Have you heard of the canary in the coal mine?
Miners took canaries underground because the birds react to toxic gases before people do, so a canary in distress meant it was time to get out.
I wanted to find something like that for software, a signal that shows whether software is solved.
Because I run ScreenshotOne, a popular screenshot API for developers. And according to social media, every developer and every company is now building their own with agents, replacing tools like mine.
But in my business, the results are mixed. I do have customers who build their own solutions, but I had those before agents too.
I also have customers who choose ScreenshotOne, and I often ask them why, in the support chat and whenever else we talk: why don’t you build this yourself? The answer is basically always the same. They don’t want to spend time on it. Competition is brutal now, and they want to focus on their own product and its core features.
It’s only what I see from inside one company. Before the day nobody needs my product, I want to know if we’re heading in that direction, or at least where we are now.
I searched, but I haven’t found a quantifiable answer. So I built one: IsSoftwareSolved.
By solved, I mean agents can take simple requirements and a conversation with humans, then create the software, deploy it, maintain it and fix its bugs and issues, almost autonomously. In some cases they’d even be proactive, deducing requirements from what people do and building tools on the go.
I’m not trying to predict whether software will be solved, and I’m not asking you to take this number on faith. You can check how the score works.
I’m tired of the narratives, the opinions and the vibes. I just want a real conversation about all this, and now we have a basis for one.
But even if we don’t talk, for me, researching the numbers and thinking all of this through was a way to build an intuition for where we’re going, and I got a lot of joy out of it. I still have mixed feelings. Sometimes I do see software being solved. Sometimes I don’t.
I hope you have fun with it and find something to argue about. If you don’t trust a number, think there’s a mistake, or have any ideas to improve the project, please open an issue on GitHub.