Goals to attempt for the rest of 2025

Four Thousand Weeks

I read Oliver Burkeman’s “Four Thousand Weeks” and it changed how I view life. The book is titled Four Thousand Weeks because if you live to the age of 80, that’s around how many weeks you have in your life. And if you’re like me, you’re already done with a third of that.

Just like every other productivity junkie before me, I picked up the book thinking that it was gonna give me a guide on how to maximize those weeks; that at the end of the book I would have a better idea of the kinds of things that I should be focusing on in order to live my life “correctly”.

What I got instead was the realization that I am severely limited as a human, and that it’s ok.

I can only do the things that I can do. I won’t get the chance to do the vast majority of things that I wanted to do. I will only achieve a few of the things that I wanted to achieve. My life will unfold in a way that is completely different to what my imagination crafted. Things take the time they take. And all of that is ok.

Our ability to see the future is severely limited, and the best we can do is to pick the seemingly next best thing and try to do it.

All of this is a preface to what I want to attempt (!) to do for the rest of 2025.

Here are the things I will attempt

Go gym more. I feel like this is everyone’s goal. My goal is to get exercise 3 times a week. I’ve found that if I set a high bar for myself, I end up psyching myself out and don’t do the things that I should be doing. 3 times a week is nice and reasonable.

Get better at SQL. I come from a machine learning research background where everybody loves dataframes. But working at a big company, you realize that the data that you want to use doesn’t fit nicely in a dataframe. You have to go to the data warehouse and use SQL.

I’ve been using ibis + Trino to circumvent the need for SQL, but I still feel that it’s an important skill to have. I wanna be able to look at somebody’s complicated SQL query and actually understand what’s happening.

Learn a new programming language (that isn’t SQL). The first programming language I learned was C back when I was 14 (shoutout to CS50 and David Malan from Harvard), and I remember how rewarding it was to write stuff in C because you had to be a lot more careful than with a language like Python.

I’ve gotten this itch in the last couple of months listening to people like Chris Lattner (creator of LLVM, the Clang compiler, and Swift), Bjarne Stroustrup (creator of C++), and The Primeagen (goofball, ex-Netflix) talk about programming languages, their design, as well as how they differ to other languages. I’ve listened for long enough. It’s time to learn.

I think I’ll start with Rust or Go. I’ll probably just pick randomly.

Revisit fundamental concepts in experimental design. The scientific method doesn’t change no matter how complicated the things that we’re analyzing are. What is the hypothesis that you’re testing? What are the metrics? Are the metrics the right ones to use? Given two different sets of outcomes (one with the treatment and one without), how can you tell whether there is a difference that is statistically significant?

I saw this tutorial on Experimental Design and Analysis for AI Researchers from Michael Mozer that was a great starting point. I want to be more active about using correct experimental design and analysis in my work.

Get better at giving feedback. This is something that I’ve yet to figure out. The feedback model that I’ve been taught at work is Situation Behavior Impact (SBI) where you outline the situation where you observed the behavior and the impact that it had on you. This grounds your feedback in reality.

But to me, this is the equivalent of somebody giving you the lyrics and tune for Creep by Radiohead. Sure, you can technically sing it. But only Thom Yorke can bring a whole crowd of sad boys to tears with the emotion he brings to the song.

That’s what I think I’m missing: the emotional piece. The tact.

One time after a presentation I got feedback from a senior engineer. I said something during the presentation that could be perceived as a personal attack on another team. I didn’t mean it that way, which he understood, but nevertheless it was a mistake on my part.

But the way in which he gave me feedback was honestly the most impressive part. He was Thom Yorke singing Creep. I don’t know what it was, but at the end of our conversation, I was extremely glad that he told me and I didn’t feel bad at all about my mistake.

That’s what I want to achieve in terms of my feedback skills.

I might not achieve these goals

I’ll add to this list if new things come up. I might not achieve all of these goals, but it’ll be a fun exercise to come back here at the end of the year and see how far I got!