The Six Irreducible Truths for Hard Conversations

Truth 1: AI Transforms Bottlenecks, Not Headcount Requirements. When a machine eliminates one bottleneck, the scarce resource shifts to what the machine cannot do—judgment, integration, and systems thinking. Reducing humans without redesigning roles around the new bottleneck produces invisible technical debt that compounds until it collapses.

Truth 2: Team Size Is a Resilience Parameter, Not a Cost Line. A single point of failure is not a team, regardless of how it is drawn on an org chart. Sustainable velocity requires knowledge distribution, which requires minimum critical mass. The cost of a team too small is captured in attrition, incident duration, and architectural fragmentation—not the quarterly budget.

Truth 3: Strategy Is What You Choose Not to Do. A strategy without a "not doing" list is a backlog. A backlog is not a strategy. Every organization evaluates the same tools; the ones that win make explicit choices about which problems to ignore.

Truth 4: Speed Is a Subtraction Problem, Not an Addition Problem. Velocity is constrained by validation, not production. Adding production capacity without adding validation capacity increases throughput of unvalidated work, which increases rework, incidents, and operational load. True speed requires stopping work, not starting more.

Truth 5: Machine-Generated Code Requires More Governance, Not Less. Code written by a system with no intent, no accountability, and no explanatory capability cannot be reviewed with the same assumptions as human-written code. The faster the code arrives, the more rigorous the validation must become.

Truth 6: Technical Expertise Without Translation Is Ineffectiveness. A technology leader who cannot translate technical reality into the language the room already thinks in—revenue, risk, retention, runway—is not unheard because the room is unintelligent. They are unheard because they are speaking the wrong language.