Adapted from my App.js 2026 talk in Kraków. Some things were said in front of a microphone that I wanted to write down too, so the rest of the room — the one this talk wasn't given in — can have it.
I want to start with something personal. I think it's important.
Late September, 2022. Tucson, Arizona.
A nursing home, near the end of a twenty-year fight with Parkinson's.
My dad.
By that point, the disease had taken almost everything. He couldn't speak. His hands had clenched shut. He couldn't leave his bed. The physically imposing man I'd known my whole life had been narrowed — physically, irreversibly — into one room and one bed and a body that no longer answered when he asked.
I'd come and sit with him. I'd feed him. I'd hold his hand. I'd stroke his head. And he would look at me with what I can only describe as pleading eyes. He had things to say. He had things he needed. There was no channel left for any of it to come out.
And here's the part that's hard to admit.
I had spent my whole career learning how to build things. How to ship products. How to solve real problems for real people. And sitting next to my dying father, none of that mattered. The most loving thing I knew how to do was hold his hand and try to read his eyes.
I felt completely useless.
A day and a half of React
Eventually — and it took longer than I'd like to admit — my brain worked past the shock and started doing what brains like ours do. It started solutioning.
There was a problem in front of me. I had tools. What could I actually build, right now, in this hospital room, to give him a voice again?
I sat down next to him and opened a new React project.
A day and a half later, I had something I called Speak Easy. It was, effectively, the world's slowest human-operated LLM. Letters scrolled across the screen one at a time. He'd hit the space bar to pick one. Then only the letters that could plausibly come next would scroll past. Over time it learned his most common requests.
Slowly, one keystroke at a time, my dad got words back.
He wasn't locked in anymore. We had a few more days of him.
After he passed, I had time to sit with what had happened. The thing I keep coming back to is this: the open-source community, and specifically React, gave me the ability to change the trajectory of someone's last days. I didn't need a team. I didn't need a runway. I needed an evening and a framework.
That's the thing I want you to hold in your head for the rest of this piece. Because I think it's the same thing happening right now — just at a much bigger scale.
The real story underneath the noise
The industry feels weird right now. The discourse is loud. Every week there's another model, another benchmark, another thread about who's about to disappear.
I get it. The pace is absurd. A year ago, most people were still debating whether AI coding assistants were even useful. Today, one person can prototype an app in a weekend that would have taken a small team months just a few years ago. That changes things. It changes the math of who can build, what they can build, and how fast.
I think most of the discourse is focused on the wrong parts.
The dominant story right now is that developers are disappearing. Junior devs are doomed. Seniors become AI managers. I don't buy it.
The real story isn't disappearance. The real story is leverage.
Leverage. Velocity. Momentum.
Leverage. That's the word.
Tools give people leverage. And right now, we are watching millions of developers suddenly gain momentum. That is what's actually happening. Not erasure. Acceleration.
Here's how I think about it.
Tools, by themselves, give you leverage — you can move something heavier than your arms alone could move.
Good tools give you velocity — you start moving fast.
Great tools — and I mean React, Expo, WebGL, React Native Skia, TypeGPU, this whole stack we work in — give you momentum. You can't be stopped. The work compounds on itself.
That's where we are. Not at leverage anymore. Not even at velocity. Momentum.
The gap that used to kill ideas
To understand why momentum matters, you have to remember what software has always been.
Historically, software has had a huge gap between imagination and execution. People had ideas. People had vision. People had real insight into problems worth solving. But turning an idea into a real product required enormous coordination — designers, backend, frontend, DevOps, mobile, infrastructure, release cycles, QA, funding, and time.
Most ideas died somewhere in the middle. They never made it across.
AI shrinks the distance between "I want to make this" and "this exists."
That sentence sounds simple. The implications can't be overstated.
Because once you collapse that distance, the bottleneck of software isn't engineering capacity anymore. It's something else.
What gets cheaper. What gets more valuable.
AI doesn't eliminate engineers. It changes what engineers spend time doing.
The mechanical parts of the job get cheaper. Boilerplate. Wiring config. Rebuilding the same CRUD app for the fiftieth time. Translating designs into JSX. All of that gets cheaper, and fast.
Look at the curve.
The distance between an idea and a shipped product used to be measured in quarters. In 2023, eighteen months from kickoff to launch was a perfectly reasonable timeline. Year over year, that distance has been shrinking. By the end of 2025, it started to trend toward weeks. Today, a motivated person with the right tools can take an idea on a Friday night and have something real in the world by Sunday.
And as the mechanical parts get cheaper, the human parts get more valuable.
Judgment becomes more valuable. Taste becomes more valuable. Leadership. Communication. Product intuition. The stuff that doesn't show up in a benchmark.
That's where the value is now, and it becomes more important every month.
Who thrives
So who thrives in this era?
Not the people who memorize the most syntax.
The developers who thrive are the ones who can understand systems, direct tools effectively, identify meaningful problems, make good decisions under ambiguity, and turn scattered potential into coherent products.
Read that list again.
That is engineering. That has always been engineering. We just used to also have to do a lot of typing.
A more creative industry, not a less creative one
If the amount of friction required to build things goes down, what fills that space?
Ambition.
This is becoming a more creative industry, not a less creative one. For years, a huge percentage of your time as an engineer was spent fighting tooling, wiring infrastructure together, writing boilerplate, debugging configuration. A lot of that friction is disappearing. What fills the vacuum is ambition. Experimentation. Smaller teams attempting things that used to require companies.
That's the trade we're making. I think it's a good one.
The audience for software just got bigger
We've talked a lot about speed. The part of this I care about most is something else.
It creates room for more people to build.
My north star is helping others bring their ideas to life. There are brilliant people everywhere with ideas they never had the resources to pursue. There are a lot of problems in the world that need solving.
Now a designer can build. A musician can prototype an app. A teacher can automate workflows. A teenager can ship software globally from a bedroom.
That's extraordinary. That's the actual story underneath the noise.
The audience for software creation just got bigger by an order of magnitude.
So
The industry feels weird, yes. The discourse is loud, yes.
But underneath the noise, the actual story is that the constraints on building have collapsed. You are unleashed. The ceiling is getting higher.
One person with strong judgment, strong taste, and the right tools can now create things that used to require entire organizations.
Think about that for a second.
It means builders are everywhere.
And it means it's a hell of a time to be one.
Go build.
Thank you to Software Mansion and Expo for App.js 2026, and to everyone who came up to me after the talk to tell me about the things they're building. The post-talk hallway was the best part. — Seth
