For the last seven years or so, my website has been working via a Hugo static server, building pages every time I publish something. That was working well and it had the desired effect of being in control of the content, the way it was published, the structure of the site, rather than relying on a third party content management system.

Three years after implementation, it was already getting a bit creaky, and I had to start compromising on some of the functionality to reduce pages. Build times were increasing, it was all just getting a bit sluggish and temperamental. The compromises weren’t big, and each one didn’t really matter, but they did start adding up to the point that I was sure mrschristine.com deserved yet another solution.

A helping hand

The trouble is, my handy round-the-house technical expert has been a bit too busy to get involved. I’ve been carrying on with Hugo, sitting through those long build times and knowing it was time for a change but unsure which direction to turn. Finally, I realised the answer was probably there all along: AI. I’ve seen the evidence online of people getting help with their code, maybe it was time for me to dip my toe in the water.

So, I turned to Claude. I have the same concerns as so many people about AI – environmental impact, jobs at risk, mistakes it makes, social issues it causes. But I can’t deny that when you ask it do a task like rework my website so it sits on a database instead of an aging old static site, it really does the business.

Good relations

I’ve been working through lots of different areas with Claude and the to do list has only gotten longer rather than shorter, but the main function of the website is there. It now sits on a Cloudflare D1 database and serves via some kind of magic that isn’t entirely clear to me. The database part of it is game-changing, though.

So much of my website is data driven, particularly on the reviews side. Artists, directors, actors, authors, genres, scores, tags. All of it was there in one form or another on Hugo but limited because that’s not how static sites are supposed to work. The relational part of it is what databases are for. And at long last I can bring film and book series back in as well. It’s very exciting.

Working together

Sometimes I have no idea what Claude is talking about and have to get it to “repeat with much less jargon” please. We’ve surfaced lots of bugs and issues that were there all along (duplicated urls, inconsistency of headers, directors with multiple spellings of their names). But the highs are worth the pain. New specific people pages are gorgeous, and the search actually functions now with added bells and whistles.

As I say, still much more to do, and I suspect there will be plenty more posts through the journey, but I wanted to update for now: if you see anything looking wonky, it should only be temporary. Claude and I are on it!