, but they tend to find different things.

A team with different thinking styles in it will probably catch things it would otherwise miss, which seems like reason enough to make room.

AI handles the mechanics, not the thinking

AI is already good at a lot of work that used to eat hours: generating code, explaining unfamiliar concepts, writing docs, summarizing, suggesting fixes, grinding through repetitive stuff. It gets things wrong sometimes, but it changes how much time that kind of work takes.

If I need a basic form with validation, I can have AI draft a decent starting point instead of writing every field and error message by hand. Whether that form is the right solution is a separate question. Do we need to collect this information at all? Are we adding friction? Is there something simpler? Are we solving the user’s problem, or just doing what the ticket said?

AI can help me explore those questions, but somebody still has to know which ones are worth asking and judge the answers.

That’s where thinking for yourself counts: challenging assumptions, linking things that don’t look related, noticing odd patterns, imagining another way. Those skills get more useful once a conventional first draft is no longer the hard part.

So I don’t think the future belongs only to the people who work fastest with AI. It also belongs to the people who can tell when the obvious answer is the wrong one.

The other part of the article that stuck with me comes from psychologist Daniel Wendler, who points out that many workplaces were built around neurotypical ways of working, and that AI gives us a chance to rethink some of that. This matters as much as the creativity argument.

If you have great technical ideas but struggle to turn them into a structured document, you can give AI your rough notes and get something readable back, with the ideas intact. A big project that’s hard to break into tasks can start with a first plan from AI. A message you have trouble getting across in writing can be tidied up before sending. And if you lose the thread in meetings, you can read a transcript or summary afterwards.

These are small things, but small things add up. If you spend a big chunk of your energy on the mechanics of work instead of the work itself, taking some of that away makes a huge difference.

I can’t say for sure what’s going on, but I know that tools which take some of that friction away make a real difference to how I work.

That’s where AI has been genuinely useful to me. Not because it does my thinking for me, but because it helps me organize what I’ve already been thinking. I can throw a messy collection of thoughts at it and get something I can actually work with. It gets me from having an idea to actually doing something with it.

And these tools help everyone, not just neurodivergent people, which usually means the old process wasn’t great to begin with.

Better tools won’t fix a rigid workplace

AI won’t make workplaces inclusive by itself. A company can hand everyone the best tools available and still expect them all to communicate the same way, work in the same setting and prove themselves through the same rigid hiring process, then wonder why some talented people don’t do well.

Hiring shows it clearly. We expect candidates to think fast, keep eye contact, explain themselves well and perform under pressure in an unfamiliar room. How much does that tell us about whether they can design software or build something useful?

For some jobs, handling pressure really is essential. For plenty of others it has very little to do with the work. A practical task, a conversation about a real technical problem, or a chance to show how someone approaches things would probably tell you more.

That isn’t lowering the bar, it’s measuring the right thing.

It’s the same after someone is hired. Not everyone does their best work in an open office full of interruptions, and not everyone has their best ideas in a meeting. Some people need long stretches alone to think, others need to talk it through. A good team can handle both without treating one style as the default.

Different doesn’t have to mean exceptional

I wouldn’t want anyone to take from Karp’s argument that neurodivergent people need some special talent to earn a place at work.

They don’t.

Not every autistic person is a genius programmer, and not everyone with ADHD is unusually creative. A diagnosis doesn’t settle what someone is good at, and expecting people to be exceptional is just another unrealistic standard.

I don’t know whether AI will give neurodivergent people an advantage in the long run. But I do think it’s creating an opportunity to change what we value in people, and that’s a good thing.

For a long time, being good at your job has also meant being good at the way work is organized. Those aren’t always the same thing.

If AI can help separate the two, maybe more people will finally get to show what they’re actually capable of.