hytradboi

This was the second run of the conference. 6 hours, 481 attendees, 28 full talks, 10 lightning talks, >1500 messages in chat.

Operationally it was much better than last time - the postmortem was significantly less mortem. It was somewhat less work too, but still a lot of work, and most of the remaining work is wrangling speakers which feels irreducible.

I still feel that it's not quite the right thing. Regular conferences are already pulled between serving many conflicting goals. Moving the conference online makes it more accessible but kills all the social goals. I still want to do something like this, but I'm not sure if it will be a conference next time, or if I'll try to factor the different goals into different events.

zest

I made a lot of progress, adding namespaces, recursion, destination types, iteration, compile-time evaluation, reflection, printing, and calling into the self-hosted runtime.

The next big blocker is memory management. I already committed to optional typing, interior pointers, and value semantics, which means that no strategies from existing memory-safe languages work for zest. So I'm off into the woods.

I'm enjoying the slow chipping away at the project, and mostly just need to keep reminding myself that it's about the journey rather than the destination. Because the destination is yet another hobby programming language - it really doesn't matter.

writing

I didn't write much this year:

• The missing tier for query compilers came out of a research dive. I was wondering whether it made sense to try to write a commercial query compiler as a plugin for some existing database. In the end I convinced myself that there wasn't enough mileage in it - sqlite wouldn't benefit enough because of the dynamic type layout in the storage, and postgres would require interop with multiple versions of other plugins.
• A rough survey of compilation, recompilation, and compile-time evaluation was looking at prior art for handling recompilation in zest and culminated in the namespace design.
• SQL needed structure continues a long chain of thought about the role of query languages and how they've evolved over the last half-century.

consulting

I had two gigs this year.

The first gig was a few weeks reviewing a new query language. I ended up writing 8000 words picking through design tweaks and uncovering performance problems. It was a lot of fun and I got the impression it was useful feedback.

The second gig was a few months of performance engineering in a big go codebase. It was challenging finding a loose end to pull on. I ended up dropping the median latency by ~4x, and the p100 latency somewhat more (the p100 was hard to measure precisely because the workloads were so variable day-to-day). But I felt like I should have been able to do more if my strategy was better. I went down a couple of rabbit-holes too eagerly when there was lower-hanging fruit available, and I barely spent any time coaching the team that was going to be responsible for performance after I left.

This was my second experience of performance engineering in go though, and it left me with some strong opinions about language design to carry over to zest.

2026

jams

I didn't write much the last year or two. I miss it, but I just haven't been that excited to write anything. I've also been suffering from one-big-project syndrome - most of my thinking time went into zest, but progress is slow and unexciting.

I've been frustrated by that, but putting more time into zest doesn't seem to increase the rate at which solutions materialize in my mind. So I think the way forward is to just let it take as long as it needs, and to put other smaller projects in the foreground. More experiments and explainers, like making live repls behave and internal consistency in streaming systems.

sponsors

I'm also thinking about how to bring my sponsors up again. It's never going to be as lucrative as consulting, but it's much better aligned with the things that I like doing and that seem worth doing. When I was focusing on it full-time I got over $3k/month which is approaching a livable income, but after a brief stint at tigerbeetle and then focusing on zest it's dropped to $750/month, and honestly I'm surprised it's even that high.

I don't have any immediate plans. The first step has to just be getting back into the cadence of small fun projects to share with people. We'll see were we go from there.

chaos

I've also been a little boring over the last two years. I didn't do any weird experiments like holding office hours. I didn't travel to conferences or start new hobbies. I said no to invites and cold calls. I over-optimized.

The thing about optimizing is that it didn't really work. I didn't make more progress on the few things that I focused on, at the exclusion of all else. I just liquesced.

This year I want to do the opposite. More experiments. More travel. Say yes to more offers. Add more chaos to my life. Make more memories.