It Started With a Screenshot at Midnight

Started With a Screenshot

4:59

By the time dinner was ready, we had a publication....Around midnight, Adam texted me a screenshot. He'd been kicking around an idea for a quantum computing news site and wanted my opinion on which domain to buy. I saw the message while I was winding down — probably watching TV — and mentally filed it under "I'll deal with that tomorrow."

There was no business plan. No technical architecture. No product roadmap.

Just a screenshot of a domain name.

The next morning we had planned to work on an entirely different project. We were waiting on something before we could move forward, and around 11:30, with time to kill, Adam asked: "What do you think about that quantum domain?"

I looked at it again. "It's actually pretty good. We should register it."

Then I added something that, in hindsight, changed the trajectory of the afternoon.

"I can probably vibe-code that pretty quickly."

His response was immediate. "Let's do it."

Here's the part that's still a little surreal.

The very first prompt that launched the project wasn't some 10-page product specification or meticulously engineered prompt library. It wasn't in one of my custom GPTs. It wasn't even a particularly sophisticated ask.

It was regular ChatGPT.

The reason is simple. The night before, while I was watching Adam's text sit unread on my phone, I'd been experimenting with ChatGPT's cross-chat memory and realized it had become dramatically better than I remembered. Good enough that I wanted to see how far I could push it without giving it much direction.

So I typed exactly this:

"thedailyquantum.com is a domain we just got... we want to put an 'everything quantum news' blog there and we're gonna build in Base44... help develop the concept."

That was it. One paragraph. I copied the response into Base44. A few minutes later, I connected the newly purchased domain.

The site was live.

From there I spent some time hardening security, improving the editorial experience — replacing the article editor with a richer WYSIWYG that supports embedded videos, formatted content, hyperlinks — and then I built something that I think is actually the interesting part of this story. An autonomous research and publishing harness.

It continuously discovers quantum news. Researches it. Prepares articles. Then surfaces them to an editorial approval queue before anything goes live.

I didn't hand over the keys entirely. There's a human checkpoint built in — editorial judgment stays in the loop. But the system does the discovery, the research, and the drafting. I do the approval. Then it publishes and distributes. Research. Analysis. Editorial judgment. Publishing. Repeat.

Every day. Automatically. What began as a screenshot sent just after midnight became a fully functioning publication before dinner.

That's the part that keeps sticking with me — and not for the reason you might expect. Not because we built a website. Websites have been easy to build for years. What's different now is that we built an editorial system. The website is simply where the intelligence shows up. The real product is the flywheel behind it. And that flywheel — the autonomous research, the curation, the editorial queue, the publishing pipeline — would have taken a team, a budget, a sprint cycle, and probably a few months two or three years ago. We did it in an afternoon.

Two people. One domain registered same-day. One paragraph prompt. A few hours of focused building. By the time we stopped for the day, The Daily Quantum was live, indexed, and already running.

I've been writing lately about synthetic labor and what it means to instantiate a temporary workforce around a problem. About how the bottleneck in knowledge work has shifted from execution to clarity of intent. This afternoon was the live demonstration. Adam had the intent. I had the execution pattern. We assembled the capability around the objective and shipped.

What I keep coming back to is this: the barrier to entry for building a real media company — one with a custom data model, signal-versus-hype editorial scoring, an autonomous content pipeline, role-based access controls, newsletter infrastructure, and a live domain — is no longer money, or headcount, or time. It's clarity. If you can describe what you're building with precision, you can build it. The screenshot Adam sent me at midnight wasn't a business plan. But it turned out to be enough.

---

Rich Washburn is a technologist and strategist working at the intersection of AI, infrastructure, and capital. He is Managing Partner and Chief AI Officer at Eliakim Capital and CIO of Data Power Supply.