Every year I write a review. Not for anyone else — for myself. To force the kind of reflection that doesn’t happen naturally in the rush of daily work.
2025 was a year of depth over breadth. Here’s what I mean.
What I built
The REACH automation system. The project I’m most proud of. What used to take a marketing analyst 3-4 hours — Persona targeting, coupon creation, campaign configuration — now takes 20 minutes with AI assistance. The system isn’t just faster; it’s more consistent and less error-prone.
The intelligent subsidy algorithm PRD. Three versions, each more rigorous than the last. V3 incorporates uplift modeling, causal inference, and budget optimization. It’s the most technically sophisticated product spec I’ve written. Whether it gets built is another question.
This website. Small thing, but meaningful. Having a place on the internet that’s mine — not a social media profile, not a company page — feels important. I want to write more in 2026.
A lot of small games. None shipped. All educational. I understand game loops, difficulty curves, and player psychology much better than I did a year ago.
What I learned
Causal inference is the most underrated skill in product. Everyone talks about data-driven decisions. Almost no one thinks carefully about causality. The difference between correlation and causation isn’t just academic — it’s the difference between a marketing campaign that works and one that wastes money.
AI tools are force multipliers, not replacements. I use AI heavily now. It’s made me faster at execution. It hasn’t replaced judgment, strategy, or the ability to read a room. Those are still human skills.
Depth compounds. I spent 2025 going deep on a few things — subsidy algorithms, causal inference, game design — rather than skimming many things. The compounding effect is real. Deep knowledge in one area makes adjacent areas easier to learn.
Writing clarifies thinking. I wrote more in 2025 than any previous year. Every time I tried to write something down, I discovered I understood it less well than I thought. Writing is thinking.
What I want to change
Ship more. I have too many projects in “almost done” state. In 2026, I want to finish things. Done is better than perfect.
Write publicly. I’ve been writing for myself. In 2026, I want to write for an audience. Not for the audience — for the discipline of writing for someone other than yourself.
Build more in public. The website is a start. I want to document the process of building things, not just the results.
Invest in relationships. 2025 was heads-down. I want 2026 to include more conversations with people who think differently than I do.
The thing I keep coming back to
The most valuable thing I did in 2025 wasn’t any specific project. It was developing a clearer sense of what I care about and what I’m good at.
I care about systems — how things connect, how incentives work, how small changes propagate through complex systems. I’m good at translating between technical and business thinking. I like building things that make other people’s work easier.
That clarity is worth more than any individual project. It makes every subsequent decision easier.
Here’s to 2026.