STRATEGIC THESIS · THOUGHT LEADERSHIP
Pillars
Each pillar is a thesis you are building in public. Every piece of writing ladders up to one of these positions.
PILLAR 1 · ACTIVE
The Abundance Layer
How cheap code changes everything for ecommerce
“Code went from scarce constraint to abundant enablement layer. The winners have judgment, taste, and distribution. Not the most engineers.”
THE THESIS
Code used to be the bottleneck. Engineers were expensive, queues were long, every feature required a team. LLMs broke that. Software is now cheap. The constraint shifted from engineering to judgment. Ecommerce operators who get this first will run leaner orgs, ship faster, and outcompete teams ten times their size. The winners won't have the most engineers. They'll have the best taste.
KEY CLAIMS
The cost of software creation has dropped by an order of magnitude
Agentic coding tools (Cursor, Replit, Claude Code) mean a non-engineer with domain expertise can ship working software in days. Tasks that took a sprint now take an afternoon.
Ecommerce is behind the curve on this shift
Most ecommerce orgs still gatekeep software through traditional engineering queues — built for a world where code was expensive. The mental model hasn't updated with the economics.
Operator-led software is the new competitive moat
When the person closest to the customer problem can also ship the solution, the iteration loop collapses from weeks to hours. That speed is compounding and un-copyable.
Organizational structures need to be rebuilt around this reality
The traditional dev team → product manager → stakeholder chain was designed for scarcity. In an abundance model, you need fewer people with more judgment — not more people with more specialization.
The new scarce resource is judgment, focus, and distribution
When everyone can build, the differentiator is knowing what to build, being disciplined about what not to build, and knowing how to get it in front of the right people.
Process documentation and SOPs are now a software problem, not a coordination problem
The old cost of operationalizing a process was high on both ends: engineers to build the tooling, and operators spending hours clicking through slow, generic platforms not built for their workflow. When engineering is cheap, you stop trying to make people fit the software — you build software that fits how people actually work. SOPs become living tools instead of PDFs no one reads.
PLANNED ARTIFACTS
No pieces linked yet.
Go to Studio →PILLAR 2 · ACTIVE
The Agentic Operator Premium
Why the AI-native ecommerce leader is worth more than the title suggests
“One AI-native director can run a $10M P&L with the leverage of a team. Benchmark comp against the headcount it replaces, not the title it holds.”
THE THESIS
Director comp is benchmarked against other directors. That's the wrong benchmark. An AI-native ecommerce director doesn't do the work of one person. They do the work of a director, an analyst, a coordinator, a researcher, and an ops lead. The org gets that leverage without staffing it. The right benchmark isn't what the title pays in the market. It's what the displaced headcount would have cost.
KEY CLAIMS
Headcount is no longer the unit of scale in ecommerce leadership
The traditional model assumes revenue growth requires proportional team growth. An AI-native operator compresses that ratio — running financials, R&D, operations, and agency oversight without a specialist layer underneath.
One director can run workflows that previously required 3-5 FTEs
Price-pack architecture and gross margin modeling. Amazon search volume analysis for product R&D. Customer support ticket synthesis. Agency performance tracking. Competitive review analysis. These are not tasks that happen sequentially across a team — they happen in parallel, in one day, from one person with the right tools.
Range is more valuable than specialization in an AI-native org
The traditional career path rewards depth — become the best at one function. Agentic tools commoditize depth. What remains scarce is the person who can move fluidly across finance, creative, data, and operations and know what to do with what they find. That's range, and it's what gets amplified.
Compensation should benchmark against displaced headcount, not comparable titles
If the role would have required a director plus two analysts plus an ops coordinator, the compensation envelope should reflect that total cost — not what a director title commands in isolation. The hire is cheaper than the team and more effective. That gap is where the premium lives.
Taste and judgment are the non-automatable layer — and they are what the market will chronically underprice
AI handles synthesis. The human provides the call: which margin trade-off to make, which R&D signal to act on, which agency relationship to restructure. Judgment at speed — across domains — is the actual product being sold. Most comp systems have no vocabulary for pricing it yet.
PLANNED ARTIFACTS
WRITING THAT LADDERS UP (1)
PUBLISHING POSTURE
Publish with optimism for operators, sensitivity to workers. The thesis is not “AI will take your job.” It is “the operators who understand this shift will build differently — and that creates opportunity for people who develop judgment, not just execution skills.” Acknowledge the fear. Write through it with specificity.
3 DOCUMENTS NOT YET LINKED TO A PILLAR
Assign each doc to a pillar from the Studio page.