·3 min read·startup-validation · adversarial-review · founders · decision-making

Steelman your own idea before you kill it

An opponent that attacks a weak version of your idea is worthless, and so is one that defends a weak version of the objection. The discipline that makes adversarial review honest is the steelman.

Contents

Testing your own idea by arguing against it sounds rigorous. It usually is not, because of a trap that is easy to fall into and hard to notice: you attack a weak version of your idea, defeat it, and walk away feeling like you did the work. Or you defend against a weak version of the objection, dismiss it, and feel safe. Both are theater. Neither tests anything.

The fix has a name. Before you try to kill your idea, you build the strongest possible version of it, and the strongest possible version of every objection to it. Then you let those two fight. That is steelmanning, and it is the quality bar the whole adversarial approach depends on.

#The rule, stated precisely

The discipline comes from the philosopher Daniel Dennett, adapting rules first formulated by the game theorist Anatol Rapoport. The first and hardest rule is this, verbatim:

Only after you have done that, Dennett says, do you earn the right to criticize. Applied to your own idea, it cuts both directions. State your idea's best case so well a true believer would nod. Then state the strongest objection so well a smart skeptic would nod. If your critique only defeats a cartoon of your idea, or your defense only swats a cartoon of the objection, you have learned nothing except that you are good at building cartoons.

#Why a real opponent beats a fake one

There is evidence that this is not just aesthetics. The psychologist Charlan Nemeth ran experiments comparing a genuine, authentic dissenter against several forms of assigned, role-played devil's advocacy. The authentic dissent won. A person merely told to "argue the other side" often left the group more confident in its original position, not less, because everyone could feel the objection was not real. A token opponent, or a strawman you built yourself, produces the comfortable sensation of having been challenged without the substance of it.

That is the danger for a founder running a solo pre-mortem. It is very easy to stage a debate you were always going to win. The steelman is the discipline that stops you: it forces the objection to be strong enough that beating it actually means something, and honest enough that sometimes it beats you.

#The move

So before you decide your idea survives, do the harder version. Write the single strongest reason it works, the one you would put in front of the most skeptical investor you know. Then write the single strongest reason it fails, the one that same investor would lead with. Put them side by side. If the failure case is a strawman, you have not tested your idea; you have flattered it with extra steps. If it is a steelman and your idea still stands, you have something worth building. And if the steelman wins, that is not a bad day. That is the check doing its job while it is still cheap.

An opponent that only beats strawmen is theater. The one worth having engages the strongest version of what it is attacking, including when that version is yours.

For the full approach: you don't need another validator, you need an opponent, and the Founder's Attack-Surface Checklist to run it on your own idea.

#Sources

Keep reading