Putting the "Personal" in "Personal Software"

I've often viewed "Personal Software" to mean products that have been designed and developed by professionals and released to a target market of individuals that would purchase (or acquire the open-source version) and use said software for their own interests. Think TurboTax, GIMP, and DuoLingo.

In this age of AI, we're now seeing that software can be personal in a different way. Not with respect to who the target is, but rather with respect to what features it has, its user experience, and the assumptions it makes. The first definition from Oxford Languages calls this out exactly.

AI has now made it so that many without the specific skill or time required to see their ideas come to life, are able to do so. Earlier this week I released pai-telegram, an example of such a personal tool. It makes some assumptions about workflow, tooling, OS choice, software prerequisites, and more that do not fit a general population. That's OK. The tool was designed to help one particular person (myself) and isn't concerned with flexible configuration, widespread adoption, or even ease of use. The entire concern of the tool is that it does what I want in a way that I find acceptable. I released it as open-source not to become a trending popular project, but rather as an appendix to the blog post. A proof of work if you will.

Such a focused piece of software does not need to be concerned with reliability, resilience, data governance, GDPR requests, and oh so much more. It is laser focused on the small needs of the world's smallest user base. A single person.

Curious whether anyone had put a name to this concept, I asked my AI assistant to do some research. I'm typically pragmatic in that I believe most ideas are not unique, and that almost certainly someone else has thought of it first. Reinforcing this viewpoint, here's a post from 2020, nearly a decade ago, well before AI, that compares personal software to that of a home-cooked meal. That analogy really resonates with me. When I make carnitas on the smoker for the family, they are good. We enjoy them. They aren't designed for mass market appeal. They aren't even good enough to be served in any respectable restaurant. They are however, of a flavor profile and quality that is satisfying to us, the only user base that matters.

I'm going to continue to build small, purposeful, useful tools for my user base of one. I'll release some of them as open-source software to hopefully inspire others to create their own too. What software will you cook up in your personal kitchen?

🤖 AIL LEVELS: This content’s AI Influence Levels are AIL1 for the writing, and AIL4 for the images. AI Influence Level (AIL) framework