Qurated: Introducing Plan A
The Best Plan Beats the Best Critique
Most smart people are excellent at explaining why a plan won't work. Almost none can articulate what would work instead. Scott Alexander's "Plan A" makes a simple, radical move: write down an actual plan — flawed, incomplete, falsifiable — and let it absorb criticism rather than hiding from it. This isn't just an AI safety essay. It's a template for how serious thinking should be done in any domain where the stakes are high and the discourse has calcified into pure critique.
The Critique Trap
Criticism is cheap because it's unfalsifiable in practice — you can always find a flaw, and finding flaws feels like rigor. But an ecosystem of pure critics produces paralysis, not progress. Nobody has to defend a coherent alternative, so nobody does. The intellectual incentive gradient rewards being the smartest person in the room who didn't commit to anything.
Mental model: Critique is a tax on Plans, not a substitute for them. If you've never proposed a Plan A of your own, your critique of someone else's plan should be weighted accordingly.
Why "A Plan," Not "The Plan"
Calling it "Plan A" — not "The Solution" — is doing quiet, important work. It signals:
- Provisional confidence. Committed enough to defend, humble enough to revise.
- Room for Plan B. Naming it invites successors instead of demanding permanence.
- A target for falsification. Vague hand-wringing can't be tested. A named plan can.
This is the difference between a hypothesis and a mood.
The Plan A Test
Before you criticize a strategy — in AI governance, in your career, in a relationship, in a company roadmap — ask:
- Can I state an alternative in one paragraph?
- Would I bet resources (time, money, reputation) on my alternative?
- Am I willing to have my alternative picked apart in public?
If the answer to any is no, your critique is commentary, not contribution. Useful, maybe — but not a substitute for someone else doing the harder work of committing.
Coordination Requires a Shared Object
A key insight buried in this piece: disagreement is often not about values but about the absence of a concrete artifact to disagree about. Vague fears and vague hopes don't converge. A written plan — with assumptions, mechanisms, and failure modes spelled out — gives disparate factions something to actually argue about, which is the precondition for argument to actually resolve anything.
Framework: Specificity is the antidote to tribalism. The less concrete a proposal, the more debate collapses into identity signaling. The more concrete, the more it collapses into facts.
The Takeaway
Doom-scrolling through problems without a stated Plan A is intellectually comfortable and practically useless. The braver, harder, more valuable act is naming your best guess, attaching your name to it, and inviting the world to prove you wrong faster than you'd have found out alone.
Question for the comments: What's your Plan A — for your work, your field, your life — that you've never actually written down because you're afraid of how it'll hold up?