Behind TikTok's boom: A legion of traumatized, $10-a-day content moderators
thebureauinvestigates.com
thebureauinvestigates.com
I visited a police station in person to report a $3000 fraud with clear evidence and also document forgery. Police told me to get a lawyer, gave me the fraud report but told me I'm wasting my time if I fill it out.
I've also had a stolen laptop, tracked with Find my Mac, and again, police handed me the report form and told me to contact my insurance.
The short answer... It depends.
People always assume a lot of things when they imply that video evidence makes something an easy case to solve. Identifying people in videos isn't as easy as it seems, for example. It's also not safe to assume that the person posting a video is also the person who filmed it, nor that the location data is correct. So who's the suspect?
Also, where does TikTok send the video? Do they send it to Police in a city where a video is geo tagged, or where it was uploaded, or where the user account was created?
This type of stuff seems trivial when reading an article like this, but it isn't in real life. It's all stuff that may take weeks to figure out and all that time can't be spent on other cases. So yes, sometimes (maybe even often) such reports don't lead to cases. And even if they did, there's no guarantee it would lead to anything.
The cop said something like "What's your deductible? Something like $250? Well how much work do you think we should do for $250?" (this was in the 80's).
I hate that this type of statement always pops up as soon as I see any mention of LE. While it might be true for some parts of the US (and probably other countries), it isn't true for the majority of people involved in this field. It's just incredibly hard (thankfully) to convict someone of a crime. It requires a lot of evidence and paperwork. Meanwhile there's a pretty limited amount of resources to work on this kind of thing.
It would be impossible for 'authorities' to investigate 20 million images they receive from Facebook. But implying that authorities don't investigate anything, and don't even want to, is just flat out wrong. For example, here's a story about an awful child predator that was caught in The Netherlands, because American authorities started an investigation into reported images: https://en.wikipedia.org/wiki/Amsterdam_sex_crimes_case
This case is also an example of what authorities tend to focus on, which is producers. A lot of effort is spent trying to prevent producers/abusers from causing more harm. All of this effort can't be invested in investigating brokers/spreaders of 'known' CSAM. Time is limited for everyone and it is too for authorities.
And yes, a probable majority of fraud cases don't get picked up, even if they seem obvious and easy to solve. This is because the man hours for actually getting someone convicted is huge and authorities usually focus on cases involving violence. I personally had to cut short fraud investigations, because preventing a wife from getting murdered by her husband or stopping an illegal arms shipment took preference. And I can't say I feel guilty about that.
And you shouldn't. But it's still a problem that lots of fraud, scam etc just gets ignored.
There's so much fraud out there that looks easy to stop, because it's clear who does it, it's clear that they do it (online, with plenty of evidence) etc etc. Yet when you bring it to the attention of the authorities, they'll look at the individual case and say "okay, but it's only 50€, there's not enough public interest", which I could understand, if it wasn't a commercial-grade criminal operation that does the same thing thousands or tens of thousands of times.
I get that paperwork sucks and investigations take a lot of time and often get nowhere because something just doesn't materialize, but I get the feeling that we're more and more policing non-violent crime like Google & Co do tech support: if there's not a very public issue that the media is actively reporting on, it gets ignored.
And if it's a resource issue: get more resources, it'll pay for itself many times over. A recent example in Germany where some Clan from Turkey had been defrauding the elderly for years and ran up dozens to hundreds of millions of Euros in damages could've been stopped very quickly if there had been more interest from law enforcement. I'm sure we could've paid a person or two with that kind of money.
Apple could report so many wrongdoing to authorities but instead they quietly support all those criminals.
/s
If they report me to the uk government, nothing will happen. If I never go to china, nothing will happen, so what is the basis of your thoughts on this?
Considering how many reports there are of Facebook, Youtube, Tiktok, and other sites where content moderators are suffering from having to perform their jobs, it suggests that there's a great deal of really terrible stuff going on - and worse, that the people involved are filming and attempting to share it.
This is a serious thing, and it paints humanity as being far darker than it would seem from the surface. It also suggests that the apocalyptic movies may not be so far off when they suggest that humanity will revert to open barbarism if we're faced with a catastrophy large enough.
I already knew that. I even know the percentage roughly. According to the Pareto principle, 80% of people are not totally fucked up, but the 20% that are cause 80% of the problems.
No, you don't. This "principle" is not a law of physics, it's just a rule of thumb that may or may not be wildly off the mark for an arbitrary distribution. Its common usage is like statistics horoscope.
I still haven’t found one company where this principle didn’t show up.
Very scientific.
This is where the internet fails.
We have a long way to go until everyone on the planet can grow up in conditions where you turn out 'default good/well-meaning towards others'.
Is it ethical to create a sentient AI that might suffer? -> we already make similar decisions millions of times per day
Don't treat humans as means to an end, instead treat them as ends -> we create other humans to satisfy our current needs in a world characterized by struggle
And of course as you have pointed out, some groups consider that it's better by default for a child to exist in adverse starting conditions (whilst also negatively impacting other people) than to not exist
We know the great mass of humans will carry on as before and there's little we can do to change that, but we are also aware that we have this responsibility that we can act on in our own life. Every person in the position to reproduce has a choice laid out before them. For the majority of us, it's the single most impactful ethical decision we will ever make.
Population explosion also began with agriculture, since more offspring meant more people to work your lands. Hunter-gatherers breastfed longer (so more years between ability to become pregnant again). Plus they were relatively mobile, so more having more children was not a practical goal.
Many of the diseases and other problems we have now are related to population increases in small areas (bad sanitation, etc.).
> The Younger Dryas (c. 12,900 to 11,700 years BP[2]) was a return to glacial conditions which temporarily reversed the gradual climatic warming after the Last Glacial Maximum (LGM, c. 27,000 to 20,000 years BP). The Younger Dryas was the last stage of the Pleistocene epoch (c. 2,580,000 to 11,700 years BP) and it preceded the current, warmer Holocene epoch. The Younger Dryas was the most severe and long lasting of several interruptions to the warming of the Earth's climate...
https://en.wikipedia.org/wiki/Younger_Dryas
Everything was good for a couple of million years, and then everything got really really bad for over a thousand years.
What we think of as human history is the story of our recovery from the trauma of the YD?
Think of it as the most extreme form of poverty. No clean water, no shoes, malnutrition, limited protection against the weather. Every day going about basic tasks carries the risk for death, disease and disablement.
There is a reason that human population didn't rapidly increase until relatively recently in our collective hundred-thousand-year history. It's because most humans died before reaching reproductive age. Even in the few pockets of the globe with conditions favorable to survival, periodic events (weather, disease, rival groups, over-hunting, etc) wiped out entire tribes in terrible ways.
I've read arguments in both directions and I'm extremely suspicious that they are almost all basically politically-driven. It goes like this:
If hunter-gatherers are predisposed to more or less gender or economic equality, or certain social structures, then that should perhaps inform how we construct our own modern societies.
In order to escape the clear appeal to nature fallacy, it then becomes necessary to argue that not only were prehistoric societies constructed in a certain way, but they were also extremely well-off. Therefore we clearly must "reject modernity, retvrn to monke", and embrace True Human Nature embodied by some cultural tradition or ideology.
The exact inverse would obviously imply that we must embrace a certain idea of technological or social progress in order to escape the "natural state of humanity" as fast as we can.
However, was the human hunter-gatherer experience was ever all that stable or predictable such that we can obviously draw out either of those major conclusions? I think a decent null hypothesis might be that all the extremes of human experience and social structure had to have occurred to some degree and that the overall average and distribution would fluctuate quite a bit over time according to weather, migration patterns, and accidents of cultural evolution that humanity at any level had basically no control over.
I find it really hard to believe that we can possibly have enough evidence in any relevant discipline to rule this out.
>the same is consequent to the time, wherein men live without other security, than what their own strength, and their own invention shall furnish them withal. In such condition, there is no place for industry; because the fruit thereof is uncertain: and consequently no culture of the earth; no navigation, nor use of the commodities that may be imported by sea; no commodious building; no instruments of moving, and removing, such things as require much force; no knowledge of the face of the earth; no account of time; no arts; no letters; no society; and which is worst of all, continual fear, and danger of violent death; and the life of man, solitary, poor, nasty, brutish, and short.
Hobbes, like many philosophers, did not have much faith in human nature.
"The heart is more deceitful than all else And is desperately sick; Who can understand it?" - Jeremiah
"Then the Lord saw that the wickedness of man was great on the earth, and that every intent of the thoughts of his heart was only evil continually." - Moses, about man before the flood.
"This is the judgment, that the Light has come into the world, and men loved the darkness rather than the Light, for their deeds were evil." - the Gospel of John
I don't know how accurate that 1 billion number of users is exactly, but it's probably in the right ballpark. In reality there are probably a few million "serial posters" of horrible stuff. Either way, with numbers like that it's very easy to get millions of undisputedly horrible videos.
Ask Elon Musk to offer to buy it, then we'll see how accurate those numbers are.
In reality, I never believe people's published numbers. OF COURSE they are inflated. Their entire (maybe majority?) valuations are dependent on those numbers being as high as possible.
Bytedance is pre-IPO. However, you are accusing Google, Facebook, snapchat, Twitter of securities fraud. What is the latest from Musk, on Twitter's alleged securities fraud?
Do you really believe TikTok as exactly 1 Billion users, or 999,990,000+ users? Anyone that says that is 1 Billion is padding the numbers. Nobody is going to call that 999 million.
If the number is anything close to being rounded, the number is padded. Unless they are showing you an exact number from the time the count was read, it is being padded.
I find it both depressing and comforting; it's comforting to know most people aren't that bad, and it's depressing that such a small group of assholes can fuck it all up.
If I watch TikTok for an hour a day for a week, and I watch 4 videos/minute on average, I’ll watch over 1500 videos a week. If only one or two of those are VERY traumatizing (e.g. showing death), I’d think TikTok has a big moderation problem. That’s an ultra-low number of users making it seem like there is a huge issue
Disclaimer: I don’t have TikTok
Do we need to be way more careful with how we design our society? Yes. Does it make sense to abandon possible solutions because of a movie? No.
I think we need to move to some way to identify actual people on the internet. I don't mean that in a "use your real name and upload your passport"-kind of way, but in a "we can 100% reliably tell you've been a twat in the past, whomever you are, so no new signups for you"-kind of way that also accounts for all the various interests such as right to privacy and anonymity.
Many platforms use phone numbers for that now, which is obviously far from perfect as I can easily get 100 of those today if I wanted to.
[1] https://www.insuranceproviders.com/best-and-worst-cities-for...
[2] https://www.thecut.com/2016/03/for-80-years-young-americans-...
Unless a person intentionally hides from the world, it is hard to not be aware of this. Human trafficking, cartels, pollution, there is a myriad of ways in which humans are just not "great people". Why act disillusioned by it?
It's the same kind of people who adopt[1] in order to get government subsidies but treat their charges worse than cattle --at least cattle get to graze nd do cattle things.
[1] I hope it's not necessary to state that this is a small minority of adoptive/foster parents, but they exist, unfortunately.
TikTok is much harder to censor in automated ways, clearly. People also find ways around the censors. So all that awful stuff may have found a way back onto the internet.
"Build it yourself" internet bubbles are giving people hilariously wrong impressions of society, and humanity.
These same bubbles led to many millions of people shocked and flabbergasted when Donald Trump won in 2016. There will probably be just as many flabbergasted people when Democrats get thrown out of the House in three weeks. Internet "safe spaces" and censorship of differing opinions give people that impression that society is changing, or has changed. It hasn't.
And if you're right, I hope this is all a simulation.
The fact that people are downvoting this obvious proclamation is further proof of my original claim. People are in bubbles, unaware of reality. No one should be shocked that Democrats are about to lose the Congress, and yet here we are.
https://projects.fivethirtyeight.com/2022-election-forecast/...
https://electionbettingodds.com/House2022.html
https://polymarket.com/market/which-party-will-win-the-us-ho...
yeah I think that's the theme of David Lynch movies
> It also suggests that the apocalyptic movies may not be so far off when they suggest that humanity will revert to open barbarism if we're faced with a catastrophy large enough.
I think that's disproven - I saw a study where they looked at peoples' behaviour during the WWII London bombings and it was mostly characterized by Mutual Aid
This is not to say TT doesn't contain such content, but it should be considered the feed you're seeing might be *significantly* different to someone else's feed…
What's up with this authoritarian discourse on HN? Since when are we casually calling for censorship of things that are "disliked"?
The megacorps are having videos taken down off so-called private clouds for calling out the rulling class too much. I worry for the Orwellian direction we are heading into.
Some would argue: since technological progress outpaced society's capacity to cope with it. One of the possible answers for the Fermi Paradox is "everyone invents Facebook shortly after computers", and it seems a lot more plausible than it did ten years ago.
It's hard not to see surveillance capitalism as a variant of what Orwell was warning about - Facebook etc. are the telescreens and two-minutes-hate from 1984. As megacorps start to rival governments in reach and power, who exactly controls them is somewhat less important than their overall impact on the civilization.
No. The platform is just a place just like a sidewalk is a place. The problem is society itself. But I guess that's a hard pill to swallow for people and making up imaginary enemies then laying down the regulations on those imaginary enemies is preferred to addressing issues with society itself.
There is a ton of evidence YouTube, Facebook, Twitter, etc. deliberately push inflammatory / extremist material to people to increase engagement.
My 8 year old daughter exclusively watches Let's Play and "how to draw" / DIY crafting videos on YouTube, and still gets recommended content on why vaccinations are dangerous and junk health pseudoscience, flat earth and New World Order conspiracy theories, and a ton of comparatively vanilla highly biased political "commentary."
Their algorithm tells them if they can get her interest on these videos (instead of others) she will spend more time on their site. They do not care why she is spending more time on their site, or if there is any societal harm from the content she is consuming - it's a "free country, after all."
If every time you went to your coffee shop, they introduced you to another customer peddling snake oil or handed you a newspaper full of extreme left or right wing talking points, you wouldn't say "it's just a coffee shop."
Don't forget these "public squares" are companies who's sole purpose is to show you ads to generate revenues, they're not here to provide a fair and calm public forum, they're not public utilities
It obviously is a two way street, technology shapes society as much as we shape technology. Do you think humanity evolves independently of its tools ?
People are horrible, what we made with tech equally is, media were biased before and we just have more sources and channels now. If anything you're making my point
What made Von Braun's rocket designs more evil than those by others? The transitive morality is a deeply dubious assumption.
However, this freedom is only aimed at the users themselves. In reality, venture capitalists, highly-paid software workers, digital nomads and other gentrifying forces are by and large on the side of authoritarianism in the role they play in the world. They produce systems that reward them greatly while increasing surveillance and quantification of the rest of the population.
It's not uncommon to see employees of adtech FAANG companies here wonder, without any trace of irony or self-awareness, why more people don't use ad blockers on their router or browser. There's a reason why the "make the world a better place" phrase is so heavily parodied.
Which makes it more noticeable to certain people.
Did you go into tiktok 'rubbernecking at the car crash'? If you watch bad content, it will serve more bad content to you.
I've had this reveal things about myself that I wouldnt have articulated, and aren't my idealized self perception: ex. videos of text-to-speech reddit posts overlayed over someone hopping around a minecraft obstacle course are compelling to my brain, even if my conscious mind finds them kind of dumb and silly.
It's a very manipulative platform on all sides. Creator, Consumer, and Moderator.
>Social media also normalizes behavior (and weird beliefs) to the extent that they become destructive to the rest of society.
Yes, there have always been loonies. Yes. But the issue is that social media has brought them together and given them a voice they have never had before.
When in the history of the US has there been a massively mainstream movement predicated on the fact that Democrats are literally agents of the devil, that wealthy elites harvest chemicals from tortured children, and that a shady New York businessman might be God's representative on earth?
> When in the history of the US has there been a massively mainstream movement predicated on the fact that Democrats are literally agents of the devil, that wealthy elites harvest chemicals from tortured children, and that a shady New York businessman might be God's representative on earth?
Never. Mostly because Trump hasn't been around a hundred years prior, and harvesting chemicals from tortured children wasn't a thing anybody cared about. Or did you mean crazy antisemitic stuff and literally burning people because they figured they're witches? I don't know, sounds crazy, I'm sure that never happened.
And they’ve worked flawlessly.
It is a shared problem. Colombia is not so far away, neither geographically nor as a gig-heavy economic model.
We should be able to solve this – not from a “content moderation” perspective but from a human work conditions perspective. Counterintuitively, solving content moderation probably requires solving the human work conditions. If the workers were paid considerably better and well-supported, the costs might rise to the point of successfully incentivizing the ideal automated solutions.
I would love to know from people working in this space how they handle illegal videos. Are they responsible to report it to the authorities in the respective countries? The company will not have a presence in many of the countries the app is used in, so even if they wanted to report to the authorities it's impossible to keep up. Do they just delete and move on? The legal, ethical implications here are staggering.
They were also on one month contracts. If you made it through your initial period you could choose to convert to full time. Most didn’t.
I shudder to think what the next generation of moderators is facing.
They have talented AI developers. They have a massive amount of harmful known content from moderation. They still need to surface harmful content to humans to check.
Ultimately maybe it just doesn't make sense for social networks to be centralized monoliths.
Besides, demand and supply.
Every time you hear "social media should be unmoderated and it should be impossible to deplatform people", you should think of all the traumatizing-but-not-actually-illegal material here which people want shown to everybody.
It would almost be better to say nothing.
Larger things scale less control over everything the chimp troupe with their 6 inch chimp brains have. Its the lesson of the last 2O years - we quickly loose control and enter Jurassic Park style reality.
In my experience, it takes people in positions of power, and with real skills for cultivating culture, to deeply change an existing work culture for the better. I've seen a few leaders with such skills make wonderful changes at their scale of influence. But I've seen far more people in leadership positions who act as if they are largely unaware of the nuance and importance of good company culture.
Such changes are also hard because meaningful cultural improvements often conflict with short term revenue/profit. It takes a lot of discipline in senior leadership to maintain the needed resolve given ever-present pressures to produce in the short term.
I know it's not ideal to have to go through the pain and trouble of being the good and righteous person who stands up against bad employers; but you HAVE to do it, or this is what happens.
One perspective is that large institutions have a sort of life of their own, caused by the interaction of internal factions and priorities, and almost beyond the ability of any individual to change.
Nobody in America is in favour of school shootings, and yet they get loads of them. Nobody at Google wants people to avoid their products expecting them to be cancelled, and yet they have that reputation.
Another perspective is that of course the institution's actions are the result of individuals' actions, what else could it possibly be?
That multinational that appears to have committed manslaughter would like us to believe nobody was responsible and it was just a tragic series of misunderstandings and communication breakdowns, so nobody should face any punishment. Are we fools to believe that?
Of course, but the naive part is believing all those individuals are spherical people in a vacuum, making independent decisions that are all well-thought-out and optimize for globally best outcome.
> That multinational that appears to have committed manslaughter would like us to believe nobody was responsible and it was just a tragic series of misunderstandings and communication breakdowns, so nobody should face any punishment. Are we fools to believe that?
Yes and no. It's foolish to believe individuals have much agency in this setting. Everyone, from the bottom tier to C-suite, is entangled in a web of interlocking incentives, that are ultimately anchored outside of any one company or institution[0]. Some people are handling large levers and could almost unilaterally make the company change course, but at a great cost to themselves[1], and thus it's foolish to expect them to become heroes, especially before an issue hits the news cycle. It's the same kind of error in thinking that rests behind ideas like "voting with your wallet".
That said, the conclusion that "nobody should face any punishment" is also wrong. Punishment is a powerful incentive that can cut through the tangled web. Even if you believe nobody is by themselves responsible for a bad thing, targeted punishment (or threat of it) at e.g. people holding the levers can encourage them to be more eager to pull those levers and steer the entire system away from causing the bad outcome.
----
[0] - Like, desire to keep your current standard of living, whatever it is, which may be driven not by your own need for comfort, but a desire to not disappoint your spouse and/or children.
[1] - And highly likely it would be reversed the moment they got fired for operating the lever wrong. Checks and balances :).
[1] https://biz30.timedoctor.com/average-salary-in-colombia/
Resist. Delete your social media accounts.
forcing people to do shitty jobs that make that unwell, for shit all cash isn't a great look. Arguing that its ok because they make slightly above the median income for the place they live, even less so.
https://restofworld.org/2021/the-chinese-content-farms-behin...
Same for the folks at click farms boosting those pieces of content.
https://www.theguardian.com/technology/2013/aug/02/click-far...
It's fascinating to me how many people really want to blame the cogs in the machine over the designers and owners of it.
There's a vast spectrum between "I'm Bill Gates and nothing is ever out of my reach" and "slavery". Folks in poor nations with few prospects being exploited for their cheap labor often gets uncomfortably close to the slavery end of things.
- Learned several cooking recipes and some tricks
- Cleaning advice I've actually put to use
- How to use some woodworking tools. Also, general woodworking ideas
- Cat behavioral advice
In my experience, it's far better than Reddit (mostly because there's far more niche content, and it isn't limited to "recently posted") and the discovery is far easier than on YouTube.
Can someone support that from Germany? I suspect that's highly locale-dependent.
I have given TikTok several tries, spent hours searching for interesting stuff and liking things.
I always get generic (and really bad) cooking videos and drunken people doing stupid things.
I'm based in Germany but view English content and I can at least echo your sentiments about the lack of good content. I've read multiple times about how amazing the TikTok algorithm is meant to be and how quickly it learns about what to show you, and yet that has not been my experience at all.
Twice now I've spent an hour attempting to train the algorithm but it's just endless streams of trash. The worst part for me was coming across an interesting video, for example some industrial machine process, only to be subjected to the exact same video hosted by another user. Most times they have just overlaid some music, but sometimes they crop the original or add some useless text overlays.
The part that also annoyed me was it felt like the algorithm gave up every 20-25 videos and tried to show me something sexual, even though I was not interested. I always found the subject transitions extremely jarring and annoying - I already have my porn channels sorted thank you very much, I'm looking for a different type of wood-working
I meant English language content, but with my phone being in Germany.
Thanks for telling your experience!
It's actually a fun quirk, you visit a country and you get to see their tiktok videos.
I have the following channels:
- twitch, good roi, short lived - affiliate forums, long lived, breakeven roi - youtube, fb, ig, cant ever recover investment, only good for content creation - ad networks and tik tok , totally 0 conversion rate even at what appears to be tremendous amount of traffic.
So yeah, if you arebin my niche a and a million views does not convert, then that algo and the platform are useless.
Why would you do that lol? Do you really have FOMO for the next garbage social media dopamine machine that simply wants to hijack your attention to sell ads and get you addicted to zombie scrolling like every other instance before it? Why does it have to work for you? I think you have your answer at this point.
It's just another garbage lowest common denominator mass content farm and memes. It doesn't have to be nor isn't something more profound than that, and in fact with it's short content length and more sophisticated AI it's just like everything before but worse.
You can certainly just dismiss it and occupy your limited time doing something else more productive or fulfilling.
Stop patronizing me.
Just checking that you are using a mobile (android or iOS) version of the app. I may be totally wrong (I’m basically making this up based on intuition), but I wouldn’t be surprised if the internet site version of TikTok doesn’t support the same level of input to the algorithm. There is fine tuning to your interaction data that the mobile app supports (I’m sure that it bases decisions on how long you watch videos (regardless of whether you like or follow), and I think it might even base decisions of what sections of videos make you stop watching)).
I hit it off right away with the mobile version. Off the bat: there were lots of half-nude women, gym rat videos, and other generic social content for my region. Within the hour, I was getting mostly finely curated comedy clips from standup comedians I liked (and not from others I didn’t like). It’s really amazing how well the algorithm works.
I think there is a language switch somewhere high up in their algorithm that is hard to compete against with watch metrics.
I find that very hard to believe (despite how god awful Reddit has gotten by now). Reddit is pretty much the last place on the internet where real people post long-form content that is not a single-issue niche forum. I can't imagine that the mode of TikTok with its short videos, algorithm-driven discovery and lack of community (commenting) encourages quality.
What they mean is a continuous churn of surface level stuff.
E.g. for woodworking, there's only so many times you can show how the basics of "10 Things you're doing wrong", or why the tape measure moves at the end.
You're not going to have that on the top of the subreddit that many times - maybe once a year.
On tiktok, that stuff will occasionally pop in, mixed with a bunch of other crap so the repetitiveness doesn't stick out in the short term. You also don't engage with it. There's almost no point to leaving a comment, it's all about mindlessly scrolling from video to video. This leads to lots of really easy to film but superficial stuff in large quantities.
It's playing out exactly like elsewhere on the Internet: beginner content is best to produce because they are the biggest audience, and the algorithm snowballs that advantage further because it only sees engagement, not quality.
I don't think these are hugely popular but they reach some audience. A big video will have 500k to Millions of likes.
These longer videos usually cap out at a couple thousand.
They're out there but it's not searchable.
TikTok has an incredible breadth of content but takes a few days to weeks to get the algorithm to show you what you want.
When I use it the algorithm is something I'm actively trying to manipulate. "Up vote" stuff you to see by liking and saving. "Down vote" by quickly scrolling away from a video.
Any engagement is a signal to show you more of that -- if you're tired of seeing one type of video, you should immediately scroll as soon as you know you don't want to watch it.
'Training the alghoritm' is just carrot on proverbial curated-content stick (you could also select 10 categories of content you like and refine it yourself further down the road, this is just less obvious for the same).
People were bashing Facebook/Instagram for this for a decade and many have stopped/rollbacked and just give few likes here and there for personal events of friends. I guess history even in digital world likes to repeat itself...
Yes, it is known that most people do not care about deep stuff, but about food in their belly and beeing good entertained. Panem et circenses.
Btw. about the romans, in italy the prime minister is now a person who said she thinks Mussolini was a good politician. So yes, history repeats itself, if too many people do not care.
Democracy is actually a quite fragile thing. Many things we take for granted only exist, because people care. And it will erode if people stop caring.
So you are still free to submit to FB and TikTok as much as you like and hope this does not change.
Because, you know, companies knowing all about you, but you nothing about them and their motives is just a receipe for long term empowerment of them and not you.
And in the case of TikTok, there is an actual government behind it and de facto in control, which is not famous for democracy, nor human rights.
Since you claim it’s not an AÍ in any meaningful sense, could you clarify what “training their datasets” means?
Some people put moniker AI on this, just like some people in marketing are putting AI into our TVs, toothbrushes, cars and god knows what else. Its got nothing with Artificial Intelligence per se, just a shortcut for 'something complicated I don't grok so its 23rd century magic'.
What Tiktok is definitely not doing, is "training their datasets".
I also joined FB to keep in touch with people I met in real life. It changed to push me to engage with groups and people I had nothing in common with.
I didn't join TikTok to keep up with friends that way. The TikTok algorithm does a great job of showing me new stuff that I might be interested in, but is still related. It's remarkably good.
FWIW -- It also seems to forget everything about me if I log off for a months.
---
Sure, I'm training data sets. Everything I do online is monitored and analyzed by someone at some point.
I spent a year of my life trying to "get off Google." IMO it's pretty close to impossible if you want to function in society.
I'm skeptical that the TikTok algorithm knows anything about me that my search history doesn't show.
I don't think TikTok is a "good thing" but I don't understand why it is sometimes painted as this unique evil.
The worst part of TikTok in my opinion is how much time I've wasted using it, but that's nothing new for social networks.
On the other hand, TikTok works fairly well. Videos can be up to 3 minutes long now, most of the time it's enough to at least explain the basics and tell you where to get more detail if you want. Also, the "algorithm-driven discovery" means that creators usually need to focus on the content of the videos. The "video reply to a comment" mechanic is very helpful too, as it encourages making additional first-level content from people's questions instead of burying things in comments.
there’s also a shocking amount of astroturfing going on. it used to be useful to search “[product] reddit” - and in some cases still can be - but many companies have cottoned onto this and they can extremely cheaply manipulate a reddit post for very high fidelity advertising. this is especially true for online-centric companies like VPNs and crypto exchanges. the VPN subreddit might as well be an advertising platform
I once saw a reddit post asking about VR headsets where the poster specifically stated that they do not want anything Meta/Facebook because they don’t trust them. the top 3 replies? “I recommend the metaquest 2 because it’s the best value for money” as if they’d said nothing at all
I think in that situation, the people who suggest a hammer or a screw silently perform some mental arithmetic, assume the person asking the question doesn't understand the problem because they're asking for a very costly tradeoff, and give the best answer that solves the basic problem.
There are some hilarious stories, and I haven't encountered such an individual yet. But I have heard one story where one individual has managed to evade a certain subreddit over 62 times. Enough for the moderators of that subreddit to send a public complaint that other moderators like me probably picked up on.
But they are working on improving their ban evasion tooling so who knows if that will do anything against that one dude that managed to evade over 62 times. Where there is a will..there is a way...and that's what frustrates me about Tiktok. I couldn't imagine moderating a video based platform mainly, just takes too much of your time. Imagine having to deal with ban evaders on top of that and related.
Ex. there is a moderator of major sub who likes to leave derisive comments to 'explain' moderation decisions and made a subreddit so other mods can join in on mocking users.
youtube is the biggest counter-example to this. yes there’s a lot of short-form content, but there’s also a huge amount of long-form, not that the algorithm will tell you that
I suspect as well that simply watching shorts over long form content makes the recommendation algorithms assume things about your ability to concentrate on long form content.
Even still youtube has a function where you click the dots near a video and say 'not interested' to get rid of channels and topics from recommendations that works fairly well, so if its recommending something you dislike, you have the ability to remove those recommendations and tune future ones.
PEBCAK
Well, or their model isn't very good.
To take an example, let's say I watch "Race Highlights | 2022 United States Grand Prix" on the official F1 channel
Then the algorithm recommends "The Internet's Best Reactions To The 2022 United States Grand Prix: THAT WAS EPIC" with a thumbnail of gurning man.
If the model had accurately captured my viewing habits, the model would say: Formula 1, yes; Reaction videos, no; Things that call themselves epic, no; Gurning thumbnail, no. And it would not have recommended that video.
One possibility is that Youtube knows me better than I know myself, and though I think I don't like reaction videos and gurning they've determined I secretly love them. Another possibility is that their algorithm merely knows that people who watch F1 videos watch F1 videos, and because I watched an F1 video they're showing me an F1 video.
I know which I think is more likely :)
not that this makes a good algorithm, but it’s a potential explanation
I'm not sure how to phrase this politely, but your comment comes off as deeply arrogant and unpleasant. That acronym especially just sounds like it is coming straight from an unflattering caricature of a nerd from the 90s/early 00s.
Long form text can be skimmed, organized, and referenced. The signal to noise ratio can be a little lower without being an issue, because it's trivial to sift through a certain amount of noise when you are reading text. And you can have conversational content that doesn't spiral out of time constraints.
TV is like this and was much worse. Back where I grew up, during forced socialism/communism there were 2 official TV channels and that was it. Then came computers and internet, me and many others and we greatly enjoyed that one felt in complete power what to do, what to see read etc.
What you mention (maybe I don't get it since I never used tiktok and probably never will due to apparent addictiveness) sounds again like passive random consuming of whatever the other side decides is appropriate. You can tune it, and maybe it works a bit, hey its 2022. Surreal to me (and not in positive way), but to each their own.
Fuck Tiktok
My take-away was there's good stuff on there and awful stuff on there. Similar to the internet itself. The HN comments on my post were very split between loving and despising TikTok: https://news.ycombinator.com/item?id=31360955
I hate black-box recommendation algorithms because of this. You have no idea what it's inferred from your behavior, so simply waiting too long is enough to bork your feed. How long is too long? No idea.
- Behind the scenes theatre stuff (I used to work in AV so this is relevant to my past interests) - Aviation interests - got tons of good videos around the Oshkosh fly-in - Farming videos - I grew up on a farm and enjoy many of these, I especially like comparing how this is done in different parts of the country - Ski-lift technician TikTok - videos explaining safety systems, backup power, and behind the scenes of ski-lift operation - Broadcast Transmitter engineering - folks working on terrestrial broadcasting systems which is loosely related to - Amateur radio - lots of folks showing off their shacks and mobile rigs - and more that I can't remember
So sad that this sort of job is necessary.
I’m curious though why YouTube never seems to feature? Most of the stories I’ve seen seen to be FB
Maybe necessary, but is it worth the problems that it's creating? I would argue that the armies of scarred people is worse than not moderating at all.
Except going forward they've be messed up, but with money.
Small publishing tools successfully use passive mechanisms to prevent illegal/horrific content (for example: community enforcement and platform culture, identity verification, cost/pricing). At what point do bad actors and unwanted content arrive? At what point do the passive mechanisms fail?
I didn't realize human labor was that cheap. I assume Teleperformance is taking a decent chunk of $1.8/hr. 15 seconds to view each video = 240 videos/hr.
On AWS, p4d.24xlarge is $32/hr with 8 A100 GPUs or $4/A100 GPU hr.
> “The human brain is the most effective tool to identify toxic material,” said Roi Carthy, the chief marketing officer of L1ght, a content moderation AI company.
The learning on my end, is better cheaper computer vision would be able to alleviate human traumatization. The flip side is we'd also be able to generate 1000s of human traumatizing videos.
https://www.theguardian.com/world/2011/may/25/china-prisoner...
But how much do moderators spend dealing with that rather BS reports? If you spend any time on Tiktok you'll quickly notice that reporting content is heavily brigaded and weaponized. The general process is:
1. Mass reports, probably from automated fake accounts, mass report one or more videos and/or the creator themselves;
2. As with almost all such systems, a certain number of reports triggers an automated response, such as taking down the video;
3. The creator then has to manually appeal the takedown. Sometimes they win, sometimes they lose. It seems to be really inconsistent;
4. Regardless of the outcome of the appeal, a certain number of reports will trigger a community guidelines violation, possibly locking or even completely nuking the account. The fact that every appeal is won seems to be irrelevant.
This problem is so bad that you can duet a video, say nothing and get reported for hate speech or harassment and lose your account.
What Tiktok doesn't do (like pretty much everyone else) is identify false reports and people who make false reports. It tends to be pretty easy too. Defining characteristics tend to be:
1. A default username ("user123239842");
2. Only follows 1 person;
3. No PFP.
Other sites (eg Twitch) make efforts to at least identify ban evaders. If you don't like someone's content Tiktok should make it hard for you to find it again and when you do, they should just shadowban you. Your comments don't appear to anyone else (which reduces harassment) and your reports are essentially ignored (while appearing to have gone through).
I'm actually amazed at just how easily and how often these systems, which supposedly exist to take down offensive content and protect users, are weaponized for simply saying something someone just doesn't like. This would surely reduce the time wasted for moderators (who often rule inconsistently on these issue) so the truly abhorrent content gets immediately nuked.
In fact, I think conducting this more publicly would have huge benefits. Perhaps you would attract more motivated staff, like pedophiles to go through CSAM?
That second point is kinda daft and would be an absolute PR nightmare at best.
No? The current moderators work from home, what control can TikTok have?
> Carolina, a former TikTok moderator who worked remotely for Teleperformance between June and September 2020, said supervisors asked her to be on camera continuously during her night shift. She was also warned that nobody else should be in view of the camera, and that her desk should be empty, apart from a drink in a transparent cup.
I wonder if a way to avoid traumatizing workers with entire videos would be to capture single frames from videos, and divide them across all the workers with them still rating the frame for content violations.
Obviously, some frames would be worse than others to see, but I would have to imagine it would be drastically lower in terms of impact for each employee than watching video.
FB/IG/Twitter all do this and have for over a decade.
That's true.
I'd call it western hypocrisy.
https://strategic-culture.org/news/2016/10/16/western-hypocr...
Plenty of coverage of this sort of thing at Facebook: https://www.theverge.com/2019/2/25/18229714/cognizant-facebo...
Facebook had to pay up, too: https://www.theverge.com/2020/5/12/21255870/facebook-content...
Running a story about one murder doesn't mean you don't care about the other murders. It just means the article is about one particular case.
* Identity verification for all uploaders
* Suing for punitive damages for every upload that traumatizes a worker
And that's social media, pollution, and climate change in a nutshell.