[tzs]

...and why can we still reply to comments on a flagged article? Also, normally I have a "vouch" option on flagged submissions, but not on this one.

Something is weird here. I wonder if for some reason the submitted actually included "[flagged]" in the title and so it just looks like a flagged article?

[dang]

Recent software changes. The system used to say [flagged] only when the story was both heavily flagged and dead (closed to new comments). Now it says [flagged] if the story is heavily flagged, dead or not. Flagged-but-not-dead was the case here. That's why you didn't see a 'vouch' link, btw; there's no need to vouch for posts that aren't dead, since the purpose of vouching is to unkill them.

The reason we did this: stories that have a lot of flags but also a lot of comments don't get automatically killed by the software. This led to confusion and a lot of "why is this post not on the front page" questions. Displaying [flagged] in the title in such cases should answer, and thus preempt, most such questions. Plus HN users like it when they get more information, which this is.

Similarly, we're now displaying [dupe] on stories that are marked as duplicates but not dead.

[zeno00]

Hah, I definitely didn't add 'flagged' to the submission. Looks like it disappeared off the front page too which sucks. I emailed the mods(?) so maybe my poor post will receive clemency.

[dang]

(Replying here as well as via email in case other users want to know.)

HN is for stories that gratify intellectual curiosity, which this one arguably doesn't. Plus we have rules against people using HN threads for job/hiring/seeing-work posts (except in monthly threads designated for that). Those are probably the reasons why users flagged your submission.

On the other hand, it's a borderline call, the discussion is reasonably good, and the Hawaii angle is novel, so we'll cut you some slack and turn off the flags.

[nekopa]

By the way dang, I just want to say thanks to the mods for still keeping a human hand in things.

I do understand that a site that has this volume of interaction needs automated moderation. But it's nice to see people step in and override the algos when necessary.

As a side note to people talking about AI eating the world: how could we train an AI to make these kinds of judgment calls? (This coming from me as a huge AI fan, I am really interested in possible solutions)

P.S. While I'm here, I did have to chuckle about Alan Kay's stackoverflow question about progress in CS being closed as not suitable for SO. Sometimes even human algorithms fail :)

[dang]

Thanks! Btw, if people want to make sure that we see their question they should email [e-mail]. Appeals to mods from HN comments are hit-and-miss because there's no way for us to read all the comments.

We haven't looked much into AI-style algorithmic approaches for HN moderation; we will eventually. But we're also interested in figuring out how to decentralize more moderation to the community, and what software we can build to support that.

I'm delighted to confirm that Alan's questions (and better still his answers) will always be "suitable" for Hacker News. They practically define suitable! https://news.ycombinator.com/posts?id=alankay1