Hacker News Reader

"Newspapers optimize for engagement by any means necessary. That's how their incentives are aligned. They need subscriptions to sell to advertisers, so all market tests and all headlines are constantly being changed to increase the amount of people who read their articles."

"Magazines optimize for engagement by any means necessary. That's how their incentives are aligned. They need subscriptions to sell to advertisers, so all market tests and all covers are constantly being changed to increase the amount of people who read their articles."

"Television stations optimize for engagement by any means necessary. That's how their incentives are aligned. They need viewers to sell to advertisers, so all market tests and all programming is constantly being changed to increase the amount of people who tune in."

"Radio stations optimize for engagement by any means necessary. That's how their incentives are aligned. They need listeners to sell to advertisers, so all market tests and all programming is constantly being changed to increase the amount of people who tune in."

Yes - absolutely this.

Earlier the media market was able to make the case of "free speech" and we had no way of arguing that this cycle was true or not. At the old rate of content, the harms were less prominent and the system seemed to balance itself out (although the behavior of certain media outlets indicate that it was a problem even then.)

Markets force optimization, and for media firms the optimization force is engagement. This has been a problem from before and its the same problem now, except at a content sharing speed which the old arguments were not built for.

I see the argument repeated quite a bit. I think it's easy to read this and think "Nothing new, just more of the same, maybe a little worse now". I don't think that's quite right though. A television station can only broadcast 24 hours of content a day. A newspaper can only have so many editions per day, and each edition has a limited number of pages. The same is true of television and radio. But the common factor among all of them is that there is a common reality that everyone is looking at, and there is scarcity because of that: the publication has to choose what will appear each day to the readers/watchers it serves.

The ideal formula is to draw a large group of people in with really engaging content, and then show them advertisements that are likely to be effective for the group of folks drawn to the content. This presents a three-way tradeoff:

1. Write about stuff folks want to read about (but don't make it so outlandish that you ruin your credibility!)

2. Show effective advertisements alongside that content (but not so much you drive them away!)

3. Show enough variety to draw in a large crowd (but still know your target demographic!)

Newspapers, magazines, television, and radio all faced this dynamic and had to make the right tradeoffs to survive.

But there is no common "Facebook" published each day. We can't say "Hey, did you see the top story today on Facebook?" Each article can target a tiny segment of the population and be designed just to keep only those folks engaged. And there's no need to tie a particular advertisement to a particular piece of content, advertisements and content can be selected in realtime based on models of the user's behavior. And drawing in a large crowd is easy, since there's content and advertisements for every tiny group that any entity is interested in targeting. The scarcity that forced reasonable tradeoffs in other mediums simply isn't present on social media. This is a huge difference that isn't captured at all in the "optimize engagement by any means necessary" argument.

Put another way: the problem isn't that incentives are different, it's that the tools available now remove scarcity that previously shaped the entire industry.

I think the core difference is the level of tailoring and the automation of that tailoring. Newspapers, magazines, television, and radio have all had the same incentives in the past, but architecturally the approach was one to many. At first the internet was more of the same, but today social media companies construct optimized engagements for each individual, automatically and very effectively. This is a one-to-one optimization.

Furthermore, social media supercharges the engagement by piggybacking on our social nature - people are usually responding to stuff their friends posted or commented on. This both allows disinformation to be received as more truthful (my friend sent it, why would they lie?) and simply makes it more compelling (I get to be happy/sad/outraged with my friends).

Put it all together and it really is different than before. It's fine to compare it to historical changes in media. I don't agree that it's just more of the same though, the differences make this a whole new animal.

I agree with that.

However, nothing is preventing Facebook from doing this. Facebook is incentivized to act as you describe. I do not see it as Facebook's responsibility to act in another way. Governments need to regulate this behavior.

If people start to understand the way Facebook is manipulating them, then there is also an opportunity for another platform, which does not do so, or to a lesser extent.

You don’t see it as Facebook’s responsibility?

To not actively push malignant content just because it gets views?

Everyone in this world bears responsibility to one another. Of course there should be regulation to prevent the circumstance, but to the extent we want a livable world it is also all of our responsibility to not be degenerates.

Facebook is being a degenerate and shrugging it off as “well it’s not MY responsibility to not be a degenerate!” feels hopelessly lost and without any critical introspection or thought at all.

Yes, I do not think so. I expect Facebook's goal to be to maximize shareholder value within the bounds of regulations.

I do not expect an entity such as Facebook to behave in any other way.

From Facebook's perspective, the current value system is possibly distorted, for example, because of a lack of competition. Suppose another social platform would enter the picture, and users would flock to that platform because it has a greater social responsibility. In that case, it might become Facebook's best interest to be more socially responsible as well.

However, that has not happened yet. It is making me believe that the burden to solve this issue is with regulators, not Facebook.

Regulatory capture is a thing.

Consider an aircraft company that wants to sell comparably cheap but dangerous products that cut corners and have hacky solutions to core engineering problems. Short term profit for shareholders is the only thing that matters, forever and always.

The aircraft company has three categories of options:

A) don’t murder people and take the time and energy to get it right. Maybe they’re less competitive some years than others

B) The aircraft company can be a ‘neutral’ actor and still try to cut corners, but leave the regulator alone to actually do its job. Note this is in the company’s long term interest because the success of the air travel industry depends on its long term track record of safety and consumer confidence that planes are thoroughly designed to not smash themselves into the ground at 500 mph.

C) The aircraft company can lobby heavily, get special government treatment, corrupt the regulator, and arrange things such that they can sell the deadly plane and try to get away with it. It’s all about the Benjamins, as they say; just not any Benjamins in the plane of course.

Now imagine that aircraft company not only has the money to lobby aggressively, but also controls what many, many consumers see on a daily basis, can show advertisements to the entire population for free, runs algorithms that promote ‘engaging’ (shocking, manipulative, emotionally salient, regardless of accuracy) content.

And when people say "Facebook" the really could (and maybe should) be using "Mark Zuckerburg" instead. He has complete control of Facebook and no one can remove him from that position. Mark Zuckerburg is responsible for what Facebook does more than almost any other CEO of a public company, I would think.

Well what about this:

Malignant content is very hard to define. If I shared videos of mexican gang violence, people would end up calling Facebook headquarters and saying "get this off my feed.".

Even though the content getting visibility is a success of overcoming local Mexican censors who are silencing voices talking about the issue.

This happened. Mexican videos went viral, and when people in the UK woke up, they called their friends at FB and the material got removed.

Also - Wouldn't Any algo that starts filtering for malignancy, have to become a sort of MR. Rogers algo?

I guess, how would you define malignancy? By intent?

I, for one think that, "understanding manipulation" is an oxymoron. If you really understand then it's not really manipulation, right?

All these web platforms have grown through strategic optimizations for engagement. It's simple. But the tactical mecanisms they use to achieve this permanent engagement are complex and adapt constantly. Otherwise they wouldn't be good for the purpose. The point is not to keep you engaged once. It to keep you engaged at their (platform's) will.

You are right that it's not Facebook job to be less good at what they propose to do. But it is everybody elses job to determine if we want them (and others) to keep doing these things.

>I, for one think that, "understanding manipulation" is an oxymoron. If you really understand then it's not really manipulation, right?

It sounds like someone never teared up at the moment when the music swells in a sappy dramatic movie. The human brain is weird. We can know we are receiving a placebo and still receive a benefit from the placebo effect[1]. Just because we can understand that we are being manipulated doesn't mean we can fully control it.

[1] - https://www.apa.org/monitor/2011/03/placebos

> If people start to understand the way Facebook is manipulating them, then there is also an opportunity for another platform, which does not do so, or to a lesser extent.

This assumes that a platform which behaves more ethically or is better aligned with the public good can actually compete financially with a company like Facebook. If Facebook financially benefits by not acting in the public interest, what makes you think a competitor who is acting in the public interest is able to compete? The whole point of the problem is that being a bad actor is incredibly profitable in this case.

This statement:

> nothing is preventing Facebook from doing this. Facebook is incentivized to act as you describe. I do not see it as Facebook's responsibility to act in another way.

and this statement:

> If people start to understand the way Facebook is manipulating them, then there is also an opportunity for another platform, which does not do so, or to a lesser extent.

are inherently contradictory because Facebook is explicitly incentivized to misbehave, as you mention, meaning that any competitor trying to compete whilst acting in the public interest will never succeed. If acting in the public interest was profitable, Facebook would do it.

You're right, it was wishful thinking of me to assume that a competitor would be morally responsible and more profitable than Facebook.

> The problem is that the social media architecture as it is, is optimizing for engagement by any means necessary

The problem with the problem is that seems to be what people actually want by definition

You could say the same about gambling, which is something else people actually want, but society sees fit to regulate due to its second-order effects. (like gambling addiction destroying families)

Right, we should just skip the line and go straight to fentanyl.

Sometimes what people want, and what will keep society from coming apart at the seams differs.