Like most everything when it comes to enterprise/work, these approaches are complementary. Need both the team-level dynamics/rules and the organizational norms/guardrails running in tandem to actually reach ideal work-life balance. We live in an interesting world and I love it :)

Love the paper & top-down "AI practice" as corporate norms, but feels like would take long (months, years?) to implement, while team organizing/deciding how to best work AI-first together can start tomorrow (which, I assume, close to what already happens at startups organically)

Now we just need to go out and evangelize these practices - not in startups (diff dynamics, diff work-life philosophies, if they even want it at all, small teams, figured this out naturally already?). But enterprise, where people are most likely to burn out alone :)

And alone, you do every task - prompting, reviewing, correcting, context-switching across everything. Within a (even small) team, can specialize/split tasks/loads so no one person is overwhelmed by all of it. Not doing more, but doing enough - together.

Alone, you carry every judgment call yourself - AI helps execute but the decisions are all yours. In a team, that weight is shared; someone can say "that's good enough, ship it" before you overpolish into exhaustion.

Alone with AI, there's no natural stopping point - you can always do "one more thing." A team creates social and professional boundaries around your work/workflow.

Keep thinking about this too, and - at least for me - I believe the solution is - ironically, AI-first TEAM work. Like 'going to gym together' effect, but inverted. Burnout comes from working alone with AI — carrying the full cognitive load in isolation until you hit exhaustion.