Hacker News Reader

Comment by nindalf | Hacker News Reader

I did touch on that problem. Like I pointed out Zuckerberg can’t log in on new devices anymore. That’s because of the thousands of attempts per second to log into his account. Those attempts happen because he’s a celebrity. His experience is objectively different because of who he is.

It’s the same with Neymar. How many times do you think his profile is reported for any number of violations by people who don’t like him? If an ordinary persons account got 100 reports a minute it would be taken down. Neymar’s won’t be.

I don’t know how every Integrity system could be modified to make an exception for any of these classes of accounts or how to codify it in a way that would seem “fair”. If you have an idea for a better way, you should share it.

View on HN

> It’s the same with Neymar. How many times do you think his profile is reported for any number of violations by people who don’t like him? If an ordinary persons account got 100 reports a minute it would be taken down. Neymar’s won’t be.

More to the point, after a human reviewed Neymar's conduct and it clearly violated Facebook policies about posting revenge porn, his account still wasn't taken down. And that's not a technical issue of the false positive rate -- it is a double standard.

From the article: 'A December memo from another Facebook data scientist was blunter: “Facebook routinely makes exceptions for powerful actors.”'

(Edited: changed verb tense to make it clear that this already happened and wasn't a hypothetical.)

This may be a naive question, but isn't that a problem with the content reporting system itself? That it requires blanket exceptions?

Popular users are going to have more false-positive reports than others, but when those reports deviate from the norm on individual pieces of content (say a nude photo) then the system should still be able to pick it up. It's an exercise in feature extraction.

The login blanket exception (for the CEO of the company) is a different use case and purpose than content control, one that blanket exceptions can solve efficiently.

>If you have an idea for a better way, you should share it.

I'm certain that the other poster won't present some kind of practical outline because part-way through the brainstorming they will realize they are designing a "Social Credit Score" and become too frustrated to continue.

The solution is simple. Facebook needs at least 100,000 human moderators. Clearly their engineering team isn't able to solve the problems facebook creates.

I concede that some accounts probably deserve that specific type of protection. However, it doesn’t explain the other kinds of protections these people have, including exemption from content-based rules. Those are the issues of real concern.

The fact that the content couldn’t be taken down was a bug. I feel like you, the author and most critics of Big Tech discount the possibility of bugs existing.

I’m sure you’ve read the old fable of “The Boy Who Cried Wolf.” Facebook has made voluntary decisions at the highest levels, reported over the course of the past decade, to shield certain people from its content rules that apply to the general public. So if it was a bug this time (and I have no reason to believe that it wasn’t), I’m sure you can understand people’s skepticism about it.

Moreover, I assume this bug was reported internally, probably pretty quickly. How long did it take to get fixed? If the fix wasn’t prioritized and corrected within, say, a day (along with a regression test to ensure it never happens again!), then that would be pretty damning of the company’s culture and priorities as well.

But facebook comes across as unwilling and uninterested to fix the bugs in cases like these. Sometimes, even, the bugs themselves seem to surface resulting from fundamentally misstructuring the problem.

Maybe normal users also shouldn't be able to have their accounts destroyed by a hundred spurious reports.

Currently if you register a Facebook account to manage a business and post ads, it will be banned off and on for weeks, and the recourse suggested to me by a director in the Integrity org was "try posting on facebook like a normal person."

> If an ordinary persons account got 100 reports a minute it would be taken down. Neymar’s won’t be.

Why does this disparity exist? By your own account, the number of times this happens for normal users is significantly smaller than for high-profile users, so why is Facebook incapable of having sufficient staffing to deal with this case for all users? _This_ is what has a lot of people annoyed.

> Those attempts happen because he’s a celebrity. His experience is objectively different because of who he is.

Then the rules should be different, like Twitter giving a blue check mark. If there are accounts that need to be treated differently then it should be clear why and how the rules are different. Fixing problems with tech (like not allowing Zuckerberg to login) should be the exception.

Twitter caught hell for letting Trump break their rules because they pretended the rules were the same for everyone.

I'm not sure what you're point really is. You keep harping on Zuckerberg not being able to log in on new devices but dismiss the entirety of the report and the internal review as "Yeah, well, at scale there's nothing you can do." If that's the case, shut it off. We wouldn't accept that from a manufacturer of a physical product.

Shut what off, exactly?

> We wouldn't accept that from a manufacturer of a physical product.

Not sure what the correct analog for physical product is, but we accept everything up to and including catastrophic failure resulting in deaths from physical products under the right circumstances, so you're going to have to be much more specific.

> If that's the case, shut it off.

It should be obvious why "Shut Facebook off" is not in the domain of potential solutions Facebook will entertain to this technical challenge.

No, but governments might.

"We wouldn't accept that from a manufacturer of a physical product." - why not? I would think that any manufacturer of a physical product is clearly entitled to provide better service or better product versions or better legal conditions or better pricing for some of their customers if they want. And they definitely do so for various VIPs - some people pay to wear Nike shoes, and some people get paid millions to wear Nike shoes, and we do consider that a manufacturer of a physical product has the right to do that.

There is not a trade principle of having to treat all customers equally (the sole exception being not denying service for a specific list of protected groups), and there is a general principle that people can trade with others as they please and provide different conditions for arbitrary reasons.

Reply on news.ycombinator.com