The second launch

Last weekend, I had dinner at the home of my university professor. He has spent decades in design in China, the kind of figure whose name carries weight in the whole field. I asked him the question everyone is asking right now: is AI replacing design?

He said yes. The execution, the production, even a lot of the concepts. AI can do it, and sometimes the options it generates are stronger than what a room of people would come up with. He wasn't defensive about it at all.

Then he told me about a branding project he did.

His team had redesigned a brand identity for a university. The design broke from tradition, it wasn't the usual round emblem everyone expected. The work was finished. Approved internally with the school committee. Done. And then the school announced it, the public saw it and almost everyone said no.

What happened next is the part I can't stop thinking about. He didn't go back and redesign a rounder logo. He went out and taught. He gave lectures. He sat through committee meetings. His team published a series of articles explaining the thinking behind the design, one piece at a time. They even held exhibitions around it. Months of work (none of it design work) until people could finally see what he saw.

The logo launched twice. Once as a file. Once as a belief.

AI collapsed the cost of the first launch to nearly zero. It never touched the second one.

The first launch didn't launch anything

I've felt this in my own work this year, and I bet you've felt it in yours, whatever you build.

Making things is no longer the hard part. I can go from an idea to a finished design in a fraction of the time it used to take. Founders can go from nothing to a live product in a weekend. The website goes up, the product ships, and then nothing. It's public, and it's invisible at the same time.

Because putting something into the world is not the same as putting it into people's heads.

The blocker used to be creativity, production, execution. That's where all the time went, so that's where we thought the value was. Now that part is fast, and what's left exposed is the part that was always underneath it: getting anyone to see it, understand it, and believe in it enough to act.

Here's the thing about that second launch, it doesn't compress. AI made production instant, but belief still moves at the same speed it always has: one mind at a time, through repeated exposure. That's why my professor needed a series of articles, not one announcement. That's why a launch tweet does nothing and a founder who explains their thinking every week for six months builds something real. You can buy traffic in an afternoon. You cannot buy belief. It accrues.

So the scarce resource quietly changed. It's not production hours anymore. It's belief-hours — the slow, unautomatable work of moving an idea from your head into someone else's.

AI cannot say no

This week, in a client meeting, I spent the better part of an hour refusing to do something.

The client kept wanting to add colors to the app, especially background colors. More color, more backgrounds, more everything. And I kept saying no. Not a flat no. I walked through why it doesn't work, why world-class apps don't do it, what thinking sits underneath the restraint they were trying to paint over. It was slow. It was a little uncomfortable. It was the most valuable hour of the engagement.

Here's what struck me afterward: AI would never have that hour.

Ask AI for a rainbow background and it will give you a beautiful rainbow background. Ask again and it will give you ten more. It will not argue. It will not hold a position under pressure. It will not risk the relationship to protect the outcome. Its defining feature is that it always says yes, and that is exactly what disqualifies it from the work that matters most now.

Because think about what the second launch is actually made of. My professor standing in front of a committee that didn't agree with him. Me pushing back on a paying client. A founder holding their positioning steady while everyone tells them to pivot. The second launch is built out of held positions, beliefs defended, explained, and repeated until they transfer.

And a held position only transfers because it costs the holder something. My professor put his reputation on the line when the public rejected that logo. I risk the client's annoyance every time I refuse. That cost is precisely what makes the conviction credible. AI's opinion is free, which is why it's weightless. Yes is now free everywhere. Everything valuable is a no.

But notice, the no isn't the point. The no is just the door. What my professor was really doing in all those lectures, what I was really doing in that meeting, wasn't refusing. It was teaching. Transferring a way of seeing. The refusal only works because there's a belief behind it worth transferring, and the willingness to spend the time transferring it.

Traffic is the wrong question

If you've launched something this year and watched it sit there, your instinct is probably that you have a traffic problem. Get more eyes, run more ads, post more.

I'd push on that. The logo didn't need more people to see it. Plenty of people saw it and rejected it. What he needed was for people to see what the design committee saw. The articles didn't advertise the logo; they taught people how to look at it. An ad can't do that. Only teaching can.

So the question after your first launch isn't "how do I get traffic." It's: who needs to be taught, and what can't they see yet?

That reframe changes what you do next. You stop optimizing the artifact, it's finished, the first launch is over, and start doing the slower work: explaining your thinking in public, one piece at a time. Holding your positions when people push back. Spending your belief-hours.

The first launch is done in an afternoon now. Everyone's is. Which means it decides almost nothing.

Everything you actually want — the customers, the trust, the outcome — lives on the other side of the second launch. AI will carry you to the doorstep of it faster than ever before.

Then it hands you the file, says yes one more time, and leaves.

The second launch is yours.

Studio Salt

I run Studio Salt, a fractional design partner that serves early stage startups.

Share my newsletter

Loving my content so far? I’d appreciate if you can share my newsletter to a friend 🙂