←back to thread

PG: The biggest source of stress for me at YC was running HN

(twitter.com)
482 points ilamont | 2 comments | 11 Jul 20 21:53 UTC | HN request time: 0.503s | source
Show context
_bxg1 ◴[11 Jul 20 23:14 UTC] No.23807033[source]▶
I honestly think the only solution is for individuals to recuse themselves from those networks (I say on one of those networks), lower the trust they place in digital information, etc. It's become clear that the downward spiral is intrinsic to the medium itself (or possibly just the scale). I don't believe that any amount of technology, or product-rethinking, or UX will change that. We just weren't meant to interact this way. My only hope is that people eventually get disenchanted or burned-out enough that they simply stop engaging.

I replied to the original tweet too ("what would you do if you were Jack Dorsey?"). I said I'd shut the whole thing down.

replies(3): >>23807161 #>>23813894 #>>23815451 #
asah ◴[11 Jul 20 23:33 UTC] No.23807161[source]▶
Sadly, the level headed people recuse themselves which only adds to the toxicity.
replies(1): >>23807309 #
newacct583 ◴[11 Jul 20 23:58 UTC] No.23807309[source]▶
Actually what happens is the level headed people on one side of an issue divide recuse themselves, leaving a "seemingly level-headed consensus echo chamber" behind. IMHO, that's worse. This account exists largely to counter exactly that trend. It's important (to me) that newcomers to the site don't get the idea that "hackers" are all fringe libertarians on every non-technical subject.
replies(3): >>23807375 #>>23808300 #>>23810001 #
dang ◴[12 Jul 20 00:07 UTC] No.23807375[source]▶
This site may feel like a "consensus echo chamber" but in reality it is nothing remotely close to that. I think you may be running into the notice-dislike bias: https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que.... Since you report noticing fringe libertarians, we can be sure that you dislike fringe libertarianism. We can also be sure that they have just the opposite picture of HN, since everyone crafts their picture in the image of what they dislike, without realizing that they're doing that. It just feels like an objective picture. I can list dozens of examples of this, but I'll restrain myself for once and spare you.

Unfortunately, these extremely contradictory subjective images of HN seem to be a consequence of its structure, being non-siloed: https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que.... This creates a paradox where precisely because the site is less divisive it feels more divisive—in the sense that it feels to people like it is dominated by their enemies, whoever their enemies may be. That's extremely bad for community, and I don't know what to do about it, other than post a version of this comment every time it comes up.

Thanks for caring about level-headeness, in any case.

replies(5): >>23807422 #>>23807590 #>>23808045 #>>23808650 #>>23809548 #
lefstathiou ◴[12 Jul 20 00:42 UTC] No.23807590[source]▶
Dang, I respectfully disagree.

Just the other day I noticed HN take a heavy hand on removing an article that hit the homepage about a virologist publishing a paper that suggested the only logical explanation for COVID is that it’s manufactured.

There are absolutely topics and perspectives that are not welcomed on HN, as the lead moderator it would be unwise in my opinion to think otherwise (given your biases would be the most threatening to an open forum) and you naturally would have a tough time identifying the absence of a perspective you don’t share.

As an example, I would challenge you to pick five articles that discuss unions that hit the homepage on HN and see what % of threads (and how much of the up vote they accounted for) were inherently anti union. I would also be sure to give only partial credit for threads that added boiler plate sentences saying something along the lines of “while I believe in the value of XYZ” because that’s basically a requirement to take any contrarian (to liberal / Silicon Valley ideology) or conservative view on this site. I can give you a laundry list of topics that will show this trend.

From my (biased) perspective (and from someone outside of the valley reading this site religiously for 13 years) HN is increasingly hostile to certain perspectives (and I’m not talking about social issues here). I don’t care much about it - I just opt out - which is the point.

Why not run a poll about it?

replies(3): >>23807603 #>>23807798 #>>23813620 #
dang ◴[12 Jul 20 01:20 UTC] No.23807798[source]▶
It seems to me that you're confirming my point by making a strong generalization about HN based on what you noticed and disliked. I assure you that the people who dislike the opposite things notice the opposite things and make the opposite strong generalization.

If you're talking about the covid submissions that gus_massa came up with, they were flagged by users. In one case we lessened the penalty and the other looks like one we didn't see. I do think that it's unlikely that HN can have a curious conversation about that theme, much as we might both prefer otherwise. I don't think you can validly draw significant general conclusions from that.

A poll wouldn't convince anybody. It would just reconstitute the same disagreement at a meta level. I wrote about something similar here: https://news.ycombinator.com/item?id=23239793. (Edit: and https://news.ycombinator.com/item?id=23808089 in this thread, as it turns out.)

replies(1): >>23807875 #
newacct583 ◴[12 Jul 20 01:32 UTC] No.23807875[source]▶
For the record: I rather think that comment validates my view. This is what you're hearing: "This guy is complaining that you took down a winger conspiracy theory that got upvoted, therefore HN must be balanced."

Here's what I hear: "A winger conspiracy theory goes to the top of the HN front page before being taken down! That means that the voting population of HN is horrifically skewed."

See the difference in perspective? I'm happy you take conspiracy stuff down, I really am. But I'm not happy about the population that pushes it finding a home here, which they clearly have.

replies(2): >>23808015 #>>23808595 #
dang ◴[12 Jul 20 01:55 UTC] No.23808015[source]▶
That's not what I heard.

You're drawing extremely skewed conclusions about the "voting population" of HN. One of those posts made it to #16 before being flagged, the other did not make the front page at all. It takes only a handful of votes to make the front page (much of the time, anyhow—it's complicated), and #16 is not high—any sensational story can easily get that far before being flagged down (and by the way, it was users who flagged it down, not us).

HN is a large enough population sample that you'll find that scale of upvoters upvoting anything, some of the time. You can't conclude anything significant about the "voting population" of HN from that, and the fact that you're doing so strikes me as an indication of what I'm arguing—that your generalizations about HN are determined by your own ideological priors, just as people with opposite ideological commitments arrive at the opposite generalizations, and by exactly the same mechanism.

I think you'd arrive at the same conclusion that I have, if you were forced as I have been to look at all sides of this under unrelenting personal pressure. I also think that I'd be arriving at your conclusion if I hadn't been forced to have this experience. That's a pessimistic conclusion—it means that rational discussion of this is probably not possible. (I mean that structurally—not a personal swipe but exactly the opposite, and I hope that's clear.)

replies(2): >>23808172 #>>23808382 #
nkurz ◴[12 Jul 20 02:56 UTC] No.23808382[source]▶
> One of those posts made it to #16 before being flagged, the other did not make the front page at all.

Well, presuming we're talking about the minervanett.no article, there's also this submission that made it to #2 with 24 votes and 18 comments in the 10 minutes before it was flagged to death:

[flagged] The most logical explanation is that it comes from a laboratory (minervanett.no)

https://news.ycombinator.com/item?id=23738545

http://hnrankings.info/23738545/

Then there was also a version a week earlier that got to #3 in 5 minutes before it was killed:

[flagged] [dead] Norwegian virologists suggest Coronavirus originated in a laboratory (minervanett.no)

https://news.ycombinator.com/item?id=23738264

http://hnrankings.info/23738264/

> You'll find that scale of upvoters upvoting anything, some of the time.

I found that to be a surprisingly high number of votes in a short period of time, likely indicating that there is a substantial population on HN who would have liked to discuss that story but were prevented from doing so by a (presumably) much smaller number of flags. Is it actually common for stories without a broad degree of interest to get that many votes in their first few minutes?

> it was users who flagged it down, not us

Technically, but I'm hoping that you try to review the flagged stories and recover the ones where you disagree with the flagging? Otherwise this would seem to allow a "tyranny of the minority" where a small number of people who want to prevent a discussion from taking place are able to enforce their beliefs on the rest of the larger group. I'm a lot more comfortable with you using your expert judgement than with trusting that flags will always be used appropriately without review.

replies(1): >>23808425 #
dang ◴[12 Jul 20 03:01 UTC] No.23808425[source]▶
Ah, that was the one I remember seeing. In such cases, if (a) article is substantive and (b) there is a chance of an intellectually curious discussion, we turn off the flags and (probably also) downweight the article as a counterweight to sensationalism, which always attracts extra upvotes. In this case (b) seemed hopeless, so I didn't.
replies(3): >>23808502 #>>23809072 #>>23812495 #
mistermann ◴[12 Jul 20 16:27 UTC] No.23812495[source]▶
> In such cases, if (a) article is substantive and (b) there is a chance of an intellectually curious discussion, we turn off the flags and (probably also) downweight the article as a counterweight to sensationalism, which always attracts extra upvotes. In this case (b) seemed hopeless, so I didn't.

That's the point, in my opinion at least.

Above you say:

> This site may feel like a "consensus echo chamber" but in reality it is nothing remotely close to that. I think you may be running into the notice-dislike bias...

If a certain class of articles get flagged by a large number of people who have a strong dislike for the topic, and you as the moderator are ok with it because here at HN with the culture being the way it is you allow it to be removed because it won't generate "intellectually curious discussion", then how can you say that HN is not a "consensus echo chamber" when it comes to these particular topics?

It seems to me there are some very obvious errors in your explanation above. You can run this forum however you want, but being aware and transparent about what topics are and are not allowed seems like a better way to do it than disingenuously explaining away flaws. No person is omniscient, however it may seem that way. Just as you can observe flaws in other people evaluation that they themselves cannot see, is it not possible that you too may have some flaws that you cannot see?

replies(1): >>23813652 #
1. dang ◴[12 Jul 20 18:41 UTC] No.23813652[source]▶
It's more complex than just "what topics are and are not allowed". Threads are extremely sensitive to initial conditions. One thread might discuss a topic within the site guidelines while another thread on the same topic might turn into a massive flamewar. Subtle (or not so subtle) differences in article, headline, URL, site design, and who knows what else can make the difference too. The decision I made was based on the thread, not the topic. As I mentioned above, we reduced the penalty on one of those threads. We wouldn't do that if the topic were "not allowed".

Is it possible that I may have some flaws that I cannot see? That is beyond possible, it is certain. The trouble with these arguments though is that operating this place is a lot more complicated than people assume it is, and so they say oversimplified things like "aha, you are suppressing topic X so HN is an echo chamber after all" and I have to try to fill in the information gap before we can have a sensible conversation about what the actual flaws might be. I'm super interested in the flaws—but first we have to be talking about the same world, which unfortunately is already not so easy.

replies(1): >>23835756 #
2. mistermann ◴[14 Jul 20 18:27 UTC] No.23835756[source]▶
I am well aware that it is complicated. When a controversial thread is reported, a rather large number of variables are referenced by your mind - some of these you are aware of, some of them you are not.

But at the end of the day, in the aggregate, either there is zero slant (by topic) whatsoever, or there is greater than zero. Based on my anecdotal observations over a long period of time, my perception is that there are indeed certain topics that are less welcome than others, and the assurances I've read, while plausible, do not seem adequate. If we were able to see a log of removed topics it may be more reassuring.

I'd rather HN had more freedom of topic discussion at least occasionally as an experiment, and then perhaps we could see if some modifications to guidelines (perhaps just on those threads) could keep things a bit more civilized. If no site is willing to put some effort into finding a workable approach to this problem, it seems reasonable that the world is just going to keep becoming more polarized as people spend more time at sites that are designed from scratch to be information bubbles.