Something interesting is happening in the world of AI and product development. The best AI models can now build complex software with little intervention — and this isn’t me jumping on the hype train. As someone who has spent thirty years at the intersection of design and engineering, the gap between vision and build has always been a challenge.

I’m now seeing that gap closing. It’s closing because AI is creating a new abstraction layer. Hard problems are suddenly much easier. Multiple specialist roles can be handled by a single person. An end-to-end design and development cycle can happen exponentially faster.

The business world is responding by shrinking teams to accomplish the same goals cheaper and faster. Which makes sense if you think the constraint is still capacity. But capacity isn’t the constraint anymore. When building gets radically easier, the question shifts. It’s no longer can we build this? It’s what do we do with 20x more capacity?

That’s not an optimization problem. That’s an imagination problem.

The Bottleneck Was Never the Ideas

Designers, artists, writers, product thinkers, strategists — these are people who live in the space of what could be. They’re trained in divergent thinking. They’re comfortable with ambiguity. They gravitate toward problems that matter to actual humans — not because they’re more noble, but because that’s what the work selects for. You don’t spend a career in design because you want to optimize funnels.

The limitation was never imagination. It was translation. The distance between a vision and a working thing has always required passing through layers of handoffs, timelines, and compromises that compress — and sometimes destroy — what made the idea worth having in the first place.

That bottleneck is disappearing. Creative thinkers now have tools to build their ideas directly. To get something real enough to react to, iterate, and share with the world. The distance between I can see it and here, look at this has collapsed.

The creative class has long received the short end of the stick — but now holds the skills most needed. They are the ones with the vision to look at a step-change in capacity and ask what it should actually be used for.

When AI Empowers the Whole Team Together

Right now, AI tools are largely single-player. One person, one model, compounding their own expertise in ways that would have been unimaginable two years ago. But the shape of something bigger is emerging.

What happens when a cross-functional team — engineers, designers, strategists, domain experts — are all operating at AI-augmented speed, together? When the designer can build a working app before the kickoff meeting, the engineer can explore three architectures in the time it used to take to spec one, and the strategist can pressure-test assumptions against real market data in an afternoon?

I don’t think anyone has seen this at full expression yet. But the early signals are unlike anything any of us have seen. The velocity is different. The fidelity of collaboration is different. The gap between intent and outcome gets smaller in every direction at once.

Caveats and Responsible Use

I’m generally cynical about AI and struggle with its dual nature. I’ve left a long trail of critique and concern over the years, and none of that has softened. The concentration of power in a handful of corporations, the staggering energy costs, the way value extraction tends to outrun value creation — these are real problems.

But I also don’t think we need exponential improvement or anything approaching sentience for this to matter. What we have is an abstraction layer over our existing technology stack, and it’s already delivering real value. And the emerging shift toward efficient, on-device AI — capable models running locally on hardware you already own, without sending your data to a cloud — gives me real hope that this technology can serve people rather than just extract from them. I’ve been experimenting with this ideas and it’s getting close!

For many of my creative friends, AI has become a symbol of all that’s wrong and unfair in the world. I’m definitely not dismissive of this view and strive to do something about it. For me, the most powerful act of resistance is using it to empower artists, challenge the status quo, and raise awareness of its dark side. These tools carry real harm and real potential, and I’d rather the people with conscience and craft be the ones shaping what they become.

Start Here

The greatest problems we face — in health, education, human connection, sustainability — aren’t waiting for better algorithms. They’re waiting for people with the empathy to understand them and the imagination to envision something new.

Imagination is the most important skill in the world right now. Not prompt engineering. Not learning to code. The ability to see something that doesn’t exist yet and describe it clearly enough that it can become real.

I wrote a free, open-source guide for my friends and colleagues that have a lot to contribute but are technically blocked from jumping in. This is a primer and a Claude plugin that helps you from zero — never touched a terminal or wrote a line of code — to a high-performance engineering stack. It’s not a coding course, but it doesn’t hide the sausage-making either. Think of it as a map through unfamiliar territory, with an AI companion built in to handle what feels foreign so you can stay focused on what feels like your craft.

I made this video below with my daughter Oona who has never coded. It’s all the steps to get a live web app running on a public server and the foundation of a space that’s only limited by your imagination (and a little determination).

I’m excited to hear what you build!

Read the guide →