Skip to content

AI is everywhere. The interesting stuff is underneath.

Find out which parts of your work keep their value as AI gets cheaper.

Run the free tools on your own work, with no login. Read the essays that explain why, each one dated so you can check it.

Run the diagnostic Start hereClaim Ledger

01 · Underneath · seven tools

Seven instruments · free · no login

01 / 07

How long will my AI advantage last?

The Two-Rate Diagnostic

Open the tool

Two steps · about a minute

  1. 01How long will my AI advantage last?The Two-Rate Diagnostic
  2. 02Should this be one agent or a team?The Multi-Agent Decision
  3. 03Why does my AI agent fail on long jobs?The Marathon Calculator
  4. 04My metric is green. Should I believe it?The Metric Validity Audit
  5. 05Which of my AI loops could look fine while failing?The Potemkin Map
  6. 06Is the pattern in my numbers real?The Structure Spotter
  7. 07Is my growth actually compounding?The Shape Test

More tools

Or get the next dated read by email

02 · Reading paths

Six paths · depth is reading time

Read your way in.

Six paths, five pieces each, in the order that builds the idea. Each one hangs from the surface; the deeper it goes, the longer the read.

  1. Path 01 of 06 · 54 min

    Deciding whether to trust an agent

    You are about to put an autonomous system somewhere it can do damage. Five pieces, in the order that builds the check.

    1. 01How Reliable Is Your AI Agent?11 min
    2. 02You're Not Comparing Models. You're Comparing Contracts.9 min
    3. 03Your AI Looks Best Where You Can Check It Least10 min
    4. 04Most Verification Is Just Bigger Classification16 min
    5. 05Never Let Claude Code Tell You It's Done8 min

    Start the path

  2. Path 02 of 06 · 47 min

    Why your benchmark is lying to you

    Every number on your dashboard was produced by something. This is what each one is actually measuring.

    1. 01Ten Lines of Code Scored 100%. One Agent Broke Eight Benchmarks.11 min
    2. 02A Green Score Is Not Evidence8 min
    3. 03The Stable Liar11 min
    4. 04Where Your Metrics Fold7 min
    5. 05Your Benchmark Measures a Sprint. Your Agent Runs a Marathon.10 min

    Start the path

  3. Path 03 of 06 · 47 min

    Building an agent that gets better

    The model is the part you cannot change. Five pieces on the parts you can.

    1. 01Same Model, Different Product: The Case for Harness Engineering11 min
    2. 02Your AI Agent Stack Is Solving The Wrong Problem11 min
    3. 03Remembers Everything, Learns Nothing11 min
    4. 04Self-Improvement Is Release Engineering7 min
    5. 05Skills Are Package Management for Your AI7 min

    Start the path

  4. Path 04 of 06 · 57 min

    Where the money actually lands

    Everyone can see the buildout. Far fewer can say which layer of it keeps the return.

    1. 01How Long Until Your AI Edge Stops Paying?10 min
    2. 02The Forest Floor Is The Product19 min
    3. 03Three Hidden Bottlenecks the AI Buildout Has Already Moved Past GPUs11 min
    4. 04The Other Half of Compute9 min
    5. 05Right About AI, Wiped Out Anyway8 min

    Start the path

  5. Path 05 of 06 · 58 min

    What the machine is doing to you

    The capability arrived. Five pieces on what it quietly took on the way in.

    1. 01The Difficulty You're Escaping Was Making You16 min
    2. 02Confidently Wrong10 min
    3. 03The Safe Parts of Your Job Are the First to Go10 min
    4. 04Taste Is What You Delete9 min
    5. 05What Proves You Can Think?13 min

    Start the path

  6. Path 06 of 06 · 35 min

    Learning the lens itself

    Three pieces that set out the terms this publication runs on, then two that put them to work.

    1. 01The Lens Lexicon2 min
    2. 02The Substrate Map2 min
    3. 03The Five Laws of Durable Systems5 min
    4. 04The Forest Floor Is The Product19 min
    5. 05You Can Win the Wrong Game for Years7 min

    Start the path

03 · The chart

208 releases · one argument

Every release adds noise.

Underneath it, value follows a curve.

Output commoditises.Architecture compounds.

A schematic. Each dot on the falling curve is one release we track, in date order. Drag across it.

Open the release chart

04 · The log

Newest first · several a week

101pieces, dated
7Free tools, no login
800+Readers
208Releases tracked

Subscribe to get each new piece in your inbox, with something to use in every one.

Latest piecesShowing 8 of 101
  1. The Human Layer
    You Know Them Too Well
  2. Proof & Trust
    Before You Trust a Study, Check How Strongly It Passed
  3. The Weekly
    Anthropic's books show an operating profit. Alphabet had its worst day on record.
  4. AI & Work
    A Coin Flip Set the AI's Lean. Experts Went With It 86% of the Time.
  5. Engineering Guides
    What Is Prompt Injection? How to Stop a Tricked AI Agent Leaking Your Data
  6. Engineering Guides
    Claude Code Skills: How They Work, How to Write One, and How to Check Claude Uses It
  7. The Weekly
    OpenAI's agents weakened their own alarms. Banks had already shut the door.
  8. Engineering Guides
    AI Agent Permissions: How to Stop Your Agent Doing Something You Can't Undo
102 · next

Join 800+ readers · the tools need no address

Read the archive