Distinguish steelmanning from sycophancy
Charitably interpreting a position is not the same as agreeing with it, refusing to disagree, or softening your critique to avoid conflict.
Key takeaways
- What it is: Charitably interpreting a position is not the same as agreeing with it, refusing to disagree, or softening your critique to avoid conflict.
- Why it works: Steelmanning requires intellectual courage: you must construct the strongest version of a view and then honestly engage with it — which means saying clearly when even the strongest version fails. Sycophancy produces the appearance of charitable engagement without the substance: the "steelman" is constructed to justify agreement, not to find the strongest challenge. The distinction is that a genuine steelman increases the probability of updating your own view; a sycophantic one is designed to preserve it.
- Evidence: Practitioner experience / self-report; no formal study.
- Avoid: Concluding after each steelmanning exercise that your position was right all along, without tracking whether you ever actually update — a consistent pattern of no-update is strong evidence of sycophantic steelmanning.
Why it works
Steelmanning requires intellectual courage: you must construct the strongest version of a view and then honestly engage with it — which means saying clearly when even the strongest version fails. Sycophancy produces the appearance of charitable engagement without the substance: the "steelman" is constructed to justify agreement, not to find the strongest challenge. The distinction is that a genuine steelman increases the probability of updating your own view; a sycophantic one is designed to preserve it.
How to do it
- 1After constructing a steelman, ask: "Has this steelman actually changed anything I believe or added anything I hadn’t considered?"
- 2If the answer is consistently no, you are likely constructing steelmans designed to fail.
- 3If the steelman reveals a genuine problem with your position, revise before continuing.
- 4Track whether your positions ever actually update after steelmanning exercises — that is the calibration signal.
What the evidence says
AnecdotalThe sycophancy failure mode is documented in AI systems and is a recognized risk in human social reasoning — social pressure creates a systematic incentive to validate rather than challenge. The distinguishing criterion here (does the steelman produce genuine updates?) is a practitioner diagnostic rather than an experimentally validated technique.
Honest caveat: Deciding whether an update is "genuine" or motivated is itself subject to motivated reasoning. External accountability partners can help calibrate.
Common mistake
Concluding after each steelmanning exercise that your position was right all along, without tracking whether you ever actually update — a consistent pattern of no-update is strong evidence of sycophantic steelmanning.
IX Coach tracks whether your positions shift over time in response to challenges, distinguishing genuine open-mindedness from the performance of it.
Practice this with IX Coach →