Bet on time

A page, a million dollars

In 2005, Alex Tew, a broke 21-year-old student, put the pixels of a web page up for sale. One dollar each, a million in total. The next year he was a millionaire. The Million Dollar Homepage still exists today, untouched, like a graveyard of logos. I stumbled on it by chance, not long after it launched. Twenty years later I still find it brilliant — not for the money, but for the cleanness of the gesture: an economic object that exhausts itself, once and for all, and can never be redone on the same surface.

The idea lodges somewhere. It waits.

The PHP draft

Back when I was learning PHP, among the first sites I cobbled together, there was one with a single function: the last visitor wrote the message. A single cell, overwritten on every visit. No time-slicing, no archive — just last-write-wins. It was dumb, it was mine, and it was the same intuition as Alex Tew’s, without my knowing it: a public space whose ownership is fleeting. Three visitors total, including me. I forgot.

The idea matures, in reverse

Over the years I kept coming back to it in my head, and the idea refined itself on its own. Not just any time: rare moments — eclipses, solstices, the last day of a decade. Not first-come-first-served: an auction. With curation, staging, a permanent archive. It had a name in my head long before it had a file: Kairos, the Greek word for the opportune moment.

When you daydream, refining is free. When something has to ship, I did the opposite — I simplified. One minute, one message, anonymous, first-come-first-served. Sttew. It’s the classic amateur-creator inversion: the refined version writes itself first in the head, the poor version writes itself first on the screen.

Bolt, Claude, Vickrey

Sttew was born during a Bolt.new hackathon. Bolt was handing out generous resources and near-unlimited model access, so I spent time on it. I tested it myself: the retro digital billboard, the minutes ticking by, the global counter incrementing with no one but me. But I never promoted the site — it didn’t feel mature. A few months later, I handed the repo to Claude Code. We brainstormed, it rearchitected toward what was running until yesterday: Cloudflare Workers, D1, a cron every minute to advance the queue. What was live when this letter went into writing was no longer my hackathon Sttew; it was a Sttew rebuilt by Claude.

And since we were already there, I came back to Kairos, the version that had always seemed more beautiful to me. I wanted an auction — I could picture the highest-bidder system clearly, but it had something vulgar about it. Claude is the one who suggested Vickrey. Sealed second-price auction: you bid what you’d actually be willing to pay, and you pay what the second-highest bidder put in. More elegant, fairer, quieter. Convex on the backend, a unit test for the second-price calculation in convex/lib/, an archive page for past moments. The object existed in my repo. It was never deployed to prod.

The verdict

In the same conversation where we’d cooked up Vickrey, I asked the question I’d been avoiding: is it worth pushing further? The answer confirmed what I already knew without wanting to hear it. An anonymous public billboard that sells itself by the minute — or by the second, or by the eclipse — is, above all, a magnet for spam and a moderation nightmare. With a captive audience, it might become something; without one, it stays an archive of Polish bots and crypto scams. The verdict took a few lines.

It killed the urge.

I archived Sttew. I archived Kairos. The same AI that had helped me materialize both objects had just buried them by confirming my gut — midwife and undertaker in the same conversation. It’s a strange break, and I don’t know whether I find it sad or clean. Probably both.

A hundred minutes, archived

A hundred minutes, in the spirit of Alex Tew’s page. Hover the filled squares.

P.S.

I’m not pulling a method out of this. The idea isn’t bad in absolute terms: Alex Tew had the audience of a 2005 internet that was more credulous, more curious, more willing to spend a dollar on a joke. Redoing that in 2026, with time as the raw material, would require either a captive audience or a tolerance for moderation that I don’t have. What stays with me is that a twenty-year admiration produced two objects, and the simple act of having built them was more interesting than what they might have become.

That’s also what living with a fixed idea looks like. At some point it takes form, at some point it reveals itself for what it is, and you learn to part with it on good terms.