Hacker News Reader

Comment by trendia | Hacker News Reader

By no means do I agree with much of what is on the World Socialist Web Site (WSWS). However, from a purely algorithmic standpoint, their website has content that should rank very highly:

1) Much of their content is original (i.e. not copy-pasted from other sites)

2) A majority of the page is composed of text, rather than images or ads

3) Where there are ads, they are minimal and largely hidden from view.

4) Their website is well organized, with clear and descriptive links.

5) Their website is updated frequently and new content is posted regularly.

So, assuming that WSWS is linked by other sites on the same topic, it seems that it should rank highly in Google search queries. Maybe there is something else that would reduce their ranking, but it doesn't seem obvious to me at first glance.

View on HN

None of those things you mentioned are the core of page rank; they're just extra heuristics. None of that matters if they don't get linked by other sites with high page rank.

Page Rank is essentially deprecated and those other heuristics are more impactful than they were when people still talked about Page Rank a lot.

Certainly, things have changed. But I don't believe that the core page rank concept has been removed or even lost primacy.

(I work for Google, but not on anything related to search and I don't have any special knowledge of the algorithms used)

Spot-checking, and I agree, this doesn't at all seem to be "fake news". Much like the actual communists that I've known, I find myself disagreeing with the opinions, priorities, and judgements, but it doesn't seem to be misleading or mischaracterizing the truth.

The thing about "fake news" is that nobody can agree on an objective definition of what it is. So people just train their models on controversial pages and hope for the best.

There's a taxonomy. As far as I can see there is:

* Propaganda/Information Warfare (motive: advance foreign policy goals)

* Clickbait (motive: $)

* Stories that portray our executive leadership in unflattering terms (motive: journalism, reportage)

I'm sure there are others, including perhaps whatever classification is targeting 'voices on the margin.' Hearkening back to "When Wikileaks met Google," this seems like a feature, and not a bug to me.

I keep having this argument about the dictionary. Just because enough people use a term incorrectly doesn't make the term adopt new meaning. Language does shift and that should be respected, but if you take it too far (which many do) nothing means anything any more.

To the point, the people calling truth they don't like "fake news" aren't adding a definition to the term. They're just lying. Fake news is news which is objectively faked. Definitions of "news" shouldn't be controversial, and neither should definitions of "fake".

Fake news are objects which try to pass themselves as news but substantially misrepresent the truth through malice or gross ignorance. Or more softly, Fake News is anything like an Onion article except not (almost) everyone is in on the joke. (and maybe the joke isn't so much a joke as a lie)

"Fake news" is such a non descript term, it's essentially double speak (i.e. it means whatever I disagree with). I mean, was WMD in Iraq Fake news?

The point is, press lost all credibility a long time ago, so now we can apply this label far and wide.

No, the producers of fake news are defending themselves by using doublespeak to dilute the accusation.

from a purely algorithmic standpoint, their website has content that should rank very highly

I think there is a causality error here. Traditionally things which meet those criteria have been ranked highly, but not because of those things. Rather, those things have been easy to detect and have correlated well with quality search rankings.

But that doesn't mean that they are (or should) be the only signals a ranking algorithm can use. I'd argue that a hypothetical AI search engine would do a lot more verification of content as part of the ranking process. It is absolutely true that this remains challenging at the moment (verifiable implies truth after all), but the point remains.

> By no means do I agree with much of what is on the World Socialist Web Site (WSWS)

I also could have my nuances with some things said there. But I don't think there is so big of an ideological breach among people if the debate is proposed from a pragmatic and resolutive standpoint. Wouldn't many of us agree that a government should control or manage some resources and give free land for innovation in others? Couldn't we debate about what would be the most efficient way to organize a public governing body, area by area, and how?

> However, from a purely algorithmic standpoint, their website has content that should rank very highly

But why? None of those things are reasons to link to it for any particular terms, just reasons not to not link to it.

> from a purely algorithmic standpoint

You could use the same argument on any extremist site.

Algorithm can parse content. It probably has list of bullshit related keywords. Just the title:

> Marxist analysis, international working class struggles & the fight for socialism

Extremist is the tag that Putin bolts on every opinion he doesn't like, so now I may assume you are a putinist.

Website that constantly talks about 'struggle' and 'revolution' is not extremist at all.

"Extremist" is a weasel word for when a person didn't really do anything yet, but you are already trying to censor them.

Marxists and socialists did not do anything yet. Now you are just trolling.

These particular ones didn't, and you certainly can't prevent a person (or a group of people) from thinking improper thoughts as long as they want to.

But you would like to, won't you? You would like to get into people's heads? You consider people your peons and you want to pull switches in their heads because you know better what they should think?

In short, you are a putinist.

How are these 'bullshit'?

Reply on news.ycombinator.com