Stefan, thanks for sharing this. The “executive module” / elephant framing landed for me, especially the part about the cage being built from identity and responsibility rather than hours.
Two thoughts I hope are additive:
One underrated failure mode is treating burnout as a rest problem when the real issue may be perceived “trappedness” (identity, indispensability, heroic responsibility, community norms). That gets further locked in by a false A/B: either keep going as-is or quit. Naming that false dichotomy matters, because it blocks the moves that change the equilibrium. A question I’ve found useful: “What’s a third option?” (Not 20 options, just one concrete change.) Often it’s boring: modify one thing that’s draining you (@John Salter mention of the energy audit), renegotiate scope or success metrics, shift roles, get a contingency plan in place. If I can’t generate any 3rd option, that itself is data. (Yet likely a friend can help with options.)
Relatedly, I’ve noticed a head-forward failure mode (in myself too): assuming we’re the exception who can override signals indefinitely. In practice the bill comes due somewhere, and it’s often not output that breaks first. It’s relationships, physical health, or mental health. My take is the people who sustain high-stakes work aren’t the ones who override signals best. They’re the ones who notice drift early and make small course-corrections before arriving at a cliff. I got closer to the cliff edge than I like to admit; migraines were my wake-up call (thankfully now well managed).
I appreciate you sharing this methodology as I am diving in to look at a slice of the AI governance/ AI safety space (current simple framing: AGI preparedness x human capability layer). I’ve been contemplating how to efficiently find what’s out there—likely looking for research, past and current projects, convenings, and grants.