This is a place for my passing thoughts.
If you’re nodding or fuming, send me an email. I like both :)

• # Jan 25 2026 People will care about anything if you give them a reason to. The mind craves anomalies, paradoxes, results that feel wrong. That’s your hook. Clickbait is fine, you just have to earn it by adding context and detail later.
• # Jan 22 2026 Taste predicts status. Status predicts taste. It’s human nature to seek like-minded people, thus forcing you to be around those of similar status. Switching tastes is hard: potential mistakes, feeling like you don’t belong, and why bother when current tastes are already fun? Well, higher status might give access, but lower status could give you comfort in the bubble (*).
• # Jan 19 2026 Movie Theatres are a great example of forced attention - dark room, limited movement, big screen, and no easy exit. As kids this was everywhere (home, school). As adults, freedom means choosing one’s own constraints.
• # Jan 16 2026 Going from a millionaire to a billionaire is same as going from having a thousand dollars to a million dollars! But for every millionaire, there are about 10 people with $1,000. For every billionaire, there are 30,000 millionaires. Same multiplier, 3,000x harder.
• # Jan 15 2026 Sometimes not knowing how hard it is to do something is a good thing. The momentum gets you through for a while and then sunk cost fallacy kicks in to keep you going. What comes out could be worthwhile. Examples: Nvidia, ASML, SpaceX.
• # Jan 11 2026 When you want to convince someone about something, the details matters. The audience can often tell if something feels genuine even if they have no clue about the reality of the subject. Something good but unprofessional feels worse than something mediocre but professional.
• # Jan 8 2026 Long ago, Western empires (Macedonian, British, Portuguese, etc) raced to reach the East - India and China. Today, people from the East live and build everywhere. A reminder that people make places and shape ideas.
• # Jan 7 2026 Science isn’t brilliance. It’s the privilege of failing repeatedly, and persisting long enough to be called “creative.” Material stability is essential, but mentally you have to live with discomfort.
• # Jan 6 2026 AI coding agents are great for avoiding attachment to ideas. The longer you spend coding something, the more attached you get to the implementation. Coding agents can help you fail faster, making failures more acceptable.
• # Jan 6 2026 In biology, there are two ways to explain something: mechanical or evolutionary. Evolutionary understanding is deeper. We often ignore the evolutionary pressures that shaped computer system designs. Should one apply this lens when designing systems?
• # Jan 1 2026 There are more royalty today than ever before in human history. Yet we think of royalty as a thing of the past. Royalty has the luxury of ignorance about the system around them.
• # Dec 27 2025 LLMs model language about reality, not reality itself.
• # Dec 27 2025 If humans have been shaped so thoroughly by language that most never access raw experience (birth, death, awe - something that can’t be labeled) anyway, is the difference between human cognition and LLM cognition smaller than we’d like to believe?