SoylentNews is people
SoylentNews
Last week an article from the BBC said:
Google has said it is 'thinking deeply' about ways to improve search, after criticism over how some results - including ones discussing the Holocaust - were ranked.
[...] Google - which processes five billion searches a day - was keen to come up with a solution that was broadly applicable across all searches, rather than just those that have been noticed by users.
"It's very easy to take a search here and there and demand Google change something," explained Mr Sullivan, "and then the next day you find a different search and say, 'why didn't you fix that?' "
This week we see the results of their efforts: Google has modified PageRank to surface "more high-quality, credible content on the web":
Google's technology was changed again after people spoke out about how typing in "are Jews evil" in the autocorrect function resulted in offensive terms. Also, when people searched "who runs Hollywood?" the result, "Jews," was scrubbed last year. Google said its algorithm incorrectly gave "authority" to a site that suggested so because it was linked to over and over again.
But Heidi Beirich, intelligence project director for the Southern Poverty Law Center, said Tuesday that Google has a long way to go to "clean up its act." While searching for "did the Holocaust happen?" no longer shows one white supremacist site at the top, searching for "is the Holocaust real?" still provides a site up high that claims it's a hoax.
(Score: 5, Insightful) by SomeGuy on Saturday December 31 2016, @12:17AM
And yet with all these improvements, the top 10 sites for many obscure searches I do are still content-less crap sites. The problem here is that one person's "better results" is another persons censorship.
Last week Google de-listed a site I frequent for fuck knows why. (Vetusware.com if you really want to know).
Google has an amazing amount of power. If they don't want you on the web, they can remove you at a press of a button.
(Score: 3, Interesting) by Francis on Saturday December 31 2016, @12:20AM
These sorts of shenanigans are a large part of why I don't use Google. It's not just the delistings and the constant modification of how pages are ranked, it's that they even rank things in the first place by things other than how closely it matches my current search query.
If I wanted other things, then I should type in a different query. If somebody happens to be interested in Neo-Nazis, then they should receive the results for those sites. Making it harder to get to those sites isn't going to solve anything, but it will increase the number of theories about the Jews controlling the internet and whatever other things they control.
(Score: 0) by Anonymous Coward on Saturday December 31 2016, @12:23AM
What does this even mean? Isn't this completely subject to change, since machine learning and eventually real AI would be needed to determine "how closely it matches my current search query"? Give GOOG some credit.
(Score: 5, Insightful) by Francis on Saturday December 31 2016, @12:29AM
It's called being "bubbled" where they take the things that match what you typed in and then filter it based upon what they think you'll want.
The problem is that it leads to biased results that don't necessarily match with what you're looking for. It's a very dangerous way of handling the situation as it runs the risk of trapping you in a bubble where you think your views are informed, but they're informed by a skewed subset of the possible views on the topic.
The search engine should do it's absolute best to give me what I ask for and nothing else. It shouldn't be making editorial decisions about what I really meant to search for. If I search for the holocaust, it should include all the sites that deal with the topic, not just the ones that deal with the acceptable views on it. I can personally filter out the neo-Nazi content and similar since that's not what I'm looking for.
(Score: 0) by Anonymous Coward on Saturday December 31 2016, @12:54AM
I thought being bubbled involved a hooker and some champagne?
(Score: -1, Troll) by Anonymous Coward on Saturday December 31 2016, @01:07AM
Not for Francis, it doesn't! This is why he doesn't do google.
(Score: 5, Interesting) by aristarchus on Saturday December 31 2016, @01:57AM
Francis, Francis, my poor deluded Francis!
The problem is that it leads to biased results that don't necessarily match with what you're looking for.
Exactly how do you know what you are looking for, such that you can identify biased results? If you can recognize "biased results", you already knew what you were looking for, so you were not in fact looking for something, you were looking for confirmation, as in "confirmation bias".
Plato discussed this in his dialogue The Meno:
Meno: And how are you going to search for [the nature of virtue] when you don't know at all what it is, Socrates? Which of all the things you don't know will you set up as target for your search? And even if you actually come across it, how will you know that it is that thing which you don't know? (80d)
So if you are looking for something, and you do not already know what it is you are looking for, you will not be able to know when you have found it. Or, in your case here, you will not be able to know that what Google has given you is not what you were looking for.
It's a very dangerous way of handling the situation as it runs the risk of trapping you in a bubble where you think your views are informed, but they're informed by a skewed subset of the possible views on the topic.
The greater risk is thinking you are in a bubble, when there is nothing that distinguishes the bubble from reality other than your own private wishes about what reality should be. This is the technique that is being used to defend the alt-right. Fake news just means disagreement, discounting racist sexist fantasy is censorship! Why? Because Google and Facebook and Universities and Stephan Colbert could be lying to us. The ([(Obvious Rebuttal)]), of course, is that it is also possible they are not. How could you tell? Knowing is a bitch, Francis!
(Score: -1, Offtopic) by Anonymous Coward on Saturday December 31 2016, @03:51AM
You mean sort of how nothing intelligent can pass through your thick skull?
The greater risk here is in filtering out material in reality. If nobody is doing any filtering then there's at least some way of addressing the situation. But, if there's filtering going on, then there's no way of knowing what is and isn't being kept from the public and there's a very big problem about who draws what lines where.
Seriously, I've had to re-evaluate my views on several posters here as you've set a new all time low for lack of intellectual capacity. I'm not positive, but I think you might actually be a cucumber. Probably the one crammed up your mother's cunt.
(Score: 2, Insightful) by aristarchus on Saturday December 31 2016, @05:05AM
Seriously, I've had to re-evaluate my views on several posters here as you've set a new all time low for lack of intellectual capacity.
The first step is to admit you have a problem. Have you noticed that a lot of comments here just go over your head? Do posters use words you do not know? Can you tell me why you want to be a cucumber? What was your relation with your father like? Do you miss the interuterine state? I prescribe education! Even Community College! Not a bad place to start. You will be exposed to evidence verification, inferential reasoning, the scientific method, and if you take some philosophy, the "massively mistaken" hypothesis. Worth a shot. Better than hanging around on SoylentNews and trying to beat up on 2400 year old philosophers, where you are quite clearly, to everyone, out of your league. We're coming for you, Francis!
(Score: -1, Offtopic) by Anonymous Coward on Saturday December 31 2016, @05:31AM
The greater risk here is in filtering out material in reality. If nobody is doing any filtering then there's at least some way of addressing the situation. But, if there's filtering going on, then there's no way of knowing what is and isn't being kept from the public and there's a very big problem about who draws what lines where.
Oh, yes, since clearly you would be able to detect unfiltered reality, because of it's "truthiness". No, there is no reality, there are no objective facts, and there is no way to know whether anything lines anywhere taco bravo indulges kumquat. Ant this goes doobie for Califiorina. There are several (two) songs you should listen to:
http://www.imdb.com/title/tt0126289/ [imdb.com]
https://www.youtube.com/watch?v=KThlYHfIVa8 [youtube.com]
https://www.youtube.com/watch?v=7xEwbzEBw14 [youtube.com]
'
Satan is the source.
(Score: 5, Insightful) by AthanasiusKircher on Saturday December 31 2016, @07:01AM
Exactly how do you know what you are looking for, such that you can identify biased results? If you can recognize "biased results", you already knew what you were looking for, so you were not in fact looking for something, you were looking for confirmation, as in "confirmation bias".
While I understand what you're getting at, you have to admit that Francis has a point here too. Things called filter bubbles [wikipedia.org] do exist.
And the way I knew it was happening to ME was not by reading a Socratic dialogue, but because in 2012 I got interested in following the Ron Paul fiasco for a while. NOT because I was a Ron Paul supporter, but because I found the whole thing interesting (and while the Paul folks certainly behaved unfairly in some ways, the RNC's suppression of them was also quite a bit crazy at times).
Anyhow -- for several months back then I did a Google search for "Ron Paul" probably every other day or so to check in and see if any news appeared. Eventually, I started seeing lots of Ron Paul headlines showing up. After having a conversation with a friend, I realized when we both did searches that we were seeing DIFFERENT results, due to our search histories.
So, pace Meno, it IS possible to realize that someone is messing with your "reality," or at least your search results.
Since then I have never done a Google search from a browser where I'm logged into a Google account. But that's still not sufficient to get you out of Google's crappy system, which personalizes your results based on where it thinks you are (even if you don't tell it), etc. Yes, in some cases this can be a feature, but eventually in the past few years I quit Google as my primary search engine because I was spending more time "fighting" it to get it to search for what I actually want, rather than what it THINKS I want.
Back in the early 2000s, Google was fantastic, because you could actually do a full-text search on basically the entire internet. It was AWESOME. You could come up with a set of four search terms that you knew could uniquely identify what you wanted, and you could get a handful of predictable hits. That's just not possible anymore. "Verbatim" search is broken. (See the Google forums, if you don't believe me -- it's been broken ever since it was introduced after Google deprecated the + operator.) You can try "allintext:" as an operator, which works much better than verbatim, but it still fails in all sorts of cases. If you start using search tools, be prepared to discover all sorts of weirdness -- if you limit results to 1990 to 2000 vs. 1999 to 2000, be prepared to get different lists of results for the period 1999 to 2000, even though both should give you the same results. It's a MESS.
I know it's mostly because there are AI algorithms trying to "give people what they want" rather than what they literally search for. But it drives me crazy. And yes, it can also significantly distort the results you get.
I get that you are arguing with Francis because you are concerned he just wants to be able to see his favorite political news all the time, and he doesn't want people arbitrarily deranking it. But the larger issue here is that Google does reshape search results all the time to "personalize" them for people, and that's a huge problem that actually makes your concerns WORSE... whatever Socrates might say about it.
(Score: 2, Insightful) by aristarchus on Saturday December 31 2016, @07:29AM
While you are quite correct,
While I understand what you're getting at, you have to admit that Francis has a point here too.
No, Francis has no point, other than to be Francis, Francis who does not know things. In fact, does not know any things, but still has an opinion he must post here. So, please, spare me, and the rest of us Soylentils.
So, pace Meno, it IS possible to realize that someone is messing with your "reality," or at least your search results.
My point is, that it is not possible. It is only desireable. You get one search result. Your friend gets another. And I, of course, as a 2400 year old philosopher, get the real results. Not. You may suspect that someone is messing with your search results, but I suggest that if you have to suspect something like this, you are already beyond the Pale, you have no idea what actual evidence and facts and reality are, and since you are AthanasiusKircher, one of the few rational voices here on SoylentNews, I despair. The Francises are taking over.
But the larger issue here is that Google does reshape search results all the time to "personalize" them for people, and that's a huge problem that actually makes your concerns WORSE... whatever Socrates might say about it.
Of course. The question is what you do about it. From the beginning search engines have been suspect. Real scholars know how to do real research. I am just suggesting that Plato, through his dramaticae personae of Socrates, is still correct. Sophistry actually gains nothing by technology, and in fact may lose more than it gains. Nazis, damn nazis!
(Score: 0) by Anonymous Coward on Saturday December 31 2016, @09:32AM
I see, so you admit that you're not a scholar. This is an internet forum, not a research paper. I am not going to waste time looking up every possible thing that people could disagree with just because aristarchus has a small penis and a large ego.
Doubly so since you can't even be bothered to do more than rudimentary word mining.
(Score: 2, Insightful) by aristarchus on Saturday December 31 2016, @09:58AM
Triply so: what the fuch are you talking about? Are you a complete idiot? Or an incomplete idiot? Either way, I recommend, as usual, more education. Evidently you need help to realize how profoundly ignorant you are. Nothing to be ashamed of, we all have to start somewhere. Oh, why is it that internet forums are not held to the same standards as research papers? Is not what we are doing here peer review? Of course, the peers are somewhat less knowledgeable, so not sure that works out. But, we can aspire.
(Score: 2) by darkfeline on Saturday December 31 2016, @08:57AM
Let's say you do a lot of Ron Paul searches. Now Google has to show you some Ron Paul results. What should Google show you?
Should Google show you what it thinks you want, or should Google show you something you don't want ("the truth")? Which is more ethical? Does Google have the right to decide what "the truth" is? But on the other hand, is it ethical for Google to perpetuate your bubble of ignorance?
Ethics is a bitch, but complaining about Google is easy, because whichever choice Google makes there's room for an attack. Finding a solution? Nah, that's too hard, much easier to complain on the Internet.
Join the SDF Public Access UNIX System today!
(Score: 2, Interesting) by Francis on Saturday December 31 2016, @09:21AM
Google should show you what you ask for plain and simple. If that's not what you want, you can always refine your search criteria. But if google of some other engine makes those decisions for you, now what, the only way to know something has gone wrong is if you stumble on different results with a different browser.
To make matters worse since they do it behind the scenes it's much harder to know how to refine the search as you don't really know exactly what tgey did to the results. Ever try googling for an older article? Chances are you can't get it unless you know exactly what to type and it isn't the same as the new stuff. Knowing roughly the title doesn't cut it either.
(Score: 3, Informative) by aristarchus on Saturday December 31 2016, @10:48PM
Chances are you can't get it unless you know exactly what to type and it isn't the same as the new stuff. Knowing roughly the title doesn't cut it either.
Why would you expect it would? Have you even done old school, hard-copy, card-catalog research in a thing called a "library"? Knowing "roughly" is not knowing, Francis! You have to know how to search, and not rely on algorithms. You collect clues, you process leads, you narrow it down. If you can't find something, it means one of two things: either it doesn't exist, or you are not very good at searching.
(Score: 2) by darkfeline on Saturday December 31 2016, @10:57PM
How does Google know what you are asking for? Humans aren't very good at being unambiguous and explicit. Maybe Google could gather lots of personal data to try to improve their algorithms, but then people complain about privacy (and even then, it's hard to tell what humans want. Have you TRIED understanding a woman? (a joke, if your sensors are broken)) But if Google tries to not be all-knowing, people complain about Google not showing you what you want.
Keep complaining, maybe that will fix everything.
Join the SDF Public Access UNIX System today!
(Score: 0) by Anonymous Coward on Monday January 02 2017, @01:01AM
Or, they could implement things that we used to have back in days of yore, like, you know, the near operator. I see no evidence that they have that operator at all, they certainly don't list it with their other operators.
(Score: 0) by Anonymous Coward on Sunday January 01 2017, @04:09AM
Given modern search capabilities, we are probably just going to have to accept that these "bubbles" exist.
But the point of the main story is that there needs to be a way to let users know they are in a bubble and have some way to get out of that bubble.
Just a couple of generalized examples, if I click through 12 pages of search results to get to the one I want, a modern search engine should go "oh, THAT is what you were looking for, I'll try to do better next time".
And if the bubble is as big as described above, I would hope to see some indication that I am getting personalized results and have an option to search without or perhaps "pop" that bubble.
Usually Google SEEMS good about alerting users when links have been censored (such as by DMCA), but when they don't, that is a big problem.
(Score: 0) by Anonymous Coward on Saturday December 31 2016, @09:40AM
Ehm, one can tell a result is biased by just having cursorary knowledge in a field.. For instance - search for :nuclear incident: and tell me how many of your results where about cracked fuel claddings, misplaced containers and maintance issues of safetysystems with no consequence (this is what a nuclear incident is [INES 2-3]); and how many was about TMI, Chernobyl or Fukushima (nuclear accidents [INES 4-7]) and how many highlighted :nuclear disaster: (not even a technical term).
(And no, didn't know what I was looking for - ended up at an IAEA-page I hadn't read after putting in a few exclusions for disaster and accident)
So, above was an example from yesterday about how I got a biased result without knowing what I was looking for that wasn't about confirmation bias (I was able to tell simply by knowing what the phrase I searched for means)
(Score: 0) by Anonymous Coward on Saturday December 31 2016, @05:10PM
I don't know why search engines nowadays insist on listing pages that don't have all my search terms. I put those terms there for a reason.
At least have a nonfuzzy mode. Either that or improve your "AI" so it stops list SEO trash websites that don't contain all my search terms for very specific searches.
(Score: 1, Informative) by Anonymous Coward on Saturday December 31 2016, @12:31AM
> Last week Google de-listed a site I frequent for fuck knows why. (Vetusware.com if you really want to know).
Probably because its 100% piracy. You can argue about whether it should be considered piracy or not, but the fact is that. due to the ridiculously long length of copyright, sharing abandonware is, by law, piracy.
(Score: 2) by SomeGuy on Saturday December 31 2016, @02:09AM
If you wanted to get pedantic about it, there are copyright violations just about everywhere, so you might as well just pull the plug on the entire internet.
But there are hundreds of other sites that host abandoned/vintage software. Should Goolge also pull the plug on Bitsaver's or archive.org? What about small personal pages that host, for example, someones personal 8" disk CP/M software collection with the hope of helping other with the same machine? What about history sites that let people see and try out vintage software? Will they de-list soylentnews just for mentioning it? Really, where does it end?
I actually suspect that the de-listing may not even be because of the content, but rather their mis-use of multiple domain names. But Google isn't saying the reason.... as far as Google and most Google users are concerned vetusware simply no longer exists!
(Score: -1, Troll) by Anonymous Coward on Saturday December 31 2016, @04:11AM
> If you wanted to get pedantic about it, there are copyright violations just about everywhere, so you might as well just pull the plug on the entire internet.
Ugh. Typical reductive geek logic used to define away the question rather than address it. Its not insightful, its deflection. If you don't want to understand why google did what they did, that's your prerogative. But don't make your ignorance into a cause for righteousness.
> I actually suspect that the de-listing may not even be because of the content, but rather their mis-use of multiple domain names.
That's not true, all you gotta do is type ventusware into google and the .org site comes up.
(Score: 2) by jmorris on Saturday December 31 2016, @05:01AM
Lets really go for blood here. Youtube was explicitly built on rampant, massive, pervasive copyright violation. They encouraged and facilitated random users on the Internet to upload content all parties knew was copyrighted by someone other than the uploader. Then Google used its size to simply bully the much smaller (in market cap) content producers into signing deals to turn all that illegal content semi-legal. They have zero moral ground to stand upon.
And they, along with the bigots at SPLC have zero moral ground here as well. Google has built its reputation on using computer algorithms to generate the page rankings with minimal human intervention. That is their 'elevator pitch', it is as neutral as they can make it, only intervening to manually punish people who try to game the system and even then only until they can iterate a better algorithm to automatically punish attempts to get spam into the rankings instead of content users actually want to see. If they actually do this it will destroy them, and if they are stupid enough to be played by the narrow minded bigots at SPLC they will deserve to suffer.
Funny, somehow I doubt we will see zany conspiracy theories favored by the left getting memory holed. Bush blew the levees and drowned New Orleans. Bush blew up the WTC. The moon landing was fake. Bush lied us into a war. Chemtrails (crazies on ALL sides), The Russians did it!, Valerie Plame was outed by Dick Cheney. Need I continue?
(Score: 2) by Runaway1956 on Saturday December 31 2016, @12:42AM
vetusware? Obviously a spam site. OS2? Subversive! MS-DOS 6.0? REGRESSIVE!! Oh, God, there's Wordstar! Whacko science fiction writers are known to use that! Vetusware needs to be censored right out of the universe!
/sarcasm
The games list is pretty impressive too. Errr, I mean, LET'S STOMP OUT THAT ABANDONWARE!!!
Abortion is the number one killed of children in the United States.
(Score: 0) by Anonymous Coward on Saturday December 31 2016, @04:02AM
That most of these sites don't have a darknet presence and the foresight to nudge their collective userbases to learn to use such software to help protect themselves and the service hosters from easily tracked future enforcement actions.
(Score: 2) by Webweasel on Saturday December 31 2016, @02:11PM
Funny that.
Using a darknet or VPN instantly flags you with the authorities. Use either of these services and GCHQ/NSA will have a file on you. A lot of sites in the TOR network are honey traps made by TLA's and the like.
Priyom.org Number stations, Russian Military radio. "You are a bad, bad man. Do you have any other virtues?"-Runaway1956
(Score: 5, Insightful) by requerdanos on Saturday December 31 2016, @02:26AM
If Google doesn't want you on the web, there is precious little they can do about it.
If, however, Google doesn't want you in *their search listings*, then you don't show up there; that's how the game is played.
Thankfully, there [yahoo.com] are [duckduckgo.com] a [bing.com] few [facebook.com] other [yandex.com] popular [baidu.com] sites [twitter.com] that people widely use to publicize web sites either by search indexing or social-type sharing.
The web did fine before Google existed (Anyone remember Excite? Inktomi/Hotbot? Altavista? The December List? Yahoo's hand-curated index?) and will doubtless do fine in a post-Google world.
(Score: 1) by TrentDavey on Saturday December 31 2016, @02:49PM
... and don't forget Archie [wikipedia.org]. I would submit search queries and get results days later (if I'm recalling correctly). And I used Lynx [wikipedia.org], the text-based browser. And before that when I was younger, I had to get up off the couch, walk to the TV and change the channel via a knob. Get off my lawn! ... but I digress.
(Score: 2) by linuxrocks123 on Saturday December 31 2016, @02:53AM
That looks like it might be a mistake on Google's part somehow. Using Google's Transparency Report, I was able to find this on Lumen.org as the only complaint Google has received about vetusware: https://lumendatabase.org/notices/1847688 [lumendatabase.org]
And that was in 2014, so I don't know what happened.
Incidentally, the complaint was about an old version of SpinRite, which is a well-known snake oil scam. This guy Steve Gibson, a grandiose self-promoter who has much more marketing sense than programming sense, has managed to make a business charging $89 for a poor imitation of dd_rescue. The kicker is he uses the fact he was stupid enough to write an I/O-bound program in assembly language to market to rubes who think that will make it run faster. Philosophical question: is it still a scam if you believe your own bullshit?
It doesn't surprise me at all he'd pay some company to harass Google about a site hosting software he wrote in the 80s. That guy is basically the Donald Trump of tech.
Well ... a Donald Trump of tech. We have narcissists aplenty in our field, unfortunately. :-(
(Score: 1, Informative) by Anonymous Coward on Saturday December 31 2016, @03:57AM
And maybe we can push for a Solus(or whatever the java/apache search engine is called) with a DHT for search terms so end users can do spidering and help provide results we actually want, rather than relying on commercial sites shoving us what we don't.
There are two ways to get free: whore yourself to some company with some strings attached, and setup, support, and frequent volunteer alternatives to the commercial endeavors. The latter will take up more of your time, but if both your privacy and your quality of service are important to you, there is no better way than doing it yourself (or as part of a volunteer organization.)