Skipping to the Bad Part
When the plan delivers the thing it was supposed to prevent
Some strategies are built so their own mechanism produces the outcome they were meant to prevent, and a comforting label hides it. That label is a revealed priority: stopping the failure was never the real goal.
The “herd immunity strategy” was never a strategy for avoiding mass infection. It was mass infection, with a nicer name.
The stated goal was to stop people getting sick and dying. The plan was to let everyone get sick until enough were immune. The thing you were trying to avoid was step one of the plan to avoid it. And nobody called it the “mass infection strategy.” They named it after the destination you’d supposedly arrive at, not the road you’d take, and the name did the rest. It made a prescription to let people get infected sound like a form of protection.
I’ve started calling this move skipping to the bad part.
The Move
The pattern goes like this. You have a goal: prevent X. Some protection exists to prevent X. Someone points out the protection is leaky, it’s not perfect, it can be beaten, it has known failure modes. And instead of asking the sane question (how leaky, and at what cost?), you redesign the “protection” so that it now produces X. The failure stops being a risk the plan guards against and becomes a step the plan delivers.
The insidious part is that it still looks protective. It keeps the vocabulary of prevention while doing the opposite. That’s what separates it from honest surrender.
Not Just Perfectionism
This is not “the perfect is the enemy of the good.” That one’s old, and it’s passive: seatbelts aren’t 100%, so don’t wear one; masks leak, so don’t bother. Call it the nirvana fallacy, call it perfectionism; the point is it already has names, and it leaves you merely unprotected. You didn’t build a machine for producing the harm. You just stopped guarding against it.
Skipping to the bad part is the active version. You don’t drop the plan. You keep the plan, and you wire the failure into it as a load-bearing step. Then you name the whole contraption after the outcome you wish it produced.
A Fresh One
Here’s a current example. There’s a serious genre of writing right now about how the middle powers (the UK, Europe, Japan, Korea) avoid getting steamrolled when one country reaches superintelligence first. One proposal: if the US ever cuts you off from frontier AI, you’ll need a plan B, and the only other game in town is China. So build the relationship now. Be ready to pivot.
Look at what the plan protects against: being dominated by whoever controls superintelligence. Now look at the plan B: ally yourself with the other country racing to control superintelligence. By the argument’s own logic (winner takes all, no incentive to give the little guy a fair deal), China running the show ends you precisely the way a dominant US would. The escape hatch opens onto the same room.
Widen the Goal
It only reads as protection if you quietly swap the goal. State it as “avoid being dominated by the US,” and pivoting to China looks like a lifeline. State it as “avoid being dominated,” which is the goal you actually have, and the lifeline is just a different rope around the same neck.
That’s the general tell, and it’s worth keeping in your pocket: a skipping-to-the-bad-part plan only looks protective under a goal that’s been narrowed just enough to let the failure slip outside it. Widen the goal back to what you actually care about, and there’s the plan’s own mechanism, sitting in plain view, producing the thing you were trying to prevent.
(To be fair to the proposal, it frames China as leverage you’d never really use, a bluff to extract a better deal from the US. But a bluff only has teeth if you’d actually pull it, and pulling it walks you straight into the bad part. The leverage is exactly as real as your willingness to lose.)
The Relabel Is the Tell
Here’s the part that earns this a place on a blog about revealed priorities.
When a plan’s mechanism is the very failure it claims to prevent, and the only thing holding the illusion together is its label, that label is a revealed priority.
Think about what the relabel is for. “Herd immunity strategy” instead of “mass infection strategy” lets you skip the expensive, unpopular work (locking down, building capacity, actually intervening) while still sounding like you have a plan. The stated priority was protecting people. The revealed priority was protecting the appearance of a plan, at the lowest possible cost.
Run the same read on anything. A company ships anonymisation it knows is reversible and calls the dataset “anonymised”: stated priority, privacy; revealed priority, ship the data while being able to deny you did. Reviewers can already guess who wrote the paper, so the journal keeps calling single-blind review “blind”: stated priority, impartiality; revealed priority, change nothing.
The euphemism is where the two priorities come apart. Whenever you catch yourself (or a policy, or a country) reaching for a comforting name for the step in the plan that is the failure, that’s the seam. The stated goal is printed on the label. The real one is in the mechanism.
Read the Mechanism
So the next time you meet a strategy named after its happy ending, do one thing: ignore the label and read the mechanism. Ask what the plan actually produces, under the goal you actually hold. If those two things turn out to be the same, if the plan manufactures the outcome it was sold as preventing, you’re not looking at a protection that fell short.
You’re looking at someone who already decided to skip to the bad part, and would rather you didn’t read the fine print.