[flagged] means users flagged it.

Since a lot of people are disagreeing about that and we always try to take guidance from the community, I've turned the flags off on this submission and reduced the software penalty formerly known as the flamewar detector. But FWIW this goes against my better judgment—this thread isn't as civil or substantive as several of the recent ones on related issues have been (I mean in relative terms—they still weren't that great overall) and I super don't like the negative focus on one individual. That's not what substantive discussions are made of.

It mightn't be as substantive or civil, but I've learned more from this HN thread than most.

Reading some of the comments here confirmed for me that this woman is on to something. People are claiming that there's more to the story, or that she must be a bitch/bad at social interactions (with no evidence worth mentioning and despite a respectful and genuinely empathetic writing style).

In fairness to GitHub, there may indeed be more to the story. However, I think if the tech community didn't have the issues she's talking about, the comment section here would be very different. Nothing she said sounds remotely unbelievable, nor is she selling us something, but people here read that she was forced into a mental health institution and felt free to jump to wild conclusions and publicly speculate on her job performance.

I'd have liked to see more discussion of BI, merit strategies, ideas on humanising large organisations, commiseration, maybe job offers and suggestions. And in fairness to this thread, there is a bit of that too. But experience has shown me that when people react this defensively to polite and well though out criticism, there is a measure of projection / delusion going on.

Now, I'm not saying that all non-substantive discussions should always be deemed kosher, but I feel like it's worth pointing out the irony here - the poor quality discussion has verified the quality of the OP.

When I see people drawing conclusions from one data point and I know they could just as easily draw the opposite from a different data point, I start to feel queasy. What I've learned from working on HN is these phenomena are complex, with many counterintuitive effects, and people are much too quick to make statements that fit a narrative.

It's not that your observations are wrong, it's that there are many other observations you'd also need to account for before drawing a conclusion about "the comment section here". Few people have the patience to do this, especially not when the data start to go outside a narrative that seems to make sense. (Where by few I mean statistically no one, including me.) Actually it's the other way around: nearly all of us start with a narrative and then see the community through it. This feels obviously right and there's plenty of data you can point to to support (a.k.a. 'prove') it; the trouble is that there's also plenty of data you can point to to 'prove' the opposite—and people do. What's really going on is that the community is large and complex enough to 'prove' any view you bring to it, and so everyone is simply reading back out what they just encoded in, by weighting some things differently than others. (Given enough things to weight you can encode anything.) This is why people so adamantly say such contradictory things about HN, and I'm sure that happens even more with still larger systems.

IMO this is also why there's a large measure of what you call 'projection/delusion' going on on every side of the internet wars. I'm not sure how to fix that but I'm pretty sure it can't be fixed by the triumph of the good side over the bad side, however you draw that line. The most interesting thing that has happened on HN recently has been the appearance of new forms of dialogue across those lines. Those threads still aren't great but at least it's a sign of something new that isn't just predestined conflagration. Another welcome sign is that in some of these contexts, more women have been commenting than usually do. Those are the kind of developments we're looking to support on HN, in the hope that more interesting and substantive forms of discourse can emerge here. Flamewars are certainly not that.

When we make moderation calls that don't seem to make sense, that's probably why. We're trying to optimize for one thing long-term: intellectual curiosity and substantive online discussion, a problem that turns out to have many counterintuitive aspects. But we're not trying to impose a regime that the community doesn't want and are happy to reverse decisions we get wrong along the way.

I think in relative respects, there needs to be more thought to the flagging system in general. A thread shouldn't be penalized if flagged a lot -- it should be brought to the attention of the moderation staff and then decided upon. These should all be transparent notes, e.g., [flag submitted], [flag rejected], [flag accepted].

On the content of this story in particular: it's not a win/win. But there's an inherent risk in becoming too insular by letting a majority of users decide the outcome on posts the are about minorities. The flag button looks a lot more like a "censor this" button or a "down vote" button than it does a "this doesn't fit the guidelines" button. Maybe that's the goal, but if it is, it should be made clear that's what it is.

It might be useful to distinguish between moderator and user actions. What about "[user flagged]"? Given that moderators do have powers, it's easy to be confused about this.

The moderation is plenty transparent and if you just pay attention to what mods say, you figure out how things work around here. They are happy to explain lots of things. They just don't like to moderate in a manner that makes the moderation overly center stage. I consider that to be a feature, not a bug.