A while ago I came up with an idea for a personal project that could be useful. Since I love reading, I’m always on the lookout for new books to discover. The idea, in this case, is to create stacks of books grouped around a common theme. The theme can be anything readers can think of, like stories that bend time to group books about time jumps, books with narrators you can’t trust, or books where an alien comes to Earth and passes as human. The possibilities are endless, and as I say on the page describing Drayfus, the idea is to move away from the classic way of discovering books based on a genre or a star rating, which usually doesn’t say much.

The project itself isn’t complicated, but it’s true that if it hadn’t been for AI (Claude, in this case), I wouldn’t have been able to bring it to life. With my software development knowledge, it was fairly straightforward to create an architecture that lets me do what I want, test the code to have some peace of mind, set up a simple deployment system with GitHub, and automate the process. But that takes time, and getting all of this up and running was much faster than what I could have built on my own, considering that I have to work 8 hours a day, take care of my family, and handle the other obligations many of us have.

On top of that, Claude helped me a lot with the parts I wasn’t so sure about, like making the site SEO friendly from the start, and helping me with the frontend and UX (thinking about mobile too). The result could surely be better, but I think it turned out pretty well.

But it doesn’t matter. Once the project was launched, fully functional, and with a few stacks created in English and Spanish to start gaining traction, I ran into reality: nobody is visiting the platform.

A software product is more than its code

This, which seems so obvious, is something we software engineers often ignore. The technical side is just one of the parts that make up a product, and sometimes it isn’t even the most important one. In this case, I haven’t created a complex piece of software that has value in itself. I haven’t written patentable code worth money. I’ve built a platform that anyone could replicate in a few hours with a good AI assistant. What really matters is the execution of the idea, and whether, in the end, that idea has value for someone.

For now, I have no idea whether the idea is any good, because nobody is using Drayfus. And I know this because I check in periodically to see if I have any visits, only to find that the only visitor is me. But I’m aware that, even with no visits, this doesn’t mean nobody is interested in the concept. On Reddit there are plenty of subreddits dedicated to this kind of recommendation. They’re wonderful, and they help you discover books, but they tend to be ephemeral, and people end up asking for the same recommendations over and over. So I think that, even if it’s marginal, there’s some value here.

The thing is, the internet is huge, and getting your product noticed isn’t easy.

The product profile matters more and more

This project has led me to confirm what I already mentioned in other blog posts. It’s not certain that AIs will replace programmers, but it is clear that they’re going to replace the people who don’t know how to build product.

Small, multidisciplinary teams are going to compete with much larger companies, because they’ll be able to do far more with their knowledge. The bottleneck used to be the hours you could put into your work, but now that bottleneck is much wider, and what matters is knowing what to do with those hours. Marketing, SEO (which is also changing a lot with AI), and the other disciplines surrounding software are still important. And as a software engineer, you’re going to have to get to know them in order to be part of multidisciplinary teams where every member is a one-person band.

For now, my job is to learn how to gain traction so that someone discovers Drayfus and to validate whether the project makes sense. That’s where I am. What I’m clear about is that AI will help and guide me through the steps, but I’m going to have to keep learning a whole lot.

I’ll keep you posted.