To clarify, can you specify exactly what law you would like made? What do you want to be done exactly?
I would rather take the stance of feed algorithms and moderation logs be PUBLICLY available. Transparency instead of censorship.
I would rather take the stance of feed algorithms and moderation logs be PUBLICLY available. Transparency instead of censorship.
A site you can specify certain requirements like no doom scrolling or requirements like PUBLISH YOUR ALGORITHMS.
Why go full California with another banner? People will ignore it and you will have done nothing of substance.
You can soak the tobacco in solution which contains additives, such as more nicotine. Which is exactly what cigarette companies have done in the past (and not just the tobacco, the filters, and the paper as well).
The parallel here is filling people's feeds with divisive political news and posts, even when they have tried to opt out.
A social media website does not need doom scrolling or private algorithms for the feed, you can change how it works instead of adding a useless banner.
1) What would you expect be implemented to reduce/eradicate doom scrolling?
2) What would making the algorithm public do for us? I'm not an ML engineer, but presumably their algorithm isn't just an algebraic equation where x is how toxic the post is and y is how inflammatory it is and y is the number of kids who will think harder about suicide because of the post.
Maybe I'm just super naive and that _is_ how Facebook made their algorithm, but my understanding is that the algorithm is a little more of a black-box and is a little abstract. How is a lay-person supposed to evaluate something like that?
I do want them to publish their algorithms and moderation logs so we have insight on how they are serving and moderating content.
I don't care about organic user content, I do care if FB is pulling the strings to make it either more salacious or being biased in one way or another.
I also care if they are banning certain users or content but not others.
Some of the signals might be generated by ML also, like photo labels, but ultimately these things are very understandable if you have the model and data.
A social media behemoth does not need to use doom scrolling or private algorithms to be a social media site.
Even when it comes to encouraging positive play between users - in the new Pokemon MOBA (games known for their toxicity) there's no text chat, only communication with a few emotes you can show. Some of their decisions make for arguably worse games for "hardcore" gamers (like the way they rank users in smash, or how they focus on more casual-style in-game tournaments or make matchmaking harder) but they sacrifice that in favor of a more positive general experience, especially important since children play their games.
This part of the thread went pretty off topic but I like it! Pictochat was certainly ahead of its time, wish we stuck to things like that.
Yes, cig's cause cancer, but to say that it's because it burns tobacco is missing a big part of the story.
Were they even looking for lung cancers when people died back in 1800's?
That means mandates for data interoperability. Sadly, I have no idea how to implement that in a way that doesn't utterly stifle innovation by ossifying what sorts of data models social media is allowed to have. But at the very least we could create a sort of interoperability minimum that prevents you from locking up things like photo albums or peoples' "social graphs."
Over the longer term I'd like to see some kind of disentanglement of the protocols, standards, and data models from the front-end clients. It's obviously a lot more complicated now, but in the same way that you could access AIM, ICQ, GChat, and a bunch of other stuff from a variety of chat clients it would be good to be able to do this with everything social. Hell, ActivityPub basically tries to do this now so it's not impossible.
Good advertising[1] is limited to making sure your product is visible in comparison to competitors - having shiny cereal boxes is something I find pretty meh, but in the cereal aisle you're dealing with someone who wants to buy some kind of cereal and you're trying to convince them to buy yours. TV Advertising drives up demand for products which, by definition, means we're consuming more of that product than we otherwise would... that's great for business... and it's also great for the obesity epidemic.
1. What I'd consider to be ethical advertising, but that's like my opinion man.
I mean, ideally we could allocate all the capital we put into manufacturing, selling, and consuming beer into fitness or math education or something more harmonious with wellness and human achievement; but hey, plenty of great scientists and inventors love beer.
How about we don't let the state decide what I get to do with my body (and mind)?
Neither you or the government/state decides what you get to do with your mind. An advertising company decides what to do with it and can manipulate it however it decides best benefits itself. Not you, not society, Facebook, what makes Facebook the most money.
But a sin tax ostensibly accounts for the economic externality*. We know that cigarettes impose a cost on society beyond the individual smoker. I'm all in favor of making people pay for things that we know cause damage to society more broadly. And I hardly think it's controversial that social media is in many aspects harmful to society.
*Sin taxes are technically different than pigovian taxes, but I and I think most people tend to use the terms interchangeably.
What's that cost?
From what I've read, all economic costs smokers impose on society are more than made up for in their dying early, they actually cost less [1]. I guess everyone should smoke to save the state money!
[1] https://pantagraph.com/news/fact-check-do-smokers-cost-socie...
Generally speaking I'd rather reduce taxes but I fail to see what's wrong with e.g. an alcohol excise tax going towards rehabilitation and/or highway safety programs. "Sin tax" is just a colloquial name for an excise tax, which a state has every right to enact.
And if I choose to smoke in the privacy of my own home (or yard)? What societal harm am I causing?
As for alcohol, the societal harm caused is a laundry list of already illegal behaviors that are illegal regardless of alcohol's involvement with the exception of sin tax avoidance.
Why not outlaw the societal harm instead?
> e.g. an alcohol excise tax going towards rehabilitation and/or highway safety programs
Both of those seem like good things regardless don't they? Why do we need a special tax on alcohol for things that are generally good? It's not like only people who consume alcohol are the only ones who need rehab or they're the only problem with highway safety.
Does the tobacco tax go toward lung cancer patients? It actually goes towards funding campaigns that overstate (ie, lie) about the dangers of smoking to the point that people vastly overestimate the dangers of smoking [1].
> Sin tax" is just a colloquial name for an excise tax, which a state has every right to enact.
Of course it's legal, it's just garbage policy. Sin taxes come from the pairing politicians wanting more money and pearl clutching interest groups pleading to think about the children.
[1] https://journals.plos.org/plosone/article?id=10.1371/journal... > 99.5% of respondents overestimated absolute risk, only about 0.3% estimated it correctly (by giving an answer of 30), and 0.2% underestimated it (by giving an answer less than 30).
Weren't you just saying how you don't want the state legislating what you put in your body and mind? That's why.
Another example is driving a car, which results in thousands of fatalities and many more injuries daily. Not to mention environmental impacts which affect others. The state chooses to require drivers to have insurance and their cars to pass smog tests, rather than outlawing driving.
Smoking and alcohol use can also not harm others. Should those who smoke and drink responsibly be held responsible for those who don't? How does the tax ameliorate those harms?
How about you take full responsibility: you get to not put whatever in your body, and you agree never to take an ambulance ride or be treated by a hospital.
Don't want the vaccine? That's fine, it's your right. But now when you can't breathe, nobody coming to help.
Absolutely not. I know lots of people that manage to drink alcohol responsibly. They never drink and drive, don't regularly over indulge and it makes their and their peers lives _better_.
What negative externality are they paying for with alcohol taxes?
> How about you take full responsibility: you get to not put whatever in your body, and you agree never to take an ambulance ride or be treated by a hospital.
> Don't want the vaccine? That's fine, it's your right. But now when you can't breathe, nobody coming to help.
If I pay for health insurance, I'm already taxing myself in this instance. It would be perfectly reasonable for a health insurance company to offer incentives for people to be vaccinated just like they offer incentives to non-smokers.
If the government wants to start providing that healthcare, then they can have a say in the cost of poor health decisions.
The state acts on the collective when the public is not making good decisions for themselves and causing net harm onto themselves, usually with the public paying the price. Sometimes thats overt like with death rates from accidents without seatbelts, or cancer from asbestos exposure. Sometimes its less overt like the behavioral issues, increased incidents of mental illness, and crime rate increases from leaded pipes and gasoline.
I'm willing to bet social media causes net harm. It hasn't enabled communication that wasn't possible before; if you can get access to a facebook account you therefore have email and access to irc. But it has cost probably trillions in productivity from people staring at it so much during all their idle time, and the cost to treat mental health issues that wouldn't have cropped up without toxic social media culture.
I say we have these companies pay for these externalities if they are forcing us to pay for them otherwise. By not passing a tax on externalities like this, the state is deciding that I need to pay for facebook's ills on society whether I use the service or not, which should anger you as a libertarian as much as it angers me as someone on the left.
- The US prohibits people under 21 years old from buying alcohol, and allows those over 21 to do so.
- The US prohibits anyone of any age from driving a motor vehicle over a certain blood alcohol level.
This is something which causes health and community harm (alcohol), which we have allowed and denied to people in certain ways.
And honestly, I think struck a fair balance between individual liberty and social liberty/good.
I don't think anyone would argue that everyone should be allowed to drive anywhere, as drunk as they wanted to, at whatever age they wanted to.
I wear my seatbelt, I don't smoke, I don't drink and I'm vaccinated.
Everyone keeps talking about these "negative externalities" without being specific. Why not just make the societal harm illegal and let people hurt themselves without buying permission from the government?
Because these are all statistical precursors to accidentally killing someone with a car.
So, advertising company hurts me and the state gets compensated?
(PDF): https://www.everycrsreport.com/files/19980430_97-1053E_53c59...
It most certainly does. The retort your parent made is for those who make the argument that a tax is necessary because X (smoking, in this case) costs the country economically.
If you want to make the purely _economic_ argument, it's a benefit to the bottom line.
Very interesting idea, actually. There is evidence Social Media causes harm to some individuals' mental health (in a widespread manner causing some measurable societal harm), so a proposed tax on all social media companies with revenue going towards mental health programs seems worth exploring.
Generally I'm not much in favor of implementing new taxes (would rather close existing loopholes) but if implemented reasonably and backed by scientific evidence this seems valid.
It seems to be the near monopoly that is one of the major issues for FB. Lack of competition seems to lead to a house of horrors.
Meaning that blocking social media for kids and teens is likely on the anvil at some point.
And they do so by exploiting human weaknesses using the same psychological techniques used by casinos and other forms of gambling.
FB in some sense, but not entirely, is a form of speech, no better or worse than Grand Theft Auto or the National Enquirer. That's how I thought of it ten years ago.
Now that it is in our pockets nearly cradle to grave; a monopoly; and dependent on minutes of engagement rather than subscriptions -- it is a different animal altogether.
You tax Facebook but allow it to operate however it wants. Facebook is then incentivized to double down on its algorithms---like tobacco companies using chemical and biological techniques to make cigarettes more addictive---in order to regain the lost profits.
What is the specific harm involved here that is deserving to be taxed?
How would we measure this harm in order to know how much to tax a given company?
Should other causes of this harm be taxed/penalized as well? If not, why?
For instance, if the harm in question is some people feel varying degrees of worse after using a given product, is there any limit we as a society should set on penalizing the cause of the harm?
Should people or entities who say things that make people feel worse be fined/prosecuted by the law? If I feel worse (let's call this 'trauma' or 'anxiety' or 'depression' or 'literally shaking' or 'panic attack') after reading a book or reading a news site, should I have standing to sue the creators and medium which presents said content?
Taxing all social media or all media may have interesting implications as this again will reduce profit for all and give some revenue for governments without making any actual change. Also people making cooking/educational videos on Youtube may resent having to pay sin tax.
It doesn't notify you when someone responds to one of your posts. It doesn't send you any nagging emails (indeed, doesn't even require an email to sign up). It has the noprocrast setting to let you set limits on your own usage. It doesn't (afaik) try to optimize for engagement - dang tries to maintain civil discourse as much as possible.
These companies do awful things because:
1. They have few viable alternatives so ethical consumers end up choosing them when they might not otherwise.
2. They have enough money that it is profitable to do bad things and pay for the damage control.
3. They have enough power to defeat regulation and government oversight.
4. They are so large and monolithic that they can hide their internal workings more easily.
5. The org chart is so deep that those in power are psychologically removed from much of the consequences of their overpowered actions.
Public algorithms, or at least some 3rd party review. Ban infinite-scroll on social media platforms. Require feeds to be configurable (users can set to "newest first" or "top picks" or whatever else). I'm sure I could come up with more, that's just off the top of my head.
A click == friction.
I think that removing CDA Sec 230 protections for algorithmically-curated feeds is the answer. It's one thing to have a basic FIFO feed, it's another to hide or reveal content based on your own internal engagement/revenue targets. At the point where you're crafting bespoke engagement-maximizing feeds, you're creating a gestalt creative work that has a life of its own and that is no longer merely a passthrough for the works of users of the platform.
In other words, consider Facebook's curated feeds as Facebook's speech, not only the speech of their users, so Facebook faces liability for that speech.
All a ranking algorithm is doing is making suggestions that certain content be read before other content. That's protected speech.
I'm not talking about the legality, I'm talking about the liability. If one user posts defamatory/libelous speech and Facebook decides to blast it to 1 billion users because their secret sauce algorithm found that it really cranks the revenue dial, Facebook should share some culpability in that defamation lawsuit. A basic 'hot' algorithm that's not designed by squadrons of PhDs running machine learning on user behavior analysis isn't really the same thing, culpability-wise, imho.
A "hot" algorithm can blast high engagement libel in front of people as well as a fancy algorithm.
If the "hot" algorithm and the fancier algorithm are both content neutral, on what basis can you distinguish the two as a matter of law?
Does the hot algorithm become illegal if a PhD implements it? I'm at a loss about what distinction you're actually trying to draw.
Your post, like many others on this thread, isn't articulating exactly what about FB's conduct you find objectionable?
> Your post, like many others on this thread, isn't articulating exactly what about FB's conduct you find objectionable?
The objectionable conduct is pretty thoroughly covered in the article, but to review:
> "Facebook over and over again has shown it chooses profit over safety"
Fine. What, precisely, is the line between algorithms acceptable to you and algorithms not?
What is the precise conduct that would make liability attach to a ranking algorithm? You can emote, but you can't describe what exactly it is you would turn into a law.
I thought I made it pretty clear from the outset I was talking about removing CDA Sec 230 protections for sites using bespoke (i.e. proprietary) curation algorithms for their feeds.
HN is fully moderated, FB is barely moderated.
HN nerfs their algos. FB weaponizes theirs.
HN applies their rules fairly, evenly. FB does not.
FB has ads. HN does not.
FB weaponizes PII. HN does not.
Etc.
Comparing HN to FB is like bicycle vs freight train, firecracker vs nukes, nice bath vs a hurricane.
Your list of criteria is legally inactionable. Neither you nor anyone else can write down what precisely it is that FB should be prohibited from doing. All I see is a bunch of unproductive rhetoric about how this is bad or that is bad. Blah blah, filter bubbles, moderation, whatever.
What is the exact definition of the criteria you would use to prohibit ranking algorithms?
No targeted advertising.
No algorithmic news feed.
Voila. Fixed.
RSS feed readers can do it. So can Facebook.
> dang applies special rules to your comments if he doesn't like you.
What's FB's equivalent to HN's guidelines?
When has @dang violated his own guidelines?
Are you suggesting that automated curation, moderating is without bias?
> FB's algorithm is impersonal.
FB's algorithms are hyper personal. Machine mediated curation is sociopathic, utterly lacking empathy, not impersonal.
Repeating myself: every FB user gets their own news feed, every HN user sees the same thing.
Strong disagree. Why the hell shouldn't I be able show ads to people who want to see them?
> No algorithmic news feed.
HN is an algorithmic feed.
Nobody's stopping you. If you want to show ads to people interested in football, post ads on a football site. If you want to show ads to people interested in horses, show them on a horses site. If this type of non-invasive targeting to display ads doesn't make you as much profit, well that's a you problem.
> HN is an algorithmic feed.
You seem to be struggling with the idea of a 'bespoke' algorithm where everyone's algorithm is different, informed by an internet-wide surveillance system. This is the difference between HN and FB.
Still legitimate, and maybe even encouraged, are contextual ads, banners, sponsorships, branding, swag, paid placement, influencers, misc calls to action, etc.
Please rest your concerns. Late stage capitalism will continue unimpeded after Facebook's wings are clipped. Social media will revert to less noxious enterprises, like metafilter and craigslist.
> HN is an algorithmic feed.
Opt-in. Transparent.
What is FB's equivalent to HN's /newest?
I'm not crazy about HN's ranking. It's pretty good, non-toxic, seems fair. But someone other than me is making judgement calls.
I much prefer client side (user controlled) filtering and sorting, like with any generic RSS reader.
Repeating myself: All HN visitors see the same feed. A crucial qualitative point you pointedly do not acknowledge.
Nor do you answer any of my questions.
> Late stage capitalism
I don't subscribe to the neo-Marxist worldview that uses this "late stage capitalism" frame all the time. Capitalism is the natural order of the universe, not some temporary aberration on the way to your Utopia.
> Opt-in. Transparent.
Not transparent, and hardly opt-in: HN feed ranking is the default.
> I much prefer client side (user controlled) filtering and sorting, like with any generic RSS reader.
Okay, so people seeing different feeds is good...
> Repeating myself: All HN visitors see the same feed. A crucial qualitative point you pointedly do not acknowledge.
... and now people seeing different feeds is bad.
Make up your mind.
Different people should see different feeds because they're interested in different things. There's nothing sinister or nefarious about showing people stories about topics that interest them and not showing them stories about topics that don't.
I think even you'd agree that someone should be able to opt into seeing stories about "cooking" and opt out of stories about "motorcycles". Okay, that's good, right? Now what's wrong with using ML to infer, based on what someone reads, that he's interested in "cooking" and not "motorcycles"?
This interest inference is what FB is doing. There's nothing bad about it.
I maintain that the criticism you and others are lobbying against algorithmic feeds is logically incoherent, emotionally rooted, and unworkable as public policy.
And circling back to the original article, there is in fact something bad about it. One guy started out with a site to rate the attractiveness of women at Harvard, and now my mother is probably going to die because his customized algorithm found that showing her lots of anti-vaccine misinformation would maximize his profits.
As the scholarship advances, and both the predictive and explanatory powers they afford us, I'll update my views as best able.
Meanwhile, if you can't entertain the premise that Facebook's profit model is a major net negative for humanity, we don't have much to discuss.
Now, what controls are on the internal algorithmic dial? Apparently that's top secret, but a legal requirement to expose the interface to the users seems reasonable.
Note that this might not affect what ads you get served (that seems more on the private business side, although banning prescription pharma ads makes sense), but it would affect what shows up in your feed, what content you get served, etc. You could write your own exclude lists, for example (i.e. if you never want to see content from MSNBC, FOX, or CNN, that would be your decision - not the algorithms, etc.)
If they don't operate in America, then I don't give a fuck what they do.
The social media ecosystems aren't like that. You can't be a chat provider and plug into FB Messenger; you can't plug into Twitter, etc.
There is an open social media eco-system called the fediverse (for its federated nature), in which Mastodon is the best-known player. But it's gotten very limited traction, because of the network effect that keeps people on FB and Twitter. No such effect keeps people on Gmail.
Honestly, social media issues are for the most part a parenting issue. If you don't have access to your kids phone, or know what platforms they are on and who they are talking to and what they're sharing, I'm not sure legislating social media is going to do much of anything. New platforms will pop up, more private networks will be started and suddenly, everything becomes to fragmented to really oversee.
I would create laws that have teeth and address issues like bullying, doxxing, SWATING and other ways people weaponize social media against other people. You start to put some teeth into laws where people are facing serious consequences for bullying and pushing people to suicide, then you might see some changes.
Counterpoint: kids aren't all neurologically and socially developed enough to understand life-altering consequences for certain actions, and that's not their fault. Legal codes and law enforcement are too crude in most child-related cases, unless you're okay with incarcerating misbehaving children.
It's on adults to make sure things kids can reach are reasonably safe for—as well as from—them.
School administrators don't care about bullying and teens being driven to suicide, they care more about liability than fairness. Taking the King Solomon approach in public schools is abhorrent and injust. Kids do dumb things, but it makes it so much worse when incompetence and callous disregard create perverse incentives.
News feeds must be sorted in chronological order only. Users may selectively filter their feeds if they choose.
No infinite scrolling.
The problem is companies selling peoples data and optimizing the algorithms on probability. Instead what if everyone paid some taxes to have a social network which helps people interact and businesses promote themselves, you can get rid of the ads AND the algorithms. Let users customize settings which dictate the algorithm.
Ok let’s say that’s now the case. The FB source is now open. What changed? Any negative consequences are still occurring, and if anything, we just have had actors better visibility into what to exploit.
It looks like the recco algorithms will be carved out from the neutral platform protections they enjoy.
What if social media platforms required all minors to have their account associated with a parent account? The parent could monitor activity, institute time limits, etc.
Platforms that harbor minors and adults together will have to have different rules than platforms just for adults. But you also cannot make everything fit for kids. Government would actually try to do just that since ensuring safety here is their mandate. So a solution must be found. Normally minors should be supervised, but that is not trivial and you don't want constant surveillance.
Do something to make platforms responsible when their "algorithms" promote something. Not just hosting it, but when they promote it. Don't like it? Don't curate, then, or have a human do it so you're sure nothing you're deliberately promoting is shitty enough to land you on the wrong end of a lawsuit. "But how will tech companies show every visitor a totally different home page of content they're promoting (but in no way responsible for), and how will Youtube find a way to recommend Jordan Peterson and Joe Rogan videos next to every damn thing? How will tech companies make every part of their 'experience' algorithmically-selected, personalized recommendations of content they farmed from randos?" They won't, they won't, and... they won't. You're welcome.
Make data leaks so cripplingly expensive that no company would dare hoard personal data it didn't absolutely need to get by.
Force the quasi-official credit reporting agencies not to be so shitty. In particular, "freezes" should be free and should be the default, alerts for activity should be free, and access to one's own info should be on demand at any time, not once per year per agency. Or just outlaw the bastards completely, IDGAF.
I dunno, lots of things we could do to make the current personal data free-for-all less hellish.
It sounds like you're suggesting GDPR style regulation. They're still figuring out how to enforce that but generally I support it. Too much money is against it to get anything passed in the US, though.
Another problem is that the US government seems to like when the tech sector gobbles up data on people. It gives them new powers for social control.
That being said. No. This is no more a free speech issue than forcing food manufacturers to make their ingredients public.