psemme

Trust in the age of fake everything.

Mike Bloch recently published a clear-eyed framework for defensibility in the age of AI. His filter is simple and correct: hard to do, or hard to get? AI compresses the time it takes to do things. It does not compress the time it takes for things to happen. From that distinction he derives five durable moats:

• compounding data
• network effects
• regulatory permission
• capital at scale
• physical infrastructure.

All five share a common property: they required years of real-world elapsed time to accumulate, and no amount of intelligence can run that clock faster.

I think he’s right. And I think there’s one more layer to name.

Near the end of his essay, Bloch pauses.

“Does trust become its own moat when AI is doing more of the work? Someone has to be accountable when things go wrong, and the institution that bears that liability might become more valuable, not less.”

He flags it as a genuine uncertainty, then sets it aside, unclear whether it stands as its own category or emerges from the five he’s already named.

It doesn’t sit cleanly alongside them. It sits underneath. But it behaves like a moat in its own right.

To see why, it helps to ask a more basic question: what is trust, actually?

Trust is a decision to rely on something under uncertainty. It only matters when there is risk, when you don’t know for sure, and it matters if you’re wrong. We infer it, we don’t calculate it. We look for grounding. Was someone actually there? Has this been consistent over time? Is anyone accountable for what happened?

For a long time, those questions had natural answers. Reality produced signals, and those signals were hard to fake. AI breaks that. It becomes possible to generate outputs — text, images, decisions, behavior — that look indistinguishable from reality without ever being grounded in it. Appearance separates from origin. When that happens, the deeper question surfaces: what signals remain trustworthy?

Bloch’s five moats are one answer. They all rely on time that cannot be compressed. But that raises a harder question. If time is the constraint, what is time actually producing?

At the deepest level, it produces presence. Not intelligence. Not output. Presence. Real people in real places, doing real things, at particular moments that do not come around twice.

AI can generate artifacts. It can simulate the traces left behind. What it cannot do is go back and have been there. Either you were there or you weren’t.

On its own, presence is just a fact. It becomes meaningful when another person can attest to it. That is where witnessing enters.

A witness does more than observe. A witness attaches identity and accountability to a moment. They don’t say the evidence looks plausible. They say: I was there. I saw this. I am willing to have my name attached to that claim.

In a world saturated with synthetic outputs, that distinction becomes critical. The scarce thing is no longer just information. It is grounded information tied back to a moment of real-world presence, and to a person who can stand behind it.

This is where Bloch’s framework extends. His five moats are all accumulations, things you build up and hold over time. Data accretes. Networks grow. Capital concentrates. Licenses get issued. Infrastructure gets built. Each is the output of elapsed time.

But elapsed time is not empty. What fills it are moments of presence: people joining networks, operators generating data, regulators granting approval, builders installing infrastructure, capital allocators making commitments. The assets are what remain. Presence is what had to occur for those assets to exist at all.

Witnessing is the mechanism that makes that presence legible.

A witnessed moment cannot be precomputed or simulated after the fact. The window either opened or it didn’t. Once it closes, no amount of intelligence can reopen it. Each instance is singular. But a system built from many such instances can compound, increasing in coverage, reliability, and trust as more participants contribute.

We already have fragments of this logic. Notaries, legal witnesses, certain archival practices all operate on the same premise: that presence, identity, and accountability matter at the moment something occurs. What’s missing is a broader system that extends that logic beyond narrow institutional contexts and into everyday life, especially as synthetic outputs become indistinguishable from real ones.

That is what Notarism is for.

If Bloch’s framework is right — and I believe it is — then building infrastructure for witnessed presence isn’t a philosophical exercise. It’s structural.

The sixth moat isn’t another asset class. It’s the layer that makes them trustworthy. It’s time you can’t fake because someone was there.

This essay is an argument, not a specification. The practice and infrastructure I’m describing is Notarism. Start at notarist.org