Idea

This blog is hosted on a platform I built that allows me to quickly spin up a new blog. I have 4 that I'm trying to maintain: this one, a gaming one, a personal one, and one about cars. That's a lot of content that I need to generate, and I've been thinking about this for years now but was always offput by how much content that is. What I really want is to be able to just have a "system" that has read access to details of my life like my github, my social media, and my camera roll in my phone and can use that to generate appropriate content. Building something like that used to be a demotivating amount of work for me, but with AI I think it's much more attainable, so I'm considering how to build a content generation engine.

Process

My typical engineering process for feature implementation is broken down over a few steps:

1. Identify the problem you're looking to solve.

2. Create a requirements document around what the important parts of a solution to that problem look like, but don't choose technology yet.

3. Create an architecture for the feature that you're looking to build, suggest possible technologies and approaches.

4. Create a defined plan to test and implement the feature.

5. Implement/test the feature.

6. Push to production.

7. Smoke check production.

8. Gather feedback from production and iterate.

Caveats

Some of those steps collapse if a feature is small or the implementation is obvious, but caveat to the caveat: just because something is "obvious" doesn't mean the obvious solution is the best solution and you should usually take the time to evaluate. It is a judgment call though, because you are always balancing prioritization of your time.

The process I listed above is more of an overview, with a lot of details and considerations glossed over. As I work through this content generation engine feature I'll break down the details in future posts. Keep watching!