Back to Blogging: AI is Moving Fast

It has been a while. Last time I posted here I was still finishing up at deCODE. Since then I moved back to Iceland, changed jobs, changed jobs again, and the blog sort of stopped. Not a conscious decision, more like one of those things where each week you think “I should write something” and then you don’t, and eventually enough weeks pass that the whole thing feels like it needs a restart.

So here’s the restart.

Why now?

AI development is moving fast enough that I have something to say. I work with data and statistics for a living, so my angle is specific: what does this mean for quantitative work? The gap between the hype and what I see when I use these tools for real analysis is worth writing about.

I have these conversations at work all the time. Might as well write them down.

What to expect

I’m not turning this into a news feed. I’d rather write less often and say something. Work through an idea, share something I built, or write up something that took me way longer to figure out than it should have.

Topics I’m thinking about:

• LLM internals: how these models work, the stuff behind the hype
• AI in science: what tools like AlphaFold mean for fields like genetics, and what they don’t
• Day-to-day tooling: what’s useful versus what’s a demo
• The statistics/ML overlap: where they agree, where they don’t

Getting started again

The blog got a makeover too. Migrated from Hugo to Astro, cleaner design, dark mode that works. Small things, but they make it more pleasant to write.

If you want to follow along, there’s a subscribe link in the nav. Or find me on Twitter/X at @gumgumeo.