Skip to content
How to ThinkIn the Age of AI

The zero-based commitment test

Apply the zero-based question to each commitment during scheduled renewal reviews: 'Knowing what I know now, would I start this commitment today?' treating 'probably not, but...' as a signal that qualifiers are rationalizations requiring investigation.

When This Fires

Common Failure Mode

Accepting the hedged answer: "Probably not, but I've invested two years." The two-year investment is irrelevant to the forward-looking question. Sunk costs are sunk. The question is about future value: "Given the next year, is this the best use of my commitment capacity?" If the answer requires a "but" to become yes, the commitment is being sustained by rationalization, not by genuine value.

The Protocol

(1) During scheduled reviews, for each active commitment ask: "Knowing everything I know now, would I start this commitment today?" (2) If the answer is a clear yes → renew. The commitment passes on current merits. (3) If the answer is "probably not, but..." → investigate each qualifier: "But I've invested so much" → sunk cost. Not a forward-looking reason. "But it would be awkward to leave" → social cost. Real, but is it larger than the ongoing commitment cost? "But others depend on me" → investigate: do they genuinely depend on you, or could they adapt? (4) After investigating qualifiers: if no forward-looking reason survives → begin exit process (Exit criteria must be state+date pairs… exit criteria). (5) If a forward-looking reason survives → renew, but document the specific reason. At the next review, check if that reason still holds.