Frankly, it does not sound very hard at all. They have to write 20 posts a day, but each post is only a headline and a brief summary. A focused writer can finish that in 15 minutes.

> We had to write in the most passive tense possible. That’s why you’d see headlines that appear in an alien-esque, passive language.

Oh no! How dare Facebook strive to be neutral and passive.

> After doing a tour in Facebook’s news trenches, almost all of them came to believe that they were there not to work, but to serve as training modules for Facebook’s algorithm

Of course they were. Facebook is quite explicitly trying to apply ML throughout their site. Why should they permanently be in the business of employing writers to do something which computers could do reliably and effectively?

I was just thinking that the Facebook summary headlines are surprisingly journalistic compared to Buzzfeed, Upworthy, etc. Seeing somebody complain about that explains why their title is "news curator" and not "journalist."

[1] https://en.wikipedia.org/wiki/Portal:Current_events

It might be somewhat working with printed media and specialized news agencies — after all, agency wants something close to what the journalist wants: to have presence, for public to pay attention to them, not to some other agency. "No news today" would be a disaster for them.

I mean, it's all pretty natural, if we admit that typical news agency's goal is not to "give information". And Facebook's "trending news" section's goal, on the contrary, is quite close to that.

More on topic though, the comments section of the article is also full of "booohoo they have it sooo tough" comments but I think work conditions are relative. It's the typical job agency treatment only a couple of levels higher. If everyone around you is living the Facebook life and you're curating news in a basement/conference room...sure that's not horrible but it's still fairly indecent treatment.

Disclaimer: Huge Transmetropolitan fan, programmer not journalist :)

1. http://scholarship.law.berkeley.edu/cgi/viewcontent.cgi?arti...

They knew they were signing up for contractor position. Frankly, they were lucky to get any benefits at all.

The article fails to provide any evidence of abuse though. The worst that Facebook did was not inviting them to company happy hours.

Also, in case it wasn't clear, I have in fact worked at several media companies directly with journalists. I'm not speculating from some distant position about the attitudes of journalists.

If you have an actual point, I'd love to hear it. I don't see how writing basic summaries of the news is a valuable or artistic endeavor.

Everyone was happy to have $30 jeans which could last one year over their $70 jeans made stateside which could last five years or more... But now that it affects them, oh my, this is bad now...

My only piece of designer clothes was a $160 pair of jeans that lasted for only 3 or so years. I concluded that from now on I should stick with buying $50 jeans which last me at least 2 years.

I'm not sure this analogy works. Surely the people selling jeans would prefer to sell new $30 pairs every year rather than $70 pairs every five.

Any company that did that wouldn't make it to tomorrow.

Doesn't this imply that the customer has enough information to determine which is the superior product? In reality, it's quite easy to dress up an inferior product with lots of marketing, and gain great success in the marketplace. Heck, look at Beats headphones. They're not superior in any way. Yet, by carefully crafting a marketing message, and putting massive amounts of money behind it, they've managed to get a fairly large share of the market.

If you ask a consumer whether they want a $500 product that lasts 10 years or a $200 product that lasts 2 years, they'll happily pay extra for the higher quality product. But that's not usually the choice they're presented with.

I think a solution to planned obsolescence would be to mandate prominent labelling of lifetime guarantees.

That used to be true and is an example of information asymmetry in markets - classic example being the used car market [0]. It is much less so now, when was the last time you bought anything substantial without reading a review of it or checking the customer ratings?

>If you ask a consumer whether they want a $500 product that lasts 10 years or a $200 product that lasts 2 years, they'll happily pay extra for the higher quality product.

That's not necessarily true - value decisions are more complex than that. Depending on the market, those two price points may be in different segments based on (e.g.) disposable income. Sure, for Alice, who has $500 disposable income available for a particular purchase, this might be a no brainer, but for Bob with only $200, not so much. See also Vimes' boots theory of socioeconomic unfairness [1], hyperbolic discounting [2]

[0] https://en.wikipedia.org/wiki/The_Market_for_Lemons

[1] http://www.goodreads.com/quotes/72745-the-reason-that-the-ri...

[2] https://en.wikipedia.org/wiki/Hyperbolic_discounting

In some cases, you can simply conclude that an entire brand is either unreliable or trustworthy (e.g. Apple doesn't sell Walmart edition devices which break within 18 months) but in most cases you have to do a lot more research to know what level you're getting.

Besides the already mentioned point of reviews being made too soon to be able to reflect on product lifetime, keep in mind that easily available (e.g. on the shop site) on-line reviews are mostly bullshit anyway. It's mostly a mix of people being affiliated with producers or paid to endorse a product, with an occasional person which is too dumb to operate something correctly but quick to drop a 1-star rating with a nonsense explanation. To get any sensible reviews, you actually have to know some trustworthy reviewers - be it a site you follow or a friend you know. Either way, maintaining sources of trusted reviews requires some effort, so you won't be doing that for most of the product you buy anyway.

But reviews don't tend to judge product lifetime that well, because they're usually written within the first month a product comes out.

I think the only way you can get reliable information on lifetime is by forcing companies to provide it, and guarantee it. (Or put a big sticker on their product saying "0 years guaranteed")

>Depending on the market, those two price points may be in different segments based on (e.g.) disposable income.

I feel like I see a lot of reasonable sounding financing in the markets I'm thinking of, but perhaps I'm remembering times before 2008.

Of course they're doing that. They get away with it, because a) everybody else is doing that too, and b) they know that throwing money at marketing has much better ROI than making a good product. A little bullshit can paper over many faults.

Trump for example got tons of coverage. Is he "superior" to someone who may be more knowledgeable, skilled and would "make a better president" but got less coverage? The amount of advertisement, exposure, and just plain psychological manipulation matters.

Is sugary soda superior to healthy drinks? Is bottled water superior to tap water? So we as a society really need to throw away metric tons of non biogradable materials every minute?

Thinking that something wins out because it is "superior" is just confirming the free market ideology - whatever wins must be "superior" is a tautology which makes the word "superior" lose all meaning.

Yes, lots of crappy products beat products that performed better in practice, because of external and network effects.

But what we should be asking is about the correlation between a company's incentives and polluting the planet. In the main, they are pretty aligned, which is scary.

Anyway, most people have adopted the fashion cycle --something that only wealthy people would import from the fashion capital du jour.

See: H&M, etc.

Oh lord. That's not news. Those are sensationalist, click baity headlines making us all collectively dumber. It's anecdotal, but I'm quite sure the 'news' my girlfriend has gotten from FB has made her less informed.

I actually wrote an FF plugin to hide that right bar, since I would now and then get sucked in by that crap and waste 15 minutes reading about Lindsey Lohan or whatever.

Or what if instead of national figures FB or anyone else tried to railroad an SF Supe they didn't particularly like because they shot down their proposed development. Or Menlo Park.

I personally choose to not use Facebook specifically for these reasons. Even though I know many Facebook employees personally, I don't trust it's employees to divorce their own interests and their powerful position as Facebook's curators.

  They were also discouraged from mentioning Twitter by name in headlines and summaries, and instead asked to refer to social media in a broader context.

So anything they might do which appears to be editorialisitc in nature regarding politics would be of concern.

Which is a completely different thing than manipulating their users, which they shouldn't be doing for any reason.

In the end though, you're basically just playing at the margins.

This could be already describing what you are suggesting.

The types of trending topics that are blacklisted tend to be low quality trends. Say #FollowFriday, or take examples from the current Twitter trends, #mondaymotivation, #MondayMorning, #DressAFilm, or #MondayBlogs.

They also will blacklist a trend when it's a duplicate of another trend say #GoodWifeFarewll vs #TheGoodWife. In fact, they'd probably try to combine those two if the clustering algorithm failed, and then relabel it "The Good Wife" since it undoubtedly corresponds to a FB page.

The manipulation of the candidate set of trends is there because the trending algorithm and the clustering algorithm is too noisy to publicly surface. The trends you actually see are algorithmically chosen based on your behavior.

http://thehill.com/policy/technology/279192-report-facebook-...

http://gizmodo.com/former-facebook-workers-we-routinely-supp...

Well, there you go. Quite frankly there's a neutral editorial voice that FB wants to project. And so the main article has to be from a legitimate unbiased news source. Brietbart, Newsmax, and the like simply aren't.

Having a list of approved sources to highlight is really common I these types of products, because no one wants Stormfront to be on the front page.

Of the topics don't get out of the ghetto of subculture's echo chamber, it's never going to be approved. Even the topics cited are basically just birther conspiracy du jour.

Even if the topic gets approved with a different main link, everything is being shown on the results page, just not with a gold star next to it in the top slot.

Personally I would regard Breitbart or Newsmax are just on par with HuffPo - for you they probably aren't. That's why pretending to know what others have to perceive as "legitimate unbiased news" is such a delicate issue that should not be done in secrecy.

Non-transparency on blacklisted sites and topics will always lead to distrust and the expectation of one-sided political shilling.

Sure you could publicize your approved list of sites, but then you're just going to attract a bunch of people with an agenda to trying to promote their agenda by claiming some sort of slight by not being included. In the end though, the only honest response to this criticism is, "Screw you. Make your own site."

Topics aren't blacklisted except when they duplicate something else. Most topics simply aren't approved for trending. If you want to know what's not approved, look at Twitter trends, it's full of junk. Honestly, that's the best public example of what the trending topics algorithm detects.

You really need to read Tom Stocky's post about this, because that's exactly how it's run. But you don't really have to, because it's exactly what I told you.

Finally, if you actually click on any of the trends, pretty much everything on that page is algorithmic. Find a trend that mentions, Obama, I'm sure you're going to find a racist post on that list, because there's always scummy people.

PS. For the record, I always feel dirty if I visit HuffPo. Gawker on the other hand, I'll own up to reading that tabloid.

That would imply Facebook's influence doesn't exist, and its $340B valuation is a sham.

I guess it's plausible to me that a new company like FB lacks some sort of important company culture for reporting that the NY Times has by virtue of its long history, but this seems minor. The real issue is (1) the consolidation in the news industry and (2) its political homogeneity. FB and the NY Times are each examples of both.

"The fifth experiment is especially notable in that it was conducted with eligible voters throughout India in the midst of India’s 2014 Lok Sabha elections just before the final votes were cast. The results of these experiments demonstrate that (i) biased search rankings can shift the voting preferences of undecided voters by 20% or more, (ii) the shift can be much higher in some demographic groups, and (iii) search ranking bias can be masked so that people show no awareness of the manipulation."

[1]: http://www.pnas.org/content/112/33/E4512.abstract

The goal shouldn't be to "stop Trump" — that's not what we want a news distributor to do. I don't want bias against a particular person in my news distributors. What I want is for my news distributors to actually have ethical standards that enable them to serve the public good (and as a by product hopefully prevent lying demagogue's from gaining power).

Is it "bias" to use a standardized method or algorithm (even one that involves humans reading natural language") to filter content by accuracy and provide some interface indication not just for phishing, and for porn, but also for falsehoods?

In fact, decisions which Facebook makes this election season will impact the result regardless of the intentionality behind those decisions. Even something as mundane as the mix between links and images impacts different candidates disparately.

That said, I absolutely think that Facebook and other major media outlets should be held to a form of the Equal Time rule:

https://en.wikipedia.org/wiki/Equal-time_rule

The fact that we don't have an equivalent rule for major websites is a serious problem.

Screw 'em. If they want to play the corporate welfare game, they can abide by the equal time rule.

As for taxes, what taxes do you think were going to be assessed on this property? It was an empty piece of land before FB arrived so it was not like the city was losing much property tax to begin with. In return Facebook is paying salaries for forty or so people on site and the thousand or so who worked for over two years building the site. Oh, and because tech firms like it when others do the research and general site selection you will notice that Microsoft is dumping tens of millions into a data center in West Des Moines as well. You're welcome.

You "allowed" it because it was a good deal for a city like Altoona that otherwise was not much more than an amusement park and the last place to get gas on I-80 before you hit Des Moines.

Can you expand on this a bit? Do you mean to say when you turn on the tap in the kitchen nothing comes out and that you are no longer able to use the shower or flush the toilet?

It is crazy how much money we just give to profitable companies.

Progress hey, it's tricky.

They started the profile badges thing and folks thought that was so cool.

I had to explain to them that this is one more data point for Facebook - if I change my profile badge to a Manchester United logo, it probably means that I really, really care about Manchester United.

I can then be targeted by an advertiser that wants to sell merchandise to Manchester United fans

And why is this itself a bad thing?

Poe's law or whatever, but I can't even imagine a worst possible candidate than Trump, even Eric Cartman would be a better president than this Daily-mail-esque cartoon of human being.

Strange that people at Facebook would feel apathetic toward individuals that can't code.

It's pretty common elsewhere in SV as well. And is a cardinal error.

If Facebook is doing it's job and utilizing their biggest talent acquisition of Yann LeCun, then this is exactly what they should be doing.

A critical part of the path towards AGI is using humans to teach it.

I don't think that the way knowledge transfer is structured for humans is by default distinct from how we would do it for AI in the long term, but that's a pretty wide range of pedagogy so some tasks make more sense than others for the state of the art today.

In fact the way we train ANN/CNN is similar to how infants & toddlers do early classification. Frank Guerin at the university of Aberdeen has done some study in this area that is in it's early stages.

As far as I know there have not been efforts to set up schools to train AI's from the general population.

The closest thing to it are the efforts by Google and Facebook to train their classifiers (facial recognition, text detection etc...). We are working on an internal process that uses user input to classify objects in the home (furnishings mostly) for auto segmentation and detection.

3rd paragraph:

> grueling work conditions, humiliating treatment, and a secretive, imperious culture in which they were treated as disposable outsiders.

6th paragraph:

> “It was degrading as a human being,” said another. “We weren’t treated as individuals. We were treated in this robot way.”

And then, finally, on 10th paragraphs, we get a glimpse on the facts:

> they received benefits including limited medical insurance, paid time off after 6 months and transit reimbursement

(BTW, is it usual for contractors to receive such perks?)

> A company happy hour would happen at 8 p.m., and we’d be working

Horrible, inhumane treatment indeed.

> Over time, the work became increasingly demanding, and Facebook’s trending news team started to look more and more like the worst stereotypes of a digital media content farm. Managers gave curators aggressive quotas for how many summaries and headlines to write, and timed how long it took curators to write a post. The general standard was 20 posts a day.

20 posts during 8 hour work day is almost half an hour on one post. It is considered too little? Seriously?

So — apart from all the pretty words, I didn't really see any especially bad treatment. Hell, I'm pretty sure that your average newspaper employees have more nightmare stories.

[0] https://en.wikipedia.org/wiki/Supervised_learning

disclaimer: i worked on that fb product.

this is the most worthless piece of thing i've ever read in my life