A manifesto for the age of AI

Stupidness
is no excuse.

The answer to almost anything is thirty seconds away. Not knowing used to be an excuse. Now asking takes thirty seconds. Curiosity is free. Ignorance is expensive. This is the front door of a small, useful knowledge hub for a smarter internet.

Fig. 002 / The Stupidity Tax

What not-knowing costs.

A small, non-exhaustive ledger of expensive failures that one basic question, check, or safeguard might have prevented.

  • $460MKnight CapitalOne of eight servers missed a software update and continued running obsolete trading code. In 45 minutes, the system generated a $460 million loss.
  • ≈$125MMars Climate OrbiterLockheed Martin software produced thruster data in imperial units. NASA's navigation software expected metric units. The mismatch went undetected, and the spacecraft was lost near Mars.
  • 500M emailsNHS reply-all stormA faulty mailing list allowed one email and its replies to generate approximately 500 million messages in 75 minutes.
  • ≈$18.5MMariner 1A missing overbar in a guidance equation—combined with a guidance-antenna failure—contributed to the destruction of the Venus-bound spacecraft five minutes after launch.
  • ¥40.7B ≈ $347MMizuho fat-finger tradeA trader entered 610,000 shares at ¥1 instead of one share at ¥610,000. The Tokyo Stock Exchange accepted the order, and its system failed to cancel it.
Fig. 003 / The Excuse Detector

Paste an excuse.Get an AI verdict.

0/500

AI-generated verdict via Google Gemini. Not advice. Privacy

Need inspiration?

Public verdict history

Published excuses and answers — expand for the full archive.

Loading history…

Fig. 004 / The Manifesto

Seven tenets, no filler.

  1. 01

    Curiosity is free.

    Ignorance is expensive. The gap between not knowing and knowing is now measured in seconds, not semesters.

  2. 02

    Ask the machine before you ask the mob.

    A quiet question to a model beats a loud opinion in a group chat, every single time.

  3. 03

    Confidence is not competence.

    The loudest voice in the room is often the one that hasn't looked anything up. Verify, then speak.

  4. 04

    Read the thing.

    The contract. The docs. The label. The source. Two minutes of reading has saved more lives than any hot take.

  5. 05

    Prompt like you mean it.

    Vague questions get vague answers. Specify who, what, when, and why — and the model will meet you there.

  6. 06

    Being wrong is fine. Staying wrong isn't.

    Update your beliefs when the evidence updates. That's not weakness. That's the whole game.

  7. 07

    Stupidness is a choice. So is the other thing.

    Every day, you get to pick. Choose the interesting one.

Fig. 005 / The Field Guide

Coming next.

Vol. 01Soon

Fact-check in 60 seconds

A pocket routine for verifying a claim before you repeat it. Works on tweets, headlines, and uncles.

Vol. 02Soon

Prompt like you mean it

How to talk to a model so it stops telling you what you already knew and starts telling you what you didn't.

Vol. 03Soon

Spot the confident wrong answer

The tells — human and machine — that a very sure-sounding sentence is quietly making things up.

Fig. 006 / The Newsletter
Coming soonBest of WeekWaitlist

One useful thing a week.

The weekly letter and Best of Week roundup are not shipping yet. Join the waitlist and we will email you when the first issue goes out — one thing worth knowing, one prompt worth stealing, zero engagement bait.

Emails are stored securely for the waitlist only. No spam. Unsubscribe when the letter launches.