Things I actually do
Not a list of aspirations. These are the things I spend real hours on. Most of them are the same activity wearing different clothes: incomplete information, a decision anyway, and a scoreboard afterward. The rest are maintenance on the machine that does the deciding.
Chess
Chess club as a kid, several hours a day. I came back to it this year and I am training two- and five-minute blitz. I built myself a coach for it rather than buying one.
Poker
Same appeal as chess with the information removed. I care more about whether a decision was correct than whether the hand won, which is why the app I built grades the decision and not the result.
Football and baseball, instrumented
I watch a lot of NFL and MLB, and I could not leave it alone. An agent pulls batter-versus-pitcher history every morning at 4:30, a live rail texts me mid-game when something worth seeing is happening, and my fantasy lineups run on a daily automation. Fandom, but with a scoreboard.
Detective dramas
My fiction of choice, and it is the same shape as chess and poker: somebody with incomplete information refusing to guess before the evidence is in.
Cold, heat, and light
Heat and light most days. Cold was daily for years, and right now I am running the opposite experiment, no plunge at all, to see what changes. Testing protocols on myself is the point.
Breathwork and meditation
Daily. Two minutes of slow exhales before anything hard changes what it costs, and knowing that firsthand is worth more than citing a study about it.
Photographs
I like taking pictures. Mostly with my phone, mostly outside, mostly early.
Trails and beach walks
The Orange County coastline, most days. It is where the hard problems get worked on without a screen in front of them.
Longevity, practiced
Training, protocols, and an unreasonable number of smoothies. The goal is to still be doing everything else on this list decades from now, and I only keep the protocols I actually follow.
Reading, distilled
I have run 27 books through a distillation system that maps an author's framework onto my own. The first run caught two fabricated claims in its own output on a verification pass. That result is why I trust the system a little more and the model a little less.