I think it’s a question of time until the mobile OS do this natively as a part of responding to user request. Like those Siri app snippets (scores, maps, stocks …), but with whatever adhoc UI that’s needed to best present the result (or allow the user to further interact).

“What to do in Zagreb with a family?” will build a personalized map & content app tailored to everything the AI knows about your preferences.

AI services will want to own this, but the current OS makers have all the incentives to not let them, so I expected OpenAI hardware to be this - a smartphone with launcher basically replaced by a prompt.

Until then, you might sell Bitrig to them as a way to wrestle this approach onto competing platforms, wrapped inside their app?

As an indie app maker (Weathergraph - a beautiful hourly forecast visualization), I should probably include the Ralph “I’m in danger” gif :)

Can you imagine that question being asked thousands of times per second around the globe vs a well built app?

Also, how many years do you think it will take for AI to reach that level of creating a bugless app that won't need a real developer to look at not even a single time?

Will users stay around for hours for the app to be built and deployed after they ask their question?