weekly recap #4: oct 19 - oct 25, 2026

i have not written a weekly recap since the first week of May. ive found it really difficult to find the energy to write them because ive been trying to immerse myself in working on the problems we are solving at relay.

its very important to immerse yourself in the problem you are working on. when you are surrounding yourself with the same problem day after day, language and words stop being the mechanism of how you reason about it. you truly understand it and start to think about it abstractly. its really hard to get this point if you are splitting your focus.

for me that ends up looking like periods where i solely focus on relay. i have no other side projects or cognitive hobbies. if i want to read or write, its about relay. its intense. but i know i am putting it all on the table.

the problem is that this level of immersion, while necessary to do great work, makes you lose the distance needed to step back and reflect. to challenge how you think from first principles and revisit your base assumptions. it feels like this is a drawback of developing a sophisticated internal model of a problem. the deep immersion that causes your thinking to become automatic makes you start treating certain parts of the problem as fixed. it does not matter if your judgment itself is correct if it is working on top of incorrect assumptions.

ive seen it plenty of times in the wild. very smart people who are laser focused, but it goes to waste because they were operating under the wrong assumptions and were focused on the wrong thing. if they stopped to look around a bit and questioned their assumptions, they might have realized that and switched to work on something better. the map is not the territory.

the best way i have found to maintain good judgment is to examine how other people are thinking about things. there is a never ending amount of high quality essays where a skilled writer is working through their thought process of a problem. these are refreshing. it temporarily pauses your thinking and lets someone elses take over. then when you go back to your own thinking, you have somewhat reset your assumptions. you can notice holes in your thinking that you previously never noticed. things you got used to.

anyways, lets jump into the recap! ive been really interested recently in how other people are using agents. until recently, i have been managing agents per-thread (an outdated base assumption!) but have been trying to use manager-threads (puck from amp) and swarms instead. so ive been trying to read and listen more about agents, swarms, and recent thinking in the space.

what i read this week

• Code Review Sucks Now, but It Doesn't Have To
  • ive largely stopped reviewing code other than skimming it looking for anything obviously wrong. most of the review happens in my terminal, trying to reconstruct what descisions were made in the PR. but i love the solution proposed here to share the initial implementation thread with the reviewer.
• top five regrets of the dying
  • good reminder

  • 1. "I wish I'd had the courage to live a life true to myself, not the life others expected of me."
    2. "I wish I hadn't worked so hard."
    3. "I wish I'd had the courage to express my feelings."
    4. "I wish I had stayed in touch with my friends."
    5. "I wish that I had let myself be happier."

• Poker is a bad game for teaching epistemics. Figgie is a better one.
• Toward a Broader Conception of Adverse Selection
• the market for lemons
• Things to know if you get cancer
• The AI spend is the largest bet in the history of capitalism
• we're here on Earth to fart around
• Some final words on Foursquare
  • brutal.

  • The Foursquare acquisition was a $0 outcome for equity holders – $0 for preferred, $0 for common.

• increase your slope
  • this is targetted towards new grads, but i think this can apply to anyone, stating at any company at any point in their career
• The "whiteboard defense"
  • somewhat related tweet from tibo

    • The difference between having a tool and being a tool is subtle these days and it's easy to accidentally build yourself into being a tool. Don't be a tool.

  • try to protect your thinking. its the hardest it has ever been. read and write!
• Eliminate Sparrows campaign

  • The campaign led to surging insect populations and poor harvests, and was one of the causes of the Great Chinese Famine, which lasted from 1959 to 1961 and killed millions of people.

  • the most insane part of this, is that this is not the first time this was attempted. this was attempted in many other places before, and many of them came to the same conclusion. killing the birds leaves insects to grow out of control and destroy the crops much more than what the birds were doing.
• google patent for Using a benevolent worm to assess and correct computer security vulnerabilities
  • people were talking about abusing the PHP RCE to upgrade a websites PHP version so they wont be vulerable to the RCE from someone malicious. this actually happened before, a worm called Welchia was created to exploit the same vulnerability as the Blaster worm and try to remove blaster from the victims computer. turns out google has a patent for this
• Noam Brown - OpenAI researcher on agent swarms & recursive self-improvement
  • this video has alot of good framings on how we can think about agents working on tasks in human terms, for example:

  • I am shocked by the scale of cognitive effort that you can concentrate in such a short period of time. Think about what 130 billion tokens are. If it were a single human thinking as a full-time job, stretched back to back, 130 billion tokens would be a human thinking for 4,000 years. Eight hours a day, working a normal work week. Starting from ancient Sumeria up till today, a single sequential human thinking that long, concentrated in 88 hours.

• paul graham has such amazing writing
  • Five Founders

    • What I learned from meeting Sama is that the doctrine of the elect applies to startups. It applies way less than most people think: startup investing does not consist of trying to pick winners the way you might in a horse race. But there are a few people with such force of will that they're going to get whatever they want.

  • Taste

    • Good design is often strange. Some of the very best work has an uncanny quality: Euler's Formula, Bruegel's Hunters in the Snow, the SR-71, Lisp. They're not just beautiful, but strangely beautiful.

      I'm not sure why. It may just be my own stupidity. A can-opener must seem miraculous to a dog. Maybe if I were smart enough it would seem the most natural thing in the world that ei*pi = -1. It is after all necessarily true.

• have a life
  • all of the succesful people i know have a life outside of work. in spite of this, its odd to see the new generation in tech brag about who can spend the most hours working.
• augmenting long term memory
• When the data and the anecdotes disagree, the anecdotes are usually right.
  • tweet
• you're forgotten forever. you didn't even notice the beauty of life.