So last night at midnight I killed a pop-up social network app.

On purpose. I mean that was always the plan. Thirty days it existed, then gone. No archive, no "we're pivoting," no announcement about what comes next. Just a farewell page and a shutdown. mirage — my pop-up social network for the Coachella Valley festival season — appeared April 1 and disappeared May 1 exactly on schedule, which is more than I can say for most things I've shipped.

I want to tell you what it was and what it wasn't. But first I have to tell you about the last few months.

By the end of 2025, the tools had crossed a line.

The things I’d been telling workshop audiences were still “coming soon” were suddenly here: agents that could plan, write code, run tasks, connect systems, and make a weekend prototype feel like a small production team.

I remember sitting with my comp book at my kitchen table, watching the family putting up the Christmas tree lights thinking: I have been explaining these tools for two years. I should probably get back to building.

Which brought me to the uncomfortable part.

In 2025 I ran workshops. A lot of them. ERC. CSUSB Palm Desert. City Chambers. Leadership Council meetings. Sat Singh, making the case, showing up, genuinely believing that if I explained this well enough and often enough the community would take it as seriously as I was. And maybe the seeds of a startup community would take hold. Maybe a few seeds were planted.

But I didn't move the needle. Not the way I thought I would. People nodded. Some took notes. Most went back to what they were doing.

(I've been in enough of those rooms to know what the nod means.)

So I made a decision. 2026 was going to be the year of building. Not organizing. Not convening. Not running another session about what we might someday do with these tools. Just using them. Shipping things. Experimenting. Building in public and learning the only way you actually learn — by watching something you made not work the way you planned.

mirage was one of those things.

Not because the world needed another social app. It absolutely did not.

But because I had said 2026 was going to be the year of building, and this was one of the ideas sitting in the notebook waiting to be embarrassed in public.

So I built it.

A 30-day pop-up social network for Coachella Valley festival season. Appeared April 1. Disappeared May 1 at midnight — by design, from day one.

No accounts, no profiles, no algorithms, no data extraction. The trick was the geo-fence: if you were in the Valley, you could post. Anywhere else, you could only watch. A daily passphrase managed community. Polaroid photos, local tips, and a catchy soundtrack were the good vibes. A community bulletin board with a desert soul and an expiration date.

Sketched it out in my comp book on a Thursday. Had the bones built over the weekend. Published it within eleven days. Ran for thirty, broke in interesting ways, got picked up by KESQ News Channel 3, and disappeared exactly on schedule.

mirage cost less than $500. Bootstrapped. You can read the autopsy here.

Let’s go back a bit. I co-founded a company called Digital Family Network. An early virtual world, social gaming platform for kids and tweens.

Real tech studio — engineers in Sweden, gamification experts, designers, QA teams. Any time we had an idea for a new game inside the platform, here is what happened: a few weeks getting the concept aligned internally, a few months for engineering to build it, another few months in QA getting it ready for real kids to touch.

Six months on average from idea to published. Dozens of people. Across London, Stockholm, Malmo and Rancho Mirage.

I am not saying mirage was on par with what those engineers shipped in Stockholm. That would be a significant overstatement.

What I am saying is that with mirage, I went from idea to live product over a weekend. In eleven days it was published. In thirty days I had real users, real press, and a list of things that broke that I never would have discovered in a planning meeting. The wild told me things no studio ever would have — because the wild is the only place where you find out what your thing actually does when real people touch it.

(Turns out some of what it did was break. Also instructive.)

The point is not that AI makes everything work. The point is that AI collapsed the distance between "I have an idea" and "I can find out if the idea works."

Old math: six months and a team of dozens.

New math: a weekend, one person in Rancho Mirage, an army of agents, and a suspicious amount of confidence in a concept he had been sitting on for years.

That gap is the story. The gap, not the app.

Coachella Valley has been having the same conversation about economic diversification since approximately the invention of the golf cart. How do we attract the next Google? How do we become a hub? How do we position the region as a tech startup ecosystem?

I have sat in those rooms. I have nodded at those slides. I have never thought this was the way.

Lately the version I keep hearing is this one: the Coachella Valley could be a leader in the AI revolution.

Noted.

You might want to rethink that framing when you look at the competition. Austin has been building a tech ecosystem for thirty years. Boston has MIT and a biotech corridor and more venture capital than we have golf courses. Seattle. Salt Lake. Hudson Valley. And then there's Silicon Valley, which, famously, invented the thing we are proposing to lead.

We are not going to out-Silicon Valley Silicon Valley.

It’s not the right goal. The right goal is messier and more interesting and honestly more achievable. Not one big thing. Hundreds of small things. The restaurant owner whose tool has been on a napkin since 2021. The student whose class project could actually ship. The career switcher who's been living next to a problem for fifteen years and finally has the means to take a swing at it.

Each one taking a shot. Most failing. Some not. All of them experimenting and learning something that no workshop ever taught, including mine.

(I say this as a person who ran a lot of workshops.)

The tools exist. The cost has collapsed. The only question left is whether you do something with that or you don't.

I keep thinking: if even a fraction of the people in this valley who have been nodding at those slides for twenty years shipped one thing instead — one embarrassing first version of one idea — that might be what the startup ecosystem we keep talking about actually looks like. Not attracted here from somewhere else. Built here. By us. From the inside out.

Thirty years of conversations about economic diversification. Turns out the answer might have just gotten cheap enough to find out.

Sat Singh runs SunshineFM from Rancho Mirage. 2025 was spent in rooms explaining AI to people who nodded politely. 2026 is all about building things instead. Both approaches have their merits. Only one of them has a case study page.

If this resonates, pass it along to someone in the valley who'd benefit. This is a community project — it grows the same way communities do, one conversation at a time.