Don't tolerate assholes — but be strict about what one is

TL;DR

Don’t tolerate assholes. People who demean, belittle, intimidate, and de-energize others poison a team, and protecting them — especially when they’re talented — costs you far more than their output is worth. But apply a strong filter to the word, because it is the most easily weaponised label in engineering, and it is usually aimed at the wrong target. None of these is being an asshole: disagreeing with you, even forcefully and in public; raising new or unconventional ideas; challenging the direction; or opening an alternative PR that solves a problem you’re working on a different way. That’s not an attack — that’s the job. The dividing line is simple: an asshole attacks the person; a good colleague attacks the problem. Remove the ones who attack people; defend the ones who challenge ideas. And never let “you’re being an asshole” become the tool that silences dissent.

Context

Robert Sutton’s no-asshole ruleRobert I. Sutton — The No Asshole RuleRobert Sutton's book (and the HBR article behind it) arguing organizations should refuse to tolerate people who demean and de-energize others. His two tests: after dealing with the person, does the target feel oppressed, humiliated, or de-energized? And does the person aim their venom downward — at people less powerful than themselves?en.wikipedia.org ↗ is right: a team is better off without people who treat others with contempt, and “but they’re brilliant” is not a defense — the brilliant jerk demoralises everyone around them, drives good people out, and teaches the team that cruelty is the price of status. That part isn’t controversial.

The problem is the definition, because “asshole” is the most abusable word in a team’s vocabulary. It gets pointed at exactly the people doing the most valuable, uncomfortable work: the person who disagrees with the senior engineer, who questions the plan everyone had quietly accepted, who shows up with a competing design. Branding that as “being an asshole” is a way to make disagreement feel like misconduct — and it’s frequently deployed by the actual problem: someone with status who finds being challenged intolerable, or a fragile ego that experiences “you’re wrong about this” as an attack on the self. Used loosely, the label punishes the people a healthy team most needs and protects the people it should be removing. A strict definition isn’t pedantry; it’s what keeps the rule from being turned inside out.

Recommendation

Judge by whether someone attacks people or problems — and protect the right to challenge.

What an asshole actually is. Use Sutton’s two tests:

• They demean. After dealing with them, the other person feels belittled, humiliated, smaller, or de-energized. The behaviour targets the person — their worth, intelligence, or identity — not the work: contempt, ridicule, personal insults, condescension, dismissiveness, public humiliation, intimidation, taking credit and assigning blame, the quiet bullying that makes people afraid to speak.
• They punch down. They aim it at people with less power than themselves and behave differently toward those who can hurt them. Punching down is the tell; it separates a jerk from someone who’s just hard on everyone including their own boss.

And it’s a pattern, not a single bad day. Everyone is occasionally short under stress; the asshole is reliably, repeatedly the source of the damage.

What an asshole is NOT — the strong filter. Do not let any of these be called assholery:

• Disagreeing with you — strongly, repeatedly, in public, including with your own work. Dissent is the immune system of good engineering.
• Raising new or unconventional ideas, challenging an assumption everyone had stopped questioning, or telling you something you don’t want to hear.
• Opening a competing PR that addresses a problem you’re working on a different way. That is not an attack on you — it’s engineering, and a team where solutions compete on their merits is healthier than one where the first person to touch a problem owns it. The flash of feeling threatened by a better (or just different) approach is yours to manage; the person who wrote it is doing exactly what you want people to do.
• Being direct, blunt, or holding a high bar on the quality of the work.
• Being quiet, awkward, intense, or socially un-smooth. Not warm is not cruel.

But how you disagree still counts. This cuts both ways. You can absolutely be an asshole while disagreeing — sneering, condescending, making it personal, performing contempt for someone’s intelligence in the guise of “honest feedback.” The disagreement is never the problem; the cruelty wrapped around it is. So “I’m just being direct” doesn’t license belittling someone, and equally, “you’re being an asshole” doesn’t get to mean “you disagreed with me.” Separate the idea conflict (welcome, however blunt) from the interpersonal contempt (never), and judge only the latter.

Operating on it.

• Defend the challengers. When someone is branded difficult mainly for disagreeing or competing on ideas, protect them — and look hard at whether the person doing the branding is the actual problem.
• Remove the real ones, talent regardless. No brilliant jerks. Tolerating a high-performing asshole tells everyone the rules bend for output, and you lose more good people than the jerk could ever produce.
• Make it safe to disagree, which is the same goal as ZFN-8: people raise hard things directly only when doing so isn’t treated as an offense.

Consequences

Easier:

• The team keeps its dissenters, challengers, and competing-idea people — the ones who catch the expensive mistakes — instead of training them into silence.
• Real toxic behaviour gets named and removed without the “but they’re brilliant” escape hatch.
• “Asshole” stops being a cudgel for the thin-skinned and the status-protecting, so the word keeps its meaning.

Harder:

• The line between “blunt and challenging” and “demeaning” takes real judgement, and reasonable people will sometimes disagree about a specific case — that’s a conversation to have honestly, not a reason to abandon either half of the rule.
• Removing a talented asshole has a visible short-term cost (their output) against an invisible long-term gain (everyone they were suppressing), so it takes conviction.
• People who confuse “I felt bad” with “I was mistreated” will push to widen the definition; holding the strict line is ongoing work.
• Self-examination is required: the person most tempted to call a challenger an asshole is often the one who can’t tolerate being challenged.

References

• no-asshole ruleRobert I. Sutton — The No Asshole RuleRobert Sutton's book (and the HBR article behind it) arguing organizations should refuse to tolerate people who demean and de-energize others. His two tests: after dealing with the person, does the target feel oppressed, humiliated, or de-energized? And does the person aim their venom downward — at people less powerful than themselves?en.wikipedia.org ↗ — Robert Sutton’s rule and his two tests (demeaning; punching down) for telling a real asshole from someone who merely disagrees.
• ZFN-8 — the companion: make it safe to raise hard things directly, as yourself.
• The “no brilliant jerks” principle (popularised by Netflix’s culture deck) — talent doesn’t buy a pass on how you treat people.

Changelog

• 2026-06-12: First published as a Field Note.