What have I been up to?
I realized it’s been a while since I wrote on my blog. In fact, it’s been exactly one year to the day, which feels like a pretty good excuse to finally write another post. The past year has been a little different from the ones before it. I’ve spent a lot of time maintaining and improving systems I’ve already built rather than starting huge new projects.
It’s been lots of tech stuff, but more in the vein of small things that build on the foundation I already have. I was also putting off writing more on WordPress because I want to move off of it. I was testing different LLMs in the cloud and on my local machine, too.
In retrospect, I’ve been doing the important, less-visible work of keeping things running, making them better, and figuring out what I want to tackle next.
Tech stuff
The small things building on the big things:
- Terraforming more infrastructure
- Managing and remediating database vulnerabilities
- Troubleshooting CI/CD issues
- Reducing costs
- Making performance tweaks
It all amounts to solidifying, automating, and securing what’s already in place. That’s a great place to be after years of huge projects. I’m excited to get back into some bigger tech projects, though.
Blog hosting decisions
Still experimenting and I looked at:
- Substack – too enmeshed with other things I’m doing personally, but could still be an option
- Medium – no dark mode and not appealing to me overall
- GitHub Pages – interesting, still an option because I like using HTML and CSS
- BearBlog – very interesting because it’s in the less-is-more mode I’ve been in; really bare-bones and basic interface
Mainly, I want something affordable but user-friendly for myself and those reading it.
AI experiments
I went off on a tangent here for a while because I didn’t feel great using LLMs like ChatGPT and Claude because of concerns about their carbon footprint. I set up Ollama and LM Studio to compare local models. I wound up not liking LM Studio because it would just not run any models at all, but Ollama was nice. With only 24 GB of RAM to work with locally, the models I could run paled in comparison to ChatGPT.
The local models were interesting, and I like the idea of running things locally, but the gap in capability was still pretty noticeable for what I wanted to do.
What’s next?
So that’s the past year in a nutshell: not exactly one giant project, but a lot of small things that add up. I’m looking forward to seeing what the next year brings, and hopefully writing about it before August 2027.