Point of view

The day after the demo

demo the day after
47 great minutes, then the drop

Every AI app builder on the market is genuinely brilliant for about 47 minutes. You type a sentence, a working app appears, you ship a link to a friend. It's a great 47 minutes. The trouble is what happens the next morning.

The market has quietly split into two stories. The first — the one every landing page tells — is idea to app, fast. That story is largely solved. Lovable, Bolt, Replit, v0 and a dozen others have made the first build almost free. In controlled tests they spin up a working prototype in under an hour.

The second story doesn't have a landing page yet. It's the day after: when the demo has to become a product that survives contact with reality. When you need to change something and discover you can't see what you'd be changing. When a fix spawns two breaks. When you realise the code is a black box and you are not its author — you're its customer.

When something breaks, the hardest part is no longer understanding what the code does — it's understanding why it exists at all.

Comprehension debt is the real debt

The industry even has a name for this: comprehension debt. It's been named in talks, in security write-ups, in the agencies that have sprung up to rebuild apps people generated and then couldn't maintain. The pain is well documented. What's missing is a product built to remove it.

And it's not a small problem. Security researchers have found that a large share of AI-generated code ships with known vulnerabilities, and a striking number of generated apps go live with their database access rules switched off entirely. That's not because the people building them are careless. It's because they can't see what they built. You can't review a black box.

The doom loop

Without a map of the system and a memory of past decisions, every change is made blind. You fix one thing; you can't tell what else you touched. The newer tools have started to iterate — plan modes, project memory — but iterating without legibility is still iterating blind. You're moving faster through fog.

This is the loop that grinds people down: build, break, patch, break again, lose the thread, start over. It's exhausting, and it ends with a product nobody fully understands — including the person whose name is on it.

So we built for the morning after

Doric's whole design starts from a different question. Not "how fast can we generate an app?" but "how do we keep it legible, safe to change, and yours — for years?" That reframes everything:

The bet: the first build is a commodity. The day after is where the value — and the lock-in, and the pain — actually lives. That's the lane we're building in, and as a complete mechanism it's still unoccupied.

Why now

Because the cost of the old way is becoming visible at scale. Platforms get sunset and orphan their users. Generated apps pile up as unmaintainable sprawl. The very speed that made the first build free is what makes the day after expensive. The faster the industry generates, the bigger this problem gets.

We think the next chapter of this market isn't about generating more. It's about keeping what you generate — understanding it, owning it, and being proud of it. That's the day after. That's Doric.


Next: Why our agents run on real git branches →

Build something you'll still own tomorrow.

Start building free