Re: Why we really can't use Facebook for technical discussion.

[Phillip Hallam-Baker <[e-mail]>]

On Mon, Jun 7, 2021 at 10:31 AM Keith Moore <[e-mail]>
wrote:

> On 6/7/21 8:16 AM, Phillip Hallam-Baker wrote:
>
> What we have here is the predictable result of a company that failed to
> take moderation seriously and is now desperately throwing technology at a
> problem rather than fixing the core problem that they designed their
> environment to maximize conflict because that was most profitable for them.
>
> As much as I hate FB (I left the platform in 2016 and have never looked
> back) I think "failed to take moderation seriously" glosses over a number
> of inherent problems with social media, particularly when done on a large
> scale.
>
> One is of course that human moderation is time-consuming and therefore
> expensive if the moderators are paid.  It's also hard for a large number of
> human moderators (required to deal with large volumes of users and traffic)
> act uniformly.   On another widely used platform the moderators are almost
> completely arbitrary, despite supposedly enforcing a common set of clear
> standards.   So it's not surprising if social media platforms resort to
> algorithms.   And of course the algorithms are flawed because AI is a long
> way from understanding the many subtleties of human interaction.
>
> Unpaid human moderators can be even more capricious than paid humans,
> because the desire to impose one's own prejudices on others is a strong
> motivator of volunteers.
>
> Even under the best of conditions moderation (whether done by humans or
> machines) is dangerous both because it can easily squelch valuable input,
> and because it's often easily gamed for that purpose by people for whom
> such input is inconvenient.     No matter how noble the intent, the effect
> of moderation is often to favor established prejudices, or worse, to enable
> bullying.
>
> I don't claim to know the solution, but I don't think it's a simple matter
> of "taking moderation seriously".
>

The main reason I am on the platform is to try and understand which
approaches work and which do not.

Slashdot has a rather less respectful audience yet their karma system
produces vastly better results. But the biggest takeaway for me is that one
size fits all does not work for curation or moderation.

Yes, this is a hard problem, that is why there should be more than one
person who is the sole decider. I have never said anything on Facebook that
would not pass in an Oxford Union debate[1]. We have known that content
inspection is a failing strategy for spam filtering for over 15 years, why
apply it to enforce automatic bans on individuals?

PHB


[1] OK that might be considered a low bar since I once heard Jacob
Rees-Mogg addressed as 'Greased Frogg' but who is to say such sobriquet is
truly undeserved?