# New world as an ADHD person Six hours. I had planned to spend thirty minutes on a proof-of-concept. Engineers overrun estimates __all__ the time. But this wasn't a planning failure, it was something I'd spent my whole adult life recognizing too late: hyperfocus. The kind where time doesn't slow down, it simply stops existing. My coffee went cold. I missed lunch. I blew past my 6pm hard stop with more working code than I'd shipped in the previous three days combined. The twist: I was working with Claude Code. And I hadn't been writing much code at all. --- I was first told I had ADHD as a teenager. My mother, a pharmacist, disagreed; not with the diagnosis exactly, but with what the diagnosis implied: medication. So that chapter closed as quickly as it opened. I went on to do what a lot of undiagnosed ADHD people do: I built a career by outrunning my own brain. Long (inefficient) hours, hyperfocus weaponized as a coping mechanism, elaborate external systems to compensate for the executive dysfunction I didn't have a name for yet. In 2021, in my early thirties, I got the diagnosis again. Properly this time, with medication, with a therapist, with actual context for the previous three decades. It felt less like a revelation and more like finally seeing the subtitles to a film I'd been watching my whole life. Late diagnosis doesn't fix anything. But it does give you a map. --- So here I am in 2026, holding this map, and the territory has changed completely. AI coding agents arrived and immediately did something strange for my brain: they fit. Not perfectly. But better than almost any other tool I'd used. Think about how executive dysfunction actually works. The bridge between "I know what to do" and "I am doing it" is missing or unreliable. A structural gap in the signaling, not laziness. You can stare at a task for forty minutes, completely aware of every step required, completely unable to begin the first one. The friction of starting is *enormous*. These tools lowered that friction to almost nothing. "Hey, I want to explore this problem. Start here." You don't have to have the architecture worked out. You don't have to have the perfect name for the function yet. The act of *beginning* (which ADHD makes so hard) becomes nearly trivial. The conversation is the scaffold. The model fills in the gaps you'd normally spend executive resources on. You can start badly, vaguely, and iterate forward. This is not how most software development tools work. Most tools assume you've already started. --- But here's the part I didn't see coming. ADHD brains are famously bad at dopamine regulation. The reason tasks are hard to start is the same reason they're hard to stop: we don't have a reliable internal signal that says *enough*. The dopamine hit from making progress is chemically addictive for a brain that doesn't produce it reliably otherwise. Normally, the friction of the work itself acts as a natural brake: it gets hard, you pause, time passes, you come up for air.{{< sidenote side="alternate" >}}"air" here is also a wink at certain habits (smoking, vaping, etc.) that a lot of ADHD brains, mine included, get hooked into. Not the chemistry so much as the loop.{{< /sidenote >}} When the friction is gone, the brake is gone too. Steve Yegge wrote about this [in February](https://steve-yegge.medium.com/the-ai-vampire-eda6e4f07163) as a vampire dynamic: AI makes you 10x more productive (his number, not mine), companies capture all of that 10x, and the worker burns out at 10x speed. That's a real structural problem. But for ADHD people, the vampire was already inside the house. I don't need anyone to extract my productivity at scale. I'll do it to myself. I already do it to myself. The exhaustion isn't new, the acceleration is. What I need is a new way to answer a question my brain has always been bad at: *when is enough?* --- Here's what I've learned, imperfectly, in the past few years (feels like a decade). Hyperfocus is a superpower with an invoice attached. AI amplifies it in both directions. A task that would have taken hours of overhead to warm up now lets me get to the interesting part immediately, and the interesting part is exactly where my brain thrives. I have lost more evenings, more "just thirty minutes" that became five hours. The superpower doesn't come with a volume knob. You have to install one yourself. The invoice used to be paid differently. In my twenties I worked without a stop signal — long hours, contract gigs stacked onto the weekends, no life outside the work. Every two or three weeks the bill would arrive all at once: a Sunday I couldn't get out of bed. That doesn't work anymore. There's a kid now, and a 6pm hard cutoff, and the invoice can't be settled in one Sunday-shaped lump. Which means I have to see it building, sooner. Time blindness hasn't gone anywhere (my wife has a thing or two to say about this). Timers help. What actually works for me is externalizing time the same way I externalize memory: make it visible, make it loud, make it impossible to ignore. The agent tools won't do this for you. They have no concept of your dinner getting cold. Executive dysfunction loves these tools, and that relief is also a trap. When starting is easier, I start more things. More half-finished contexts, more "I'll just switch over to this for a second", more fragmented attention. One bottleneck removed reveals the one behind it. I'm faster at beginning; I'm not automatically better at finishing. None of this is a reason not to use these tools. It's a reason to use them with a clearer picture of your own brain. --- What the map tells me now (the one I waited until my thirties to find) is that my brain was never broken. It was mismatched: to school structures, to open-plan offices, to 9am standups, to the way most software development was expected to happen. I built compensating systems for all of it and called the effort "discipline." Some of those compensating systems are obsolete now. The friction I trained myself to push through is genuinely gone. The new question is: what does discipline mean when the friction is gone? I don't fully know yet. I suspect it means learning to stop in a world that no longer stops you. After thirty-something years of a brain that felt like it was fighting itself, having a problem that's about *too much* flow instead of *not enough* is a strange and mostly welcome inversion. I'll take it. But I'm also tired.