So I came to a depressing realization in May when I first played with Claude Code. I kind of hinted at it in a blog post. But it might be better to spell out. Again - no fun for me to point out.
If you can generate an app purely with an agent - it probably has no value. Anyone else can generate it. You’re really going to be forced to build some other hook - but the app alone won’t be it.
This does create sort of a blast radius a lot of small apps will get caught in.
This isn’t new. In the 80s there were a lot of apps you could sell that are worthless today. Same thing in the 90s. That line has been stable for a while but it does move and has probably moved again.
But that dream of “people can make apps without learning how to code?” Yeah. I mean you can make them. But building a business with them? Probably not a thing.
Like any time this has happened before there will be a very short window where people can push this stuff to the App Store before they end up drowning in the rest of the apps getting pushed.
Then because there’s no commercial value everyone moves on.
If your app gets caught in the blast radius - these LLMs are powered by gobbling up a whole bunch of source code and then they hire a ton of devs to make apps to be fed in. You didn’t get beat because you’re not smart. Someone made a clip art CD full of source code and then shipped it as a lossy database.
(If you’re a young’un you might need to look that one up.)
@colincornaby Building an app business is a bad bet but building an app for a business will still exist. If a local pizza shop wants an app, they'll probably still want someone to do it for them. And now it's actually viable for them since the cost will be so low.
I think that's an opportunity, finding where software would have been prohibitive previously and moving into those new spaces.
Colin Cornaby
Steve Tibbett