Historical thesis — October 11, 2026: This March 23 essay is one version of my recurring trust, stakes, relationships and identity argument. Read it alongside the dated original manifesto. These are responsibilities and relationships I choose to retain, not a proved boundary around everything AI can ever do or a ranking of whose work deserves to survive.
Around December 2025, I decided to move away from most of the business strategies I had been pursuing.
Author chronology clarification — October 11, 2026: The clarified Chapter 6 account says the decision to sunset SimpleDirect’s fintech business was made around December 2025, the official pivot happened around June 2026, customers were given two years to transition, and the app remains available and online. The strategic decision described here is not an immediate December service shutdown or an independently measured guarantee of operational health.
The SaaS. The acquisitions. The MRR goals. The courses and playbooks.
I thought these information, tool and advice products were exposed to AI competition. That was the concern behind my decision, not proof that every such business had become unviable.
A few years ago, you could build a business around knowing things other people didn't.
You could package expertise into consulting engagements and courses. You could be the person who had the answers. That model became less attractive in the businesses I was reconsidering; expertise can still have value in other settings.
But something survived. And understanding what survived is the most important question anyone can answer right now.
The HVAC thing
We saw a once-in-a-generation technology reshape the entire knowledge economy.
And the best advice the internet could give people was "start an HVAC business."
ChatGPT probably told you that too.
I get the logic — go where AI can't go, do physical work, own a local service business. Fine. But it completely misses the point.
The answer isn't to retreat to a category AI hasn't reached yet. That's a timer, not a strategy.
My answer was to focus on relationships and decisions for which I remain personally responsible. Not because of a technical limitation that'll be solved in the next model update. Because they require being a specific human with a specific history making specific bets.
Trust
Not "trusted brand." Not social proof. Not five thousand five-star reviews.
I mean the kind of trust where someone takes your word without fact-checking it.
Where they hand you money and don't check the math.
Where they know that if you say something, you've already done the work.
That trust lives in shared history. In promises you kept when it would've been easier not to. In a track record that exists in the memories of specific people who dealt with you over years.
A model can simulate trustworthy-sounding language. But trust isn't language. It's time.
Stakes
I post with my name on it. My clients see it. My partners see it. My future enemies see it. If I'm wrong about something, it costs me — money, reputation, relationships.
Most people giving opinions online lose nothing if they're wrong. They're background noise.
You scroll past them to the information, and you should, because why would you listen to someone who has nothing on the line?
I tend to give more weight to an opinion when I can understand the person’s stakes and track record. There's a reason you trust the founder who bet their savings over the analyst who bet their Twitter engagement.
Relationships
Not networking. Not "reaching out on X." Not the event where your name is forgotten ten seconds after the handshake.
I'm talking about the kind where someone picks up the phone at 2 am. Where there's a history that no introduction, no DM, no mutual connection can replicate.
I have maybe five people like that. It took twenty years. There's no shortcut. There's no prompt for it. It takes time and it takes being a person worth knowing.
Identity
This is the one most people skip.
Who are you. Why you. What's yours that can't be copied, prompted, or automated.
I decided to move away from a business that was working because I thought AI competition would weaken its advantage. I looked at what was left and asked what required me specifically. Not my hours. Not my hands. Me.
Most people won't do that exercise. They'd rather learn a new tool, start a new side project, build another thing. Anything to avoid sitting with the question.
If I cannot explain what responsibility or relationship I contribute, I need to reconsider my role; that does not make everyone else’s work replaceable. I can review my contribution without deciding that someone else’s work has no enduring value.
The real question
The question was never "will AI take your job." That was always the wrong framing.
The real question was always: were you ever doing something only you could do?
For my own work, I had to distinguish useful execution from decisions and relationships that required me. Useful execution and collaboration can matter even when they do not require a unique personal identity.
And when something came along that could provide those hands and hours cheaper and faster, the math changed overnight.
The people who were doing something only they could do — they didn't just survive.
They got scarcer. And I value that kind of personal credibility, while recognizing that useful execution, care, collaboration and reliable institutions can still matter.
Trust. Stakes. Relationships. Identity.
Everything else is infrastructure.