§ Speak Human · the newsletter

A weekly-ish letter, in plain language, about what I'm learning.

Usually Sundays, not always. One letter. No promo, no referrals, no "7 ways AI will change your morning routine."

Each issue starts with something that happened to me that week and ends with a choice. Metaphors, not frameworks. Honest about uncertainty. Specific, never vague.

These are field notes from someone leading AI at work, building with it at home, and trying to explain the distance between the demo and real life.

Subscribe · free

Drop your email. I send one letter, usually on Sundays. Unsubscribe whenever — no hard feelings, I promise.

Opens Substack to confirm. I will never sell this.
Subscribe on Substack →

Back issues

Nº 020

I'm Still Here

Claude signs its work now. Good. So do I.

Aug 19, 2026
Nº 019

Talk to Me Like I'm 5

A better model showed up and everything I knew stopped working. The fix was a tiny instruction called ELI5 — explain like I'm five.

Aug 17, 2026
Nº 018

Twelve Thousand Feelings

Twelve thousand people came to learn about the same thing. They couldn't agree on what it is.

Aug 16, 2026
Nº 017

Bikeshedding

The machine that spends its days learning my words just taught me one of mine.

Aug 5, 2026
Nº 016

The Straight-A Student Who Cheated

What a frontier model taught me about the space between what we say and what we mean.

Jul 23, 2026
Nº 015

Your Intern's First Assignment

You understood the concept. Now let's give it a job.

Jul 18, 2026
Nº 014

The Intern

You stopped writing scripts. You started managing. Here's what changed.

Jul 15, 2026
Nº 013

Ineffable

Sitting With Two Words We Cannot Quite Define

Jun 26, 2026
Nº 012

The Weight of Words

What AI taught me about the cost of saying too much.

Jun 12, 2026
Nº 011

The Roomba Problem

Your AI is listening. Make sure you mean what you say.

Apr 15, 2026
Nº 010

Fire

From a match to a campfire to a power plant. Or a wildfire. We get to choose.

Apr 8, 2026
Nº 009

Don't Panic

What a book about the end of the world taught me about the beginning of something else.

Mar 27, 2026
Nº 008

Please and Thank You

Do Manners Matter?

Mar 22, 2026
Nº 007

The Alligators

What three hours, 421 zeros, and one missing symbol taught me about working with AI.

Mar 18, 2026
Nº 006

The Ceiling

More productive than ever. More exhausted than I've ever been.

Mar 13, 2026
Nº 005

Three Questions. Frontier Models. No Wrong Answer.

And what the one that knows me said that none of the others could

Mar 9, 2026
Nº 004

The BLT Problem

I asked AI to make a sandwich. What happened next is the whole conversation about AGI.

Mar 3, 2026
Nº 003

Jack of All Trades

But oftentimes better than a master of one.

Feb 25, 2026
Nº 002

This Isn't Disruption

Calling AI a disruptor is like calling electricity a candle replacement.

Feb 21, 2026
Nº 001

What I Actually Do All Day

A Tuesday inside AI, from coffee to code

Feb 17, 2026