Execution has become cheap, ideas are more expensive than ever.

For most of my career the expensive part of any product was building it. You could have a great idea in a meeting on Monday and wait a month to find out if it was any good. The idea was the cheap part. Turning it into something real was the cost. That has flipped. Building is close to free now, and the scarce thing is knowing what to build. Execution has become cheap. Ideas are more expensive than ever.

I do not mean ideas as in a clever thought in the shower. I mean the whole chain that used to be slow: the decision about what matters, the judgment about what to make, the taste to know when it is right, and the strategy for where the thing goes next. AI collapsed the part in the middle, the making. What is left standing is the part that was always hardest, and now it is the whole game.

What used to take four weeks takes forty minutes

I am not speaking in the abstract. Work that used to take about four weeks now takes under an hour to get to something real enough to react to. A landing page, a working prototype, a first pass at a brand system, a data pull that used to sit in someone’s queue for days. I can have the idea and build a real version of it before lunch. Not a sketch. A thing you can click.

When the deciding and the building live in the same afternoon, the loop between a good call and the proof of it collapses. I wrote about that from the leadership side in why I still build, and about the people it favors in the most valuable designers now are the ones who build. The short version: the distance between an idea and a testable result used to be the most expensive thing in most companies, and AI erased it.

The roles are blurring, and not just in design

Here is what falls out of that. When making is cheap, the walls between the people who make start to come down. Designers are shipping code. Engineers are making real design calls. Product managers are doing both. And a growing number of people are just doing all of it, moving between research, design, and building without stopping to ask whose job it is.

This is the clearest trend in the field right now. Yahoo put it plainly: AI is blurring the line between product designer and engineer (opens in new tab), with most designers now taking on product and engineering work, and engineers and product managers pushing back into design. The design engineer is not a fad title. It is what happens when the tools stop forcing people into lanes.

When ideas can become software in minutes, the bottleneck is no longer execution. It is knowing what deserves to exist.

The Figma piece, and where they are right

Figma wrote a good piece on this, Design’s Influence Is Expanding, and Here’s Why That Feels Hard (opens in new tab). Their read is that every technology wave hands design more surface to shape and more responsibility for shaping it, and AI is no different. Their own research backs it up. Designers reported doing about 17.5 percent more work year over year, and hiring managers said the need for designers held steady or went up. When something gets cheaper to produce, people do not want less of it, they want more. Economists call that the Jevons paradox, and it is the real counter to the fear that AI means fewer of us.

Their tools are still catching up to their thinking, which is normal in a shift this fast. But the thinking is in the right place, and that is what matters here. It is a short read and worth your time.

The other thing they get right is that this feels uncomfortable, and they do not pretend otherwise. Learning a new tool, a new part of the stack, a new way of working, all of it feels awkward at first. I think the discomfort is the point. If it felt easy you would already know it, and the edge is in the thing you do not know yet. The people staying on top of this wave are not the ones who found it comfortable. They are the ones who kept paddling while it felt strange.

That is the whole idea behind how I run Surf Studios (opens in new tab). Stay on top of the wave, or get rolled by it.

So what do we actually do with this

If execution is cheap and ideas are expensive, the work that matters shifts to two things: having more and better ideas, and getting them into the world fast enough to learn from them. That sounds obvious. Most teams are still built for the old cost structure, where the making was the constraint and the thinking was assumed.

Two things I would do more of, starting now. First, have more conversations and share more ideas. The raw material of good work is a big pile of ideas you can react to, argue with, and throw away cheaply. That pile comes from talking to people, not from guarding your best thought until it feels safe. Second, put your head down and build. Experiment. The cost of trying something is near zero now, so the only real mistake is not trying enough things.

On my team I made both a habit. Everyone uses AI in their work, and everyone shows their work every week, finished or not. The weekly share is where the pile of ideas comes from.

The cost of trying something is near zero now. The only real mistake is not trying enough things.

The goal is always to save time

Underneath all of it is one goal that never changes: save time. Not to squeeze more hours out of people, to give hours back. I like to think of it as building time machines. When you automate the tedious, mundane, high-effort work, you are handing someone their afternoon back. The report that took a day now takes a minute.

That is the real prize, and it is why I point AI at the boring, heavy, repeatable work first. Let the machine do the tedious part. I wrote about what that looks like day to day in what an AI-first design process actually looks like. Clearing the tedious work is not about doing more tedious work faster. It is about spending the time you saved on the part only you can do, deciding where the thing should go.

Cheap execution does not make every idea worth building. In hardware, regulated work, or anything where a bad release is costly, the build is still the expensive part, so slow down there. And if a team cannot yet tell a good idea from a busy one, more building just makes more noise. Fix the judgment first.

Where this leaves the people who do the work

None of this makes people less valuable. It moves where the value sits. The market is not paying a premium for someone who can produce a screen anymore, because a screen is cheap. It pays for taste, judgment, systems thinking, and the ability to decide what is worth making at all. Those were always the expensive skills. AI just stripped away everything cheaper that used to hide them.

So if you are a designer, learn to build. If you are an engineer, start making the design and product calls you used to defer. If you run a team, stop organizing it around handoffs that AI just made instant. And whatever your title says, get comfortable being uncomfortable, because the tools will keep moving and the lanes will keep dissolving.

Execution is cheap now. That is not a threat, it is the most leverage any of us has ever had. The only question left is the expensive one, the one AI cannot answer for you: what is worth building? Spend your time there. That is the whole job now.