I don't understand your point here... Active users being disproportionately represented by people commenting is probably correct, but it doesn't provide any useful information about the minority of users who don't comment often.

I'm saying that an account being mostly inactive (In terms of commenting) is not at all surprising.

Someone could have been actively browsing HN for months/years without commenting, so I don't think comment activity is a good indicator of credibility when lurking is the default behaviour for almost all users.

I wouldn't be surprised if the number of comments per user followed a power law distribution.

He's not saying its an indicator of credibility. He's speaking to the probability of two infrequent commenters commenting. The density of frequent commenters in all comments is very high.