The Career That Looks Scattered
The meeting had eight people in it who’d spent their careers in federal contracting. They knew procurement regulations, compliance frameworks, the full audit trail from solicitation to close. The question on the table: how do you govern AI access to contract data?
The conversation kept circling back to the same place. How do you constrain the model? What guardrails do you put on it? What do you tell it not to do?
I kept seeing a different problem. The federal contracting world already has this answer — it just lives under a different name. You don’t constrain what a contractor can do through behavioral rules. You define what they’re cleared to access. Scope, not guardrails. Once the clearance is defined, the question of what the contractor “should” do is mostly answered by what they’re allowed to reach.
Nobody in that room had built an accounting system. Nobody had spent time in warehousing watching what happens when inventory control breaks down. I had done those things before I ever touched federal contracting — which meant I arrived with a different set of analogies.
That’s what 25 years across domains actually buys you. Not breadth. Analogies.
The career that looks scattered
Warehousing to accounting to IT to federal contracting to insurance to AI. From the outside, that looks like someone who couldn’t pick a lane.
From the inside, it looks like this: every one of those domains is fundamentally a problem of controlling the flow of information, authority, and accountability through a system with imperfect participants and incomplete data. The vocabulary is different in each. The underlying structure is the same.
Warehousing taught me that the gap between what the system says and what’s actually on the shelf is a control problem, not a counting problem. Accounting taught me that the number taking three days to produce is almost always wrong in the same direction as the incentive to be wrong. Federal contracting taught me that the compliance framework that looks bureaucratic from outside is usually load-bearing governance that someone paid for with a failure.
When AI showed up as the next domain, I already had vocabulary for the actual problems. Not the AI problems — those are surface. The governance problems. The accountability problems. The gap between what the system says it’s doing and what it’s actually doing.
What specialists can’t see
Specialists go deep in one domain and build genuine expertise. That’s not nothing — it’s usually most of what you need. But depth in one domain can make the analogies invisible.
The federal contracting team in that meeting was expert in federal contracting. They were solving an AI governance problem by extending federal contracting frameworks — which is actually correct. Federal contracting has 70 years of hard-won governance design baked into it. But they couldn’t quite see why it was correct, which meant they couldn’t see where the analogy broke down.
I could see both sides of the analogy because I’d lived on both sides.
This happens with every cross-domain move. When you’ve built an accounting reconciliation process, you recognize a data integrity problem in a different context immediately. When you’ve designed a clearance model for one type of access, you see where it applies in a different system. The domains are different. The structure of the problem isn’t.
How it compounds
Each domain you add doesn’t just give you more knowledge — it gives you more analogical surface area. The fifth domain teaches you something the first four couldn’t see. And because you have the prior four as reference points, you recognize the lesson faster and apply it further.
The other thing about this moat: it doesn’t deprecate. Depth-based specialization has a shelf life tied to the domain. Cross-domain pattern recognition doesn’t — it just accumulates more surface area. A new domain extends the prior ones. It doesn’t replace them.
The AI governance framework I’ve been building is a direct output of that accumulation. It’s not just “AI security.” It’s a synthesis of what federal contracting figured out about scope-based authority, what accounting figured out about audit trail integrity, and what warehousing figured out about the gap between the ledger and the shelf. Those aren’t metaphors. They’re design decisions.
The signal is always in there. You need enough reference points to know what it sounds like.