https://en.wikipedia.org/wiki/Tank_Man
https://www.msnbc.com/msnbc/watch/25-years-later-tank-man-st... (Note: Microsoft owns this page!)
https://www.pbs.org/wgbh/frontline/article/remembering-the-t...
https://www.bing.com/images/search?q=tank%20man%20china
(For those who don't see results, this is my screen grab of the above at 18:12 UTC : https://imgur.com/a/3tzPV49)
Also, "Tiananmen Square massacre" still returns some image results:
https://www.bing.com/images/search?q=Tiananmen+Square+massac...
Therefore, the tweak for "tank man" doesn't seem to make any sense since all the other phrases to get to the same sensitive topic are not suppressed.
HN looks like "A Quiet Place"
https://yandex.com/images/search?from=tabbar&text=tank%20man
Here's one tip for you guys, from years-long, world-weary experience: if you're coming up with sensational explanations in breathless excitement, it's almost certainly untrue.
Edit: ok, here's what happened. Users flagged https://news.ycombinator.com/item?id=27394925. When you see [flagged] on a submission, you should assume users flagged it because with rare exceptions, that's always why.
A moderator saw that, but didn't look very closely and thought "yeah that's probably garden-variety controversy/drama" and left the flags on. No moderator saw any of the other posts until I woke up, turned on HN, and—surprise!—saw the latest $outrage.
Software marked https://news.ycombinator.com/item?id=27395028 a dupe for the rather esoteric reasons explained here: https://news.ycombinator.com/item?id=27397622. After that, the current post got upvoted to the front page, where it remains.
In other words, nothing was co-ordinated and the dots weren't connected. This was just the usual stochastic churn that generates HN. Most days it generates the HN you're used to and some days (quite a few days actually) it generates the next outlier, but that's how stochastics work, yes? If you're a boat on a choppy sea, sometimes some waves slosh into the boat. If you're a wiggly graph, sometimes the graph goes above a line.
If I put myself in suspicious shoes, I can come up with objections to the above, but I can also answer them pretty simply: this entire thing was a combo of two data points, one borderline human error [1] and one software false positive. We don't know how to make software that doesn't do false positives and we don't know how to make humans that don't do errors. And we don't know how to make those things not happen at the same time sometimes. This is what imperfect systems do, so it's not clear to me what needs changing. If you think something needs changing, I'm happy to hear it, but please make it obvious how you're not asking for a perfect system, because I'm afraid that's not an option.
[1] I will stick up for my teammate and say that this point is arguable; I might well have made the same call and it's far from obvious that it was the wrong call at the time. But we don't need that for this particular answer, so I'll let that bit go.
https://news.ycombinator.com/item?id=27394925 [Flagged]
https://news.ycombinator.com/item?id=27395028 [Marked as dupe]
https://news.ycombinator.com/item?id=27394943 [Currently on page 2]
https://hn.algolia.com/?dateRange=all&page=0&prefix=false&so...
Some good threads to start with might be https://news.ycombinator.com/item?id=22902490 and https://news.ycombinator.com/item?id=21607844. Also https://news.ycombinator.com/item?id=17014869, which shows how far back political discussion goes on HN, as well as the argument about politics on HN.
If you (or anyone) takes a look at that material and still has a question that hasn't been answered there, I'd like to know what it is. Please just make sure that you've actually familiarized yourself with the past explanations, though, because the odds are good that they do answer the question.
I say that based on looking into literally thousands of such cases, and spending god knows how many hours poring over data. My comments are based exclusively on what I know about HN—I know nothing at all about other sites because I don't have their data. But I know a lot about HN, and I can tell you that the users making breathless insinuations about this stuff have literally no idea what they're talking about. The truth is just painfully boring. (Edit: here's a detailed explanation - >>27397230 .)
As for why the comments in such threads get downvoted, that is easily explained by the fact that flamewar topics attract shitty comments, and shitty comments get shittier as people get politically and nationalistically riled.
Checks out.
https://www.bing.com/images/search?q=Vehicle+being&qs=n&form...
"Microsoft divested itself of its stakes in the MSNBC channel in 2005 and in msnbc.com in July 2012."
I thought their Edge browser was still promoting links to MSNBC, but looking at Edge now, I see the links are to MSN.com which is a completely different thing and actually a Microsoft news portal.
Interesting the news channel is still called MSNBC though even though Microsoft doesn't have any ownership in it anymore.
[flagged] on submissions nearly always means users flagged it. This is in the FAQ: https://news.ycombinator.com/newsfaq.html#flag
https://news.ycombinator.com/newsguidelines.html
https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...
https://hn.algolia.com/?sort=byDate&dateRange=all&type=comme...
You don't have to believe me, of course, but if you decide not to, consider these two simple observations.
First, lying would be stupid, because the good faith of the community is literally the only thing that makes this site valuable. So, sheer self-interest plus not-being-an-idiot should be enough to tip your priors. I may be an idiot about most things, but I hope I'm not incompetent at the most important part of my job. The value of a place like HN can easily disappear in one false step. Therefore the only policy which has ever made any sense is (1) tell the truth; (2) try never to do anything that isn't defensible to the community; and (3) acknowledge when we fuck up and fix it.
Second, if you're going to draw dramatic conclusions about sinister operations, it's good for mental health to have at least one really solid piece of information you can check them against. Otherwise you end up in the wilderness of mirrors. What you see on internet forums—or rather, what you think you see on internet forums, which then somehow becomes what you see because that's how the brain does it—is simply not solid information. Remember what von Neumann said about fitting an elephant? (https://hn.algolia.com/?dateRange=all&page=0&prefix=true&que...) He asked for a mere five degrees of freedom. Nebulous internet spaces give you hundreds at least. That's way beyond enough to justify anything—even dipping in a ladle and getting one ladle's worth is enough to justify anything.
(Edit: people have been asking what Angela Lansbury has to do with this. If you don't mind spoilers, Angela will explain it for you here: https://www.youtube.com/watch?v=p3ZnaRMhD_A.)
Isn't this already a thing with the "vouch" button on said posts? https://i.imgur.com/Hp9nu58.png
https://www.hollywoodreporter.com/movies/movie-news/top-gun-...
50美分派对 also finds no results on Bing, but plenty of results on Google.
So it's not just tank man. The word filter is active for all search terms that are blocked in china.
50美分的军队 I think is another way of writing it: it shows only 3 images on bing, which all look like ads for Coca Cola.
"Because of a mistake a configuration change meant for some only some regions was also applied to Bing US. This has now been fixed." (To be clear, I just made this up.)
Edit: https://www.reuters.com/technology/microsoft-bing-raises-con...
Actual quote: Microsoft said the issue was "due to an accidental human error and we are actively working to resolve this."
At first I thought it was because safe search was enabled, but disabling it gives the same results (zero).
https://duckduckgo.com/?q=tank+man&t=h_&iar=images&iax=image...
Their regular web search shows no evident censorship though. Why is this? Does DuckDuckGo rely on a third party for image search that is itself censoring results?
The alternative explanation is that bing’s image search just sucks.
Quote from Satya Nadella Q1 2019 Earnings Conference Call "...In fact, this morning, I was reading a news article in Hacker News, which is a community where we have been working hard to make sure that Azure is growing in popularity and I was pleasantly surprised to see that we have made a lot of progress in some sense that at least basically said that we are neck to neck with Amazon when it comes to even lead developers as represented in that community..."
Mentioned here before: https://news.ycombinator.com/item?id=27293480
[dead]: https://news.ycombinator.com/item?id=27397440 (has vouch)
[flagged]: https://news.ycombinator.com/item?id=27396685 (no vouch)
https://web.archive.org/web/20210604192821/https://www.bing....
https://web.archive.org/web/20210604180506/https://images.se...
https://web.archive.org/web/20210604194355/https://www.ecosi...
https://web.archive.org/web/20210604194336/https://search.ao...
Some people probably think "nobody uses Bing", but Bing powers a lot of different search engines (Yahoo, Ecosia, AOL, DuckDuckGo, and more). It's the default search engine on millions of devices (Windows, and even if you change it, Windows search still uses it; Xbox uses it as well.)
But that's just the surface reason. The deeper reason is because of HN's design which seems to weight flags much more heavily than up votes. This means that a non-trivial minority can successfully drag a story off the homepage, even if a majority believes it is important and worthy of discussion, and even if the comments are actually constructive.
Unfortunately we are not able to see what the count of flags are, and I don't know the scoring algorithm to see what portion of minority is needed to make a story disappear, or if it is even greater than a fixed portion (i.e. whether flag count may have a greater than linear effect, compared to linear support from up votes).
I have heard the rationale of attempting to avoid flame wars and heated discussion of controversial topics, but I'm not sure that it's the only way or worth the price of making things important to most of the community disappear. So far, follow up submissions seem to still be on the home page, but as I write this comment, I can see https://news.ycombinator.com/item?id=27396783 already drop from the top position down to #6 and I wonder if it will disappear also.
I hope that HN may reconsider what options are available to keep discussion constructive and useful while still allowing important topics to be seen and discussed without being buried by a minority.
Edit - 2 additional notes a few minutes later:
- https://news.ycombinator.com/item?id=27396783 has now dropped more than halfway off the front page.
- Whoever downvoted this comment, please consider responding as well.
https://duckduckgo.com/?q=tank+man&atb=v228-1&iar=images&iax...
A Microsoft spokesperson tells me that "accidental human error" is to blame for missing images of "tank man" on its Bing search results.
"We are actively working to resolve this."
The incident comes on the 32nd anniversary of the Tiananmen Square Massacre.
https://www.bing.com/images/search?q=%2b%22tank+man%22+tiena...
A Microsoft spokesperson tells me that "accidental human error" is to blame for missing images of "tank man" on its Bing search results.
"We are actively working to resolve this."
The better HN gets, the more people want to suck its juices for their own purposes. Most haven't figured out that the above-board way to do that is simply to make interesting contributions, so they do other things, and there's probably a power law of how sinister those things are. The majority are relatively innocuous, but lame. (Think startups getting their friends to upvote their blog post, or posting booster comments in their thread.)
Users are good at spotting these innocuous/lame forms of abuse, but when it comes to $BigCo manipulation (or alleged manipulation), user perceptions get wildly inaccurate—far below 0.1%—and when it comes to $NationState manipulation (or alleged manipulation), user perceptions get so inaccurate that...trying to measure how inaccurate they are is not possible with classical physics. Almost everything that people think they're seeing about this is merely imagination and projection, determined by the strong feelings that dominate politics.
How do I know that? Because when we dig into the data of the actual cases, we find is that it's basically all garden-variety internet user behavior.
It's like this: imagine you were digging in your garden for underground surveillance devices. Why? Well, a lot of people are worried about them. So you dig and what do you find? Dirt, roots, and worms. The next time you dig, you find more dirt and more roots and more worms. And so for the next thousand places you dig. Now suppose someone comes along and insists that you dig in this-other-place-over-here because they've convinced themselves—I mean absolutely convinced themselves, to the point that they send distraught emails saying "my continued use of HN depends on how you answer this email"—that here is where the underground device surely must be. You've learned how important it is to be willing to dig; even just somebody-being-worried is a valid reason to dig. So you pick up your shovel and dig in that spot, and you find dirt, roots, and worms.
Still with me? Ok. Now: what are the odds that this thing that looks like a root or a worm is actually a surveillance device? Here my analogy breaks down a bit because we can't actually cut them open to see what's inside—we don't have that data. We do, however, have lots of history about what the "worms" have been doing over the years. And when you look at that, what do you find that they've been up to? They've been commenting about (say) the latest Julia release or parser combinators in Elixir, and they've been on HN for years and some old comment talks about, say, some diner in Wisconsin that used to make the best burgers. And in 2020 they maybe got mad on one side or the other of a flamewar about BLM. (Nobody please get mad that I'm using worms to represent HN users. It's just an analogy, and I like worms.)
Or, maybe the history shows that the person gets involved in arguments about China a lot. Aha! Now we have our Chinese spy! How much are they paying you? Is it still 50 cents? I guess the CCP says inflation doesn't exist in China—is that it, shill? If @dang doesn't ban you, that proves he's a CCP agent too!
But then you look and you see that they've been in other threads too, and a previous comment talks about being a grad student in ML, or about having married someone of Chinese background—obvious human stuff which fully explains why they're commenting the way they are and why they get triggered by what they get triggered by.
This ordinary, garden-variety stuff—dirt, roots, and worms in the analogy—is what essentially all of the data reduces to. And here's the thing: you, or anyone, can check most of this yourself, simply by following the public history of the HN accounts you encounter in the threads. The people jumping to sinister conclusions and angrily accusing others don't tend to do that, because that state of mind doesn't want to look for countervailing information. But if you actually look, what you're going to find in most cases is enough countervailing information to make the accusations appear absurd...and then you'd feel pretty sheepish about making them.
I'm not saying the public record is the entire record; of course it isn't. We can look at voting histories, flagging histories, site access patterns, and plenty of other things that aren't public. What I'm saying is that, with rare exceptions [1], what we find after investigation of the private data is...dirt, roots, and worms. It looks exactly like the public data.
And here's the most important point: the accusations about spying, brigading, shilling, astroturfing, troll farms, and so on, are all exactly the same between the cases where the public data refutes them and the cases where the public data is inconclusive. I realize this is a subtle point, but if you stop and think about it, it's arguably the strongest evidence of all. It proves that whatever mechanism is generating these accusations doesn't vary with the actual data. Moreover, you don't need access to any private data to see this.
There are also trolls and single-purpose accounts that only comment in order to push some agenda. That's against the HN guidelines, of course, and such accounts are easy enough to ban. But even in such cases, it doesn't follow that the account is disingenuous, some sort of foreign agent, etc. It's far more likely that they're simply passionate on that topic. That's how people are.
[1] so rare that it's misleading to even mention them, and which also don't look anything like what people imagine
---
Still, power laws have long tails and one wonders what may lie at the end, beyond our ability to detect it. What if despite all of the above, there is still sinister manipulation happening, only it's clever enough to leave no traces in the data that we know of? You can't prove that's not happening, right? And if anyone is doing that it would probably be state actors, right?
You might think there's nothing much to be said about such cases because what can you say about something you by definition don't know and can't observe? It seems to get epistemological pretty quickly. Actually, though, there's a lot we can say, because the premise in the question is so strong that it implies a lot. The premise is that there's a sort of Cartesian evil genius among us, sowing sinister seeds for evil ends. I call this the Sufficiently Smart Manipulator (SSM): https://hn.algolia.com/?dateRange=all&page=0&prefix=false&so....
There are two interesting things about the SSM scenario. The first is that since, by definition, the SSM is immune to anti-abuse measures, you can't postulate any technical measures for dealing with it. It's beyond the end-of-the-road of technical cleverness.
The second interesting thing is that, if you go in for this way of thinking, then either there already exists an SSM or there eventually will be one. And there's not much difference between those two cases. Either way, we should be thinking about what to do.
What should we do in the presence of an SSM? I can think of two options: either (1) give up, roll over, and accept being manipulated; or (2) develop a robust culture of countering bad arguments with better ones and false claims with true information. Of those options, (2) is better.
If you have such a culture, then the SSM is mitigated because the immune system will dispose of the bad parts of what they're saying. If there are any true bits in what they're saying, well, we shouldn't be rejecting those, just because of who said them. We should be big enough to accommodate everything that's true, regardless of where it comes from—just as we should reject everything that's false, regardless of where it comes from. We might prefer to reject it a little more rudely if we knew that it was coming from an SSM, but that's not a must-have.
The nice thing is that such a culture is exactly what we want on HN anyway, whether an SSM exists or it doesn't. The way to deal with the SSM is to do exactly what we ought to be working at as a community already: rejecting what's false and discovering what's true. Anti-abuse measures won't work forever, but we don't need them to—we only need them to last long enough to develop the right habits as a community. If we can reach a sort of (dare I say it) herd immunity from the viruses of manipulation, we'll be fine. The answer to the Sufficiently Smart Manipulator is the Sufficiently Healthy Community. That's what the site guidelines and moderation here are trying to nurture.
Edit: I should add that I'm not 100% confident that this can work. But it's clear that it's the best we can do in that scenario, and the good part is that it's what we ought to be doing anyway.
https://www.bing.com/images/search?q=Unknown+Protester+&form...
https://yandex.ru/video/search?text=navalny%20putin%27s%20pa...
It looks like microsoft caught on and is now returning generic results of "tank" or "man":
(I did some searching.) This is from a report here:
https://www.msn.com/en-gb/news/world/microsoft-bing-raises-c...
Of course, the search result is an MSN link.
The most disgusting aspect is the blatant lying and hypocrisy: https://news.microsoft.com/on-the-issues/topic/defending-dem...
Makes me wonder what other pitfalls there are to using DDG.
https://en.wikipedia.org/wiki/Tank_Man
and though the image searches do show a bunch of other things (all including "tank man" in the name at least), they _do_ also include the iconic photo at least once:
http://www.maryscullyreports.com/wp-content/uploads/2017/06/...
which I did not see in the bing results...
https://twitter.com/josephfcox/status/1400913178125553665?s=...
Microsoft has acknowledged this is in error...[1] (though the error seems to be that censorship meant to just apply to China is being applied everywhere)
[0] https://news.ycombinator.com/item?id=27395635
[1] https://www.vice.com/en/article/qj8v9m/bing-censors-tank-man
I hope Microsoft has a good write up on how this happened, both on the search term blanket ban and the safe search results thing.
To be clear, I don't care about $BigCo, only HN; and if there's any actual evidence that this whole thing was anything other than randomness playing its usual tricks on the hivemind, that would actually be interesting and on-topic for HN. But note those words "actual" and "evidence".
> DuckDuckGo distinguishes itself from other search engines by not profiling its users and by showing all users the same search results for a given search term.
Also note that DDG doesn’t just rely on Bing.
> DuckDuckGo's results are a compilation of "over 400" sources, including Yahoo! Search BOSS, Wolfram Alpha, Bing, Yandex, its own web crawler (the DuckDuckBot) and others.
https://www.bing.com/search?q=Unknown+Protester&search=&form...
Edit: I got to it! See the lower portion of https://news.ycombinator.com/item?id=27398725, after the "---".
If you or anyone notices something wrong with the argument, I'd like to hear what it is.
Plenty of results on Google: https://www.google.com/search?q=tank+man&tbm=isch
Google: 1, Bing/DDG: 0
So much for "privacy/simplicity", eh?
The pattern seems clear that these users are flagging the more sensational kinds of submissions that tend to lead to predictable discussions and flamewars. There's room for competing opinions about which of those are/aren't on-topic for HN, given the site guidelines; if you or anyone want to understand how the mods look at it, I recommend the explanations at the links below. But clearly the flagging behavior in this case was in good faith.
https://hn.algolia.com/?dateRange=all&page=0&prefix=false&so...
https://hn.algolia.com/?dateRange=all&page=0&prefix=false&so...
https://hn.algolia.com/?dateRange=all&page=0&prefix=false&so...
https://hn.algolia.com/?dateRange=all&page=0&prefix=true&sor...
https://hn.algolia.com/?dateRange=all&page=0&prefix=true&sor...
Btw I'm not necessarily agreeing that those were bad flags; but in cases like this, the community has the final say.
Apparently this is folklore: https://en.wikipedia.org/wiki/NBC_chimes#%22General_Electric...
I do not want to single out a single company, but would like to use this particular example to ask you the following: Please keep in mind the level of manpower and persistence, some of these corporations can call upon for their strategic objectives..
In 2020 Microsoft had, apparently, 106 lobbyist companies working on its behalf: https://www.opensecrets.org/federal-lobbying/clients/lobbyis...
and 94 in 2021 https://www.opensecrets.org/federal-lobbying/clients/lobbyis...
Looking at the website of some of these companies, offered services include and quoting: "Third party influencer outreach" :-)
I don't doubt you believe that too. But as far as I can tell, it's what people in western forums believe too as long as those Chinese people don't opine in a way that's similar to the government that 95% of them support (otherwise they're bots, coherced or brainwashed, or otherwise not full humans).
https://news.harvard.edu/gazette/story/2020/07/long-term-sur...
https://hn.algolia.com/?dateRange=all&page=0&prefix=false&qu...
While I certainly agree (if I hear you correctly) that truth matters and there is such a thing as loving the truth for its own sake, I don't agree that public discourse can be divided in the way you posit. Quite the opposite.
DDG's main search results mostly come from Bing, as the source for that Wikipedia statement reveals:
https://help.duckduckgo.com/results/sources/
> We also of course have more traditional links in the search results, which we also source from multiple partners, though most commonly from Bing (and none from Google).
Anyone can see this for themselves by comparing a number of Bing and DDG searches.
In a hilarious twist of fate, searching for that term brings up papers on either medical research or peer-review reliability problems in general[0]. You try to find data on a potentially abstract, complex societal issue, and come up with what can only be described as attention grabbing HN-catnip.
0: https://www.cambridge.org/core/journals/behavioral-and-brain...
https://upvotetracker.com/post/hn/27395635
Related Show HN by @janmo: https://news.ycombinator.com/item?id=27092770
Edit: I'm not sure I'm a big fan of his current sponsoring link however... maybe worth a look too.
"A significant percentage of the Microsoft employees who work on Bing are based in China, including some who work on image-recognition software, according to a former employee."
China is known to require search engines operating in its jurisdiction to censor results, but those restrictions are rarely applied elsewhere."
https://www.history.com/topics/us-presidents/tank-ride-video
> What does [flagged] mean?
> It means that users flagged a post as breaking the guidelines or otherwise not belonging on Hacker News.
I quite like this article from 2019 about Wandering Earth which tries to explain two facets of the contemporary Chinese right wing, that is the 小粉紅 Little Pinks and 工業黨 Industrial Party. Or, as they choose to translate it, "Young Cyber-Nationalists" and "Prometheans". In English: https://chuangcn.org/2019/08/wandering-earth/
Edit: oof, that link does look awful doesn't it. Most "how to get on HN's front page" content is terrible, it doesn't work and induces people to post dross and pull tricks that just degrade the site. I've got a set of notes about how to write for HN that I want to publish someday. If anyone wants a copy they can email hn@ycombinator.com.
I can't comment on this particular slick content marketing course because apparently you have to buy it to find out what it says, but previous ones I've seen have been entirely unreliable, and the look and feel of the ad certainly seems antithetical to the spirit here.
Tao Te Ching, Ch. 17,
With the best kind of rulers
When the work is complete
The people all say
"We did it ourselves."
In re: Option 2:
https://xkcd.com/810/ "Constructive"
> [[A man is talking to a woman]] Man: Spammers are breaking traditional captchas with AI, so I've built a new system. It asks users to rate a slate of comments as "Constructive" or "Not constructive". [[Close up of man]] Man: Then it has them reply with comments of their own, which are later rated by other users. [[Woman standing next to man again]] Woman: But what will you do when spammers train their bots to make automated constructive and helpful comments? [[Close up of man again]] Man: Mission. Fucking. Accomplished. {{Title text: And what about all the people who won't be able to join the community because they're terrible at making helpful and constructive co-- ... oh.}}
Cheers dang.
Western intelligence generally rush to protect their assets, or at least they used to...
These regime change attempts have been ongoing for about a century, opium/heroin dealers/criminals and their western partners/suppliers didn't like getting shutdown by the Chinese government, primarily Mao.
When people ask, "Why is China a Surveillance State?", this is why...
https://en.m.wikipedia.org/wiki/Concert_for_Democracy_in_Chi...
Was the primary driver of the Tiananmen Square protest caused by the CIA? Every credible source I’ve found indicates that the CIA was trying to figure out what was going on, but not initiating the protest.
You are saying that if a CIA regime change pattern is detected and the CIA doesn't publicly disclose its covert/overt operations (even as it does expected post-op cleanup), it probably didn't happen? Quite interesting to maintain a permanent state of denial when history says the opposite is more appropriate...
After looking at timeline of CIA destabilization operations (using NGOs, rebel groups, terrorists, etc...) around the world and the constant funding of protests and extremism in China/Tibet/HK... Anyone that can do elementary level pattern recognition can see this for what it is...
This regime change operation extends to the repeated HK protests (many participants are paid to attend) and Xinjiang extremism (participants funded, armed, and trained by the CIA and their partners in Turkey/Syria/Iraq).
Paid HK protestors - https://www.rt.com/op-ed/525596-hong-kong-paid-protests/
Uyghur extremists as a China destabiliser - https://m.youtube.com/watch?v=00Cvx0R8iDo
Uyghurs militants conveniently allowed into Syria to help the US overthrow Assad - https://www.timesofisrael.com/uighur-militants-in-syria-look...