> The contention is "(jerk present) → ¬ (success)"

The contention in the article is not "you can't ever succeed if you have a jerk on the team" or "a jerk will always cause your team to fail". A property doesn't have to be universal to be common enough to be worth writing about.

So no, a single instance of succeeding despite a jerk does not "disprove" anything relevant here.

> To me, the business of software development is way too complex to apply a simple "no jerks" policy. Brilliant jerks are often to be found at the beating heart of successful software.

That's what you replied to with "but survivorship bias!". And the article itself says:

> A “no jerks” policy must be preached and practiced from the highest levels.

... which is a universal and normative call to action.

English isn't formal logic, and treating it as such creates a strawman. An article recommending a "no jerks" policy is not a formal theorem that "jerks universally lead to failure", nor can it be refuted and dismissed by an anecdote of "but we had jerks and didn't fail".

If jerks universally led to failure, it would hardly need writing about more than once, as everyone would very quickly have to understand and deal with it. However, because you can sometime succeed in spite of the presence of jerks, that makes it a problem particularly worth writing about.