Live data from Hacker News

Diseconomies of scale in fraud, spam, support, and moderation

danluu.com

21–30 of 216 posts

Re: Diseconomies of scale in fraud, spam, support, and moderation

#21
post #12

Way back in 2014 when Jeff Atwood (aka codinghorror) switched from Stack Overflow to creating Discourse, he gave a talk about it (see notes [1], sadly I can't find a live link to the recording anymore). He gave a pithy little explanation of why they built the Discourse trust levels system the way they did that stood out to me: > The only thing that scales with the community is the community. The point being, you have…

I wonder how that would work for Facebook, for example. Facebook Groups have admins, but for general posts to the public (or to friends) - could they be user-moderated?

This strategy seems to be especially effective for focused communities as they grow from 50-5000 users. As you approach facebook scales and a more generalized audience ("the public") I'd imagine the optimal strategy may shift some. A federation (small "f") of smaller, more focused communities might be the best you can hope for.

Re: Diseconomies of scale in fraud, spam, support, and moderation

#22

The author seems to be mistaking the benefit of selection effects with competence. Those who work anti-fraud for platforms know that there is something like a market for fraud. There is a market clearing quantity of fraud that will be present on any platform, dictated by the gains from learning to execute the exploit, the effort necessary to do so, and the number of times one might be able to do it. Small platforms h…

I don't see any mistake.

> Small platforms have way less fraud because they are way less valuable to exploit, in general.

Yes.

> The big platforms actually are radically more competent at anti-fraud / anti-spam / etc, but the scope of their operations make them such juicy targets that there is a lot of it anyways.

If their business model (scale above everything) results in shitty customer experience I don't really care how competent they are. They are fighting the problem they created themselves: a platform too valuable for scammers.

Re: Diseconomies of scale in fraud, spam, support, and moderation

#23
post #8

Amazon selling obviously fake SD cards / flash disks is baffling. It’s super easy to spot and fake disks cost users data, so you’d think they’d take that seriously. https://a.co/d/auXcxrZ ($29 for 1TB as their “overall pick”)

I imagine they make a ton of it back from users who never need anywhere near that much space.

I have the non-technical people of my life insist on a terabyte for everything. One of them bought an iPad Pro just to have a terabyte of space. After 8 years of their prior iPad, they had used up all the space (64gb I think) and wanted to be sure they wouldn’t run out.

Re: Diseconomies of scale in fraud, spam, support, and moderation

#24
There's a lot to unpack in this article, but the key idea is the "myth of big" and that small is beautiful. It should be no surprise that Big-Tech succeeds in part by creating its own powerful mythology. All giant organisations, nations states too, build myths of unique competence, infallibility, and exceptionalism. At the same time it hardly needs saying that scale brings inefficiency, corruption, sluggishness, and insensitivity. If it weren't for protectionist measures in place to sustain big monopolies they'd be torn asunder in real competitive conditions.

Re: Diseconomies of scale in fraud, spam, support, and moderation

#25

The author seems to be mistaking the benefit of selection effects with competence. Those who work anti-fraud for platforms know that there is something like a market for fraud. There is a market clearing quantity of fraud that will be present on any platform, dictated by the gains from learning to execute the exploit, the effort necessary to do so, and the number of times one might be able to do it. Small platforms h…

OK, fraud exists.

Irrespective of whether or not a platform is small or large, can they make their customer whole 100% of the time? Why or why not?

Re: Diseconomies of scale in fraud, spam, support, and moderation

#26
The claim that small platforms are better at moderation needs some nuance.

In terms of content that is annoying or worse to the forum’s users, it’s probably broadly true. But it’s not necessarily true for content that the users of the forum are fine with, but society more broadly has issues with.

Take for instance forums which are devoted to sharing “revenge porn”.

Re: Diseconomies of scale in fraud, spam, support, and moderation

#27

What all this says to me, in total, is that greater competition where you have a bunch of potentially medium sized companies or smaller but no clear massive (50%+ market share) leader that power can’t congregate means they have both incentive to provide a better experience and having multiple targets makes it harder on negative actors, effectively. Thats my takeaway here is that we should as a matter of regulation /…

It is the general consensus that competition in the market place is good for consumers and propels business to find creative solutions that create value in the econony. I would say that isnt always the case, Rockefeller had something like 90% market share before litigation of antitrust began, by the time the case was adjucated their market share had already slipped; what this indicates is that even when one party holds 90% of market share (and probably total wealth for that matter), the market will eventually cut them down to size. At the same time, rockefeller was responsible for bringing a groundbreaking innovation to society. So, it was good that initially all that capital flowed to him, because he was using it to move society forward (at least in this regard), and it was also good to see that capital flow away on its own, because it was needed elsewhere.

I would also point out that the centralization of companies into todays megacorps is influenced by many factors, economies of scale may be one, but so are intellectual property laws (we really dont need them!). In fact, I would venture a guess that most regulation is written by the companies that are being regulated, that keeps competition at bay, and they wont regulate themselves downwards in the food chain. Finally, I would say the centralization really took place during WWI/WW2, it was convenient for the government to shout orders at a whole 3 people to build tanks instead of automobiles; this centralization is a strategic goal of the united states government, and enacting it through secret deals (rather than law) is a preferable method to preserve state secrecy.

Re: Diseconomies of scale in fraud, spam, support, and moderation

#28

Earlier quoted context omitted.

Not really, because the entry barrier to create a small town taxi service is incomparably lower than creating a new Uber. 20 years ago my small town had around five fiercely-competing taxi services with cars arriving within 4-5 minutes of a phone call. In many respects that was better than the new times. Especially when it came to customer service.

How small a town are we talking about? I grew up in a 100,000 person city and we only had two taxi services that would take ages to arrive, depending on where you were.

Around 100k. It was in Russia though and owning a car wasn't as common as say in US. Public transport wasn't very pleasant to use either.

It was an interesting case, a sort of a breakthrough in transportation habits achieved by loads of 15-years old right-hand-drive Toyotas converted into taxis. The cheapest ride was around $1 and people literally started riding taxis to the next building on the same street.

Re: Diseconomies of scale in fraud, spam, support, and moderation

#29
post #12

Earlier quoted context omitted.

I wonder how that would work for Facebook, for example. Facebook Groups have admins, but for general posts to the public (or to friends) - could they be user-moderated?

This strategy seems to be especially effective for focused communities as they grow from 50-5000 users. As you approach facebook scales and a more generalized audience ("the public") I'd imagine the optimal strategy may shift some. A federation (small "f") of smaller, more focused communities might be the best you can hope for.

Has Facebook perhaps shifted towards a federation of communities?

My Facebook feed as it used to be, general life updates by various friends, is pretty much dead. Nobody posts these things anymore. However, neighbourhood groups, Buy Nothing groups, and other local special interest groups thrive, I see several daily posts in each. Plus Facebook constantly recommends me more groups. This is a n=1 experience but I wonder whether Facebook is 'falling apart' into these groups.

Re: Diseconomies of scale in fraud, spam, support, and moderation

#30

Earlier quoted context omitted.

> If I'm uber I just ignore it. It has nothing to do with efficiency, it's just getting the monopoly power to treat your customers badly and get away with it. People don’t use Uber because they’re a monopoly. They use it because it’s convenient, it’s a known quantity, it’s mostly predictable, and they can use it almost anywhere they go. People like to complain about Uber, but my pre-Uber experience with taxi services…

> People like to complain about Uber, but my pre-Uber experience with taxi services was far worse Did you live in a small city back then?

I’ve used plenty of small-town taxi services, yes.

If you’re only remembering the gold standard small-town service that did everything perfectly, you’re looking back with rose colored glasses. The typical experience back then was not good.

Post reply on HN