Sheet 02 · a commonplace book

People

Lines by other people that I keep coming back to. Nothing famous on purpose.

Fig. A
08 lineswritten between the 5th century and 2019, kept here because they keep coming back
01

To come to what you do not know, you must go by a way you do not know.

Orig.Para venir a lo que no sabes, has de ir por donde no sabes.

A mystic from my city wrote the definition of a startup four hundred years early.

02

If you say enough, you are lost.

Orig.Si autem dixeris: sufficit, et peristi.

Bezos calls it Day 2. Augustine got there sixteen centuries earlier.

03

There is no victor but God.

Orig.ولا غالب إلا الله

Nasrid motto ↗carved across the walls of the Alhambra, Granada13th–15th c.

Humility carved into the most ambitious building in my city. Build like you will win, knowing it is not you who wins.

04

We must know. We will know.

Orig.Wir müssen wissen. Wir werden wissen.

The answer to 'you cannot know if this works': you can. Measure it.

05

If a thing is worth doing, it is worth doing badly.

Ship the ugly first version. He said it in 1910.

06

Grow bigger in the face of obstacles.

Orig.Crécete ante los obstáculos.

The only valid answer to an investor's no or a bug in production.

07

Wanderer, there is no road; the road is made by walking.

Orig.Caminante, no hay camino, se hace camino al andar.

A new market has no category until someone walks it.

08

No human is limited.

The impossible limit fell to a plan. Market limits are hypotheses too.

09

Left blank for the next one.

Schedule of reads

16 items · essays and papers I would hand to a friend

Ambition, and what to do with your twenties

  1. R-01You and Your ResearchOpen door, important problems, courage. The talk on how to do work that matters.Richard HammingBell Labs talk1986
  2. R-02What Should You Do with Your Life?Cold emails change lives at twenty.Alexey Guzeyguzey.com2021
  3. R-03Some things I believeSlow is fake.Nat Friedmannat.org2022
  4. R-04FastEmpire State Building: 410 days. Calibrate how fast is possible.Patrick Collisonpatrickcollison.com2019
  5. R-05QuestionsPatrick Collisonpatrickcollison.com2019
  6. R-06If your product is Great, it doesn't need to be GoodPick three things. Make them perfect. Ignore the rest.Paul Buchheitpaulbuchheit.blogspot.com2010
  7. R-07How Technology GrowsDan Wangdanwang.co2018

AI, the ones people skip

  1. R-08The Scaling HypothesisWhat being right too early looks like.Gwerngwern.net2020
  2. R-09Agentic Misalignment: How LLMs could be insider threatsRequired reading if you deploy agents.Lynch et al., Anthropicanthropic.com2025
  3. R-10Scaling MonosemanticityTempleton et al., Anthropictransformer-circuits.pub2024
  4. R-11On the Biology of a Large Language ModelLindsey et al., Anthropictransformer-circuits.pub2025
  5. R-12The Illusion of ThinkingShojaee et al., ApplearXiv2025
  6. R-13Three ObservationsCost per token falls 10x a year. Build accordingly.Sam Altmanblog.samaltman.com2025
  7. R-14The Most Important CenturyHolden Karnofskycold-takes.com2021

Under FixedGap

  1. R-15MediaPipe Hands: On-device Real-time Hand TrackingThe paper underneath FixedGap.Zhang et al.arXiv2020
  2. R-161€ Filter: A Simple Speed-based Low-pass FilterHow we stopped the hand from shaking on screen.Casiez, Roussel, VogelCHI2012