January 5, 2025: Me Day - Age 39

The biggest thing that happened this year is that I handed over the reigns at CodeCombat, which has been my work baby for 12 years, in order to work on AI safety to reduce existential risk from future superintelligent AI systems. I've been doing Me Days annually for just a few months longer than that, with the first one in October 2012, and CodeCombat started in January 2013. Feels a bit unreal to be moving on, but duty calls.

Technically, this year's Me Day personal inventory should have been done on October 4, 2024. But, I apparently am bad at making time for myself, in a surprise to no one but me. I captured some of this data around October-ish, but didn't sit down to really think, reflect, and plan until two months later–the weekend before I start a new job, in a new leadership role at an AI safety research startup, Gray Swan AI.

In the world of AI, there's so much to do, and timelines are short, so this year's entry will also be a bit accelerated! Let's oh so briefly pause and reflect.

Recall the mountain.

Each year in the summer or early fall, I backpack up a mountain for a couple days. Reminding myself of the recent trip helps me look up, figuratively, and ideate some mountaintop experiences for next year, and then work back from there to plan the year's training. This past year, George and I camped at Upper Wildcat Lake, the hike for which which crosses Snow Lake on the way up. My foot endurance was low–I needed to practice wearing shoes more–but it was a gorgeous trip.

Snow Lake during 2024 hike to Upper Wildcat Lake

This next year, we are aiming for attempting a Mount Rainier summit, which will go well beyond the hiking I'm used to and into some very aspirational levels of mountaineering. We did just Camp Muir a few years back, and that was a cardio challenge at altitude for me. A lot of training is in store for me this year to try to keep up with George and Zach to go all the way.

Measure physical stats and reflect on fitness.

Third year of working out on the Tonal home gym, 245 total workouts. Got upper and lower body strength a little more balanced, but didn't do as many workouts this year and didn't make much progress (strength score 1314 -> 1341). My main limiting factor is still endurance, as measured both by my observed workout stamina during strength workouts and hiking and by measured VO2 max. That, and lower back health–I recently threw out my back sitting too much after doing a long workout with poor overhead press form. Approaching 40 and rethinking how I approach fitness–maybe there are other aspects I should balance other than seeing how much I can deadlift until failure. Yes, yes, so obvious. I'm hiring an online coach to fix blind spots in my training and help keep me honest on preventative mobility work.

Also unpleasantly, for the first time in my life, I gained some unwanted weight, going from 143 lbs to 150 lbs over the year. I think I mostly haven't been training enough, but I may need to pay some attention to nutrition–even, perhaps, eating something other than MealSquares. Lame!

Weight: 149.9 lbs +6.7 lbs
Body fat: 15.5% +2.3%
Chest: 33.75" +0.50"
Right arm: 12.00" +0.25"
Left arm: 11.75" +0.50"
Waist: 29.25" +1.00"
Hips: 36.00" +0.75"
Right leg: 21.50" +0.50"
Left leg: 21.00" +0.75"
VO2 max: 38.7 mL/kg/min +2.8

Goals for next year, which I hope to achieve by setting and sticking to a more balanced training plan with the coach:

  1. Strengthen lower back to prevent injury
  2. Build up endurance for a Rainier hike this summer
  3. Restore body composition


Measure my max push-ups and pull-ups.

44 push-ups, +5: fast, strict, chin touching the ground.
21 pull-ups, +3: fast, no kipping, chin above bar.

This is very similar to two years ago; last year seems to be just a temporary dip. This is still without training push-ups or pull-ups, just similar moves on Tonal. I was considering starting to report Tonal deadlifts here on a recurring basis (40x200), but now I've decided this sets the wrong incentives.

Read from a favorite book.

I didn't take any special time to read a favorite book this year, although I really liked most of the books I did read. No 5-star books this year. I did finally give up on finishing Liu Cixin's Dark Forest trilogy in the original Chinese and read it in English–that was perhaps my favorite of the year.

Year   Books   Pages

2008   28      8,946

2009   18      7,244

2010   12      3,312

2011   16      6,469

2012   35      13,985

2013   33      11,760

2014   15      5,796

2015   30      18,694

2016   16      8,487

2017   17      6,824

2018   14      7,423

2019   21      13,901

2020   7       3,408

2021   3       1,538

2022   6       13,460

2023   22      17,459

2024   19      9,541

Kid photo break

I continue to enjoy the kids more the older they get. We did a lot more in the way of family activities, less in the reactive "take care of young kid" mode.

Get blood tests.

TODO: update this when I get results back.

Take stock of finances.

The year went ok. I handled the money side according to my convictions.

(If you're me, you get to see finance stats here. Since you're not me, you see nothing.)

Do Quantified Mind testing.

I had to resurrect Quantified Mind on a new domain, and it's working, but a bit buggy, so the age variable didn't capture correctly, and so automatic trendlines didn't work. Still, we can see that I didn't really see any signs of cognitive decline yet. Random new best ever score on choice reaction time, within-range scores on other tests.


Note that all these graphs have cut-off vertical axes, and in most cases very little variation compared to population variation.

Have a spectacular dinner with Chloe.

Chloe really outdid herself this year with the cutest, fanciest, local-est Fliippino/Pacific Northwest fusion place in Seattle. I couldn't get enough, and then at the end of the tasting menu, they were like, who's still hungry?, brought out some extras, and then I did get enough! Generally, frequent date nights with Chloe have been fantastic this year, and the kids are no longer even acting out with the sitters.

Reflect on the past year

I followed my convictions to change industries from edtech to AI safety, which was more strategic than I've been in a long time.

What went well? This past year:

  • Handed off CodeCombat leadership responsibilities over a slow, smooth transition
  • Learned a lot about how to do ML work
  • Built more connections in the AI safety ecosystem
  • Got my first real job ;)
  • Had more and more fun with family
  • Continued to stay healthy and not fall ill despite all the school-aged kids around
  • Redid my website to be hosted on the same stack as my blog
  • Built a bunch of fun AI side projects and got super productivity boosts from AI

What didn't go well?

  • Didn't work out as much as I wanted to
  • Didn't do as much socially as I'd like
  • I wrote two years ago here: "AI is an increasingly dire threat, and I haven't done anything about it". It's... even direr, so I'm trying to do something about it, but the overall situation could be better!

How are my relationships?

  • I need to spend more time here!

Add at least one tradition.

Planning the big summer mountain hike is a good motivating anchor for the year's fitness goals.

Destroy at least one habit.

Getting rid of the "take a long walk" section in favor of the mountain thing.