The coming era of DIY software

The advancement of AI in the last couple of years, and particularly the acceleration of its coding/math capability in the last 9-12 months, are both amazing and scary. On one hand, I’m a bit concerned about my job. On the other hand, I’m excited about the fact that I now can do something that I was not able to do earlier: quickly implement software programs that serve just my needs.

As an example, I have recently finished wrapping Excalidraw as a desktop app using AI. I have been wanting a diagram tool that runs locally like this for a long time. It is nothing fancy; Excalidraw already provides an npm package that does all of the heavy lifting, and the desktop app is just a small wrapper around it. You might argue that this is a piece of cake for an experienced front-end developer. But the thing is, I’m not an experienced front-end developer! Without AI, I didn’t even know that it was possible for me to implement something like this if I spent perhaps a year learning JavaScript, let alone implement it within a couple of hours. Folks might say that this is AI slop, but that’s ok with me; I don’t really care. The app serves my need. If it also serves yours, then use it; if not, it’s cool, no big deal!

So this is what I think is coming: an era of DIY software! Not happy with the text editor you’re using? You can write a new one! Feel that MS Excel is too bloated? You can make your own lightweight version of it! It would be hard to sell your software or even convince other people to use it because i) they will think your software is AI slop; and ii) they have their own DIY versions too. But it’s perfectly fine if you are its only user.