Andy Burns wrote:

The original article makes the very crude leap from one filament with a
claimed 0.0001% probability of failure

Grrrr! 0.01% in both cases