Private notes, made public. · 21st Aug, 2026

Hey, it’s Gabriel.

These are my private notes this week on building wealth, raising my family, and breaking generational patterns.

This week
⚡ I’m building an AI that takes money off my card when I drift
❤️ no letter last week, because I was in Bangkok with my wife
💰Stripe paid over $7 billion for an AI company, and their leaked letter says why
🧬 why “three times a week” never worked, and daily did

Here is the problem nobody warns you about.

AI did not just make me faster. It made everything possible. Every idea I have is buildable this afternoon, and when everything is buildable, everything is startable. So I start. I jump from task to task all day and end it having touched nine things and finished none.

That is my biggest challenge right now. Not capability. Completion.

So I am building something I have started calling a drift system, and Adam, my chief of staff agent, runs it. I know, I’m building something to help me complete building something. 🤣

Adam watches my screen. When I drift off the one thing I said I would finish, he calls me. Not a notification I can swipe away. A phone call. If I do not answer, he emails. If I still do not respond, he deducts money from my credit card and donates it to an organisation I would hate to fund.

The rule underneath it is the part that actually matters: I cannot start the next thing until the current thing is done.

Long-time readers have met Adam (renamed from Sonny) before. Back in #008 I told you I had built my own employee. That was an assistant. This is an employee I have given permission to punish me, because I have learned that my good intentions at 9am are not evidence of anything at 2pm.

And yes, I see the irony. I am using AI to protect myself from what AI did to my attention. That is the point. The tool that handed me infinite options is the only thing fast enough to notice when I take all of them at once.

Straight with you: it is not finished. Some days it works, some days it stays quiet when it should have called. But when it works, I am giving it to my Two Hour CEO community for free.

When did you last finish something before you started the next thing?

Building the machine is the easy half. Inside Two Hour CEO, I show you how to build agents like Adam that actually run things, step by step, even if you’ve never touched a line of code.

That is the honest reason your inbox was quiet. Zoey and I took time away, just the two of us.

We have a two-month-old at home. Three kids now. We are genuinely blessed to have help, and knowing that, we still chose to go. I want to say why plainly, because I think a lot of parents get this backwards.

Your marriage matters more than your kids.

If you truly love your children, then the most important thing you can give them is two parents who are still in love with each other. Love does not arrive in your kids directly. It overflows into them from what is happening between you and your spouse. A strong marriage is not time stolen from your children. It is the source they drink from.

What the trip showed us was not comfortable. Our current lifestyle is no longer sustainable. We have three children and we are still living like we have none: late nights, late mornings, no energy left. And when you wake up late with no energy, you skip the gym, you skip reading, you skip the work that mattered. The whole day inherits the night before.

So we made a pact. Lights out by 9:30pm. Up at 5:30am. Together, not each of us separately, which is the part that makes it work. One week in, it has gone well. We missed one morning.

But the bedtime is not the interesting part. Anyone can pick a bedtime. What we did differently was sit down and name, in advance, every single thing likely to break it. And then talk about what happens when we fail, because we will.

This is where grace comes in, and I mean that practically, not as a nice word. When we break the pattern we do not condemn ourselves and we do not sit in guilt. Guilt feels like accountability and does the opposite: the moment you decide you are the kind of person who fails at this, you stay in the pattern to prove yourself right. So we give each other grace when we fail, reset, and go again the next day.

I checked whether that is just something I tell myself. It is not. In the habit research, missing a single day has no measurable long-term cost to a habit you are building. The damage never came from the miss. It comes from what you decide the miss means.

Then we add a guardrail, so the same thing cannot take us out twice. Here’s a real example. Our biggest pattern-breaker is meeting friends on weeknights: the conversation is good, so it runs late, so the morning is gone. So we made a call. No friends on weeknights for the next few months. Friday only, because Saturday is the one day we can sleep in.

That is not willpower. That is removing the decision before it gets made at 9pm with people we love sitting in front of us.

What is the one thing most likely to break your plan this week, and what is your rule for it?

The News: Stripe, a payments company, agreed this week to buy OpenRouter, an AI company, for a reported $7.5 billion. OpenRouter takes a request and routes it to the best of more than 400 AI models from over 80 providers, choosing on cost, speed and reliability. In May it was valued at about $1.3 billion. Roughly five times, in three months.

The Leak: A letter from Stripe’s founders to their investors got out, and it is the most interesting document I have read this month. Their stated reason for the deal, in their own words: they “decided that January 1 marked the beginning of the singularity” and have been operating on that basis since. But the line I keep thinking about is their framing. Every company now runs two pipelines, one for money and one for intelligence. Stripe already owned the money pipeline. This week they bought the intelligence one.

A payment gateway bought an AI company. Sit with that for a second.

The Thesis: “The agent economy” gets talked about as one thing. It is really five, and they arrive at different speeds.

Already real: a manager agent delegating work to smaller agents inside one company. That is me, today. No crypto, no marketplace, just coordination.

Beginning now: a company hands its agent an approved supplier list, a spending limit and permissions, and lets it choose. Automated procurement.

Early: machines paying other machines in small amounts. The plumbing works and real transactions are running, but the dollar volume is still tiny next to how loudly it is discussed. I want to be honest that I know that.

Very early: an agent finding an unknown provider in a public registry, judging it, hiring it, paying it. That needs fraud, identity, reputation and disputes solved first. This is where the exciting projects aim and where most of the money will be lost.

Further out: an agent that finds its own customers, earns, covers its own costs and reinvests. Demonstrable in pieces. Legally, someone still owns it.

The Position: This is why I am investing more in Cloudflare long term, which you have watched build across several letters. It is also why I am opening a starter position in Coinbase tonight. Small on purpose: about 2% of my portfolio, with room to add as the thesis proves out.

My reason is not that crypto is back. It is that crypto’s real customer was probably never us. Human payments were built around a person opening an account, typing card details, clicking confirm. Agents cannot do that and do not want to. They need small, frequent, programmable payments, an identity of their own, and a wallet with a limit on it. Crypto was accidentally good at exactly that, and it spent years hunting for a user. The user may turn out to be software. Coinbase is building for that, and unlike most companies talking about this, they are shipping.

The Confession: This week I got to the gym almost every morning. Almost. There were mornings I could not get up, because my body is still adapting to the new 5:30 wake-up and has not caught up yet. Hopefully next week is cleaner.

The Backstory: Missing them reminded me how long the gym was a problem for me. For years I kept trying to start, and every time I started the same way: three times a week. It never held. Not once.

What finally worked was 2022. I was in Bali for an extended stretch and made myself a different promise. Not three times a week. Every day. That was the beginning of actually having the habit, and I have had it ever since.

Why It Worked On Me: A warning before the mechanism, because this one is genuinely not for everyone.

My reason is ADHD. Somewhere in how my brain works, I do not want to think. Thinking is the expensive part. Give me a system and I will follow it, because a system means there is nothing left to decide.

“Three times a week” is not a system. It is a negotiation with openings in it. Every morning it asks me a question, and every morning “maybe tomorrow” is available and reasonable.

“Every day” asks me nothing. So it stacks onto what I already do, with no decision in the middle: wake up, brush teeth, wash up, gym clothes, out the door. The habit rides on the back of a habit I already had.

The Evidence: I checked whether this is just me, and the research backs daily harder than I expected. One large fitness study put the minimum dose for a gym habit to become self-sustaining at around four sessions a week for at least six weeks, after which showing up stops depending on you deciding to. Which means the “three times a week” every beginner gets told to start with, me included, sits just underneath the threshold where the habit would have carried itself. No wonder it kept collapsing.

The Honest Part: Daily does not mean an hour, and that is the trap that stops people trying it. On a bad day, daily means ten minutes, or a walk. Both beat not going, and both keep the chain intact. I train about an hour now, but that is the output of years of showing up, not the entry fee. Do not copy my hour. Copy my ten minutes.

As promised, my private notes. No fluff.

See you next week,

Gabriel Judah

P.S. The drift system, Adam, the loops that run my company: they are all built the same way. Inside Two Hour CEO, members are building their own, step by step. You don’t need to be technical. When the drift system works, they get it first, and free. Grab a seat here.