Hacker News Reader

Comment by johnchristopher | Hacker News Reader

It tells more about what kind of content you are liking and watching than anything else.

Mine is filled with skaters, mom-daughter dumb jokes, silly lesbian/transgenders sheniningans.

There's sometimes weird things like a priest telling stupid jokes alone in his church and that korean guy doing asmr of burning bread in the oven.

On facebook I saw the death of participants in the events that happened on the 6th but that was in a journalist/news/media analyses group.

I see a lot of good vibes content on Tik-tok but that's what I am looking for on the platform. No doubt I could soon end up on indian train deaths if I wanted. Just like that time on Youtube when I saw that out of nowhere, I was looking for something totally unrelated to trains or india though.

View on HN

People say things like this in an attempt at shaming the original poster but one thing not considered is the location of the person using TikTok.

I live in a big city and most of the feed is highly sexualised, highly political, or highly prank-related.

My friend that lives in a small town about 100 miles away has to essentially send me fun tiktok videos and, no matter how many likes I do of those videos, the base diet of sexual, political, or prank-related videos are always there.

Not a tiktok user, but I know instagram also fills your explore feed with the content that your friends like. So even if you're tailoring your likes to appropriate posts, your friends may not be.

That may just be the initial period after a user has signed up and hasn't engaged with enough content for IG to show anything under "Discover". They may use friends' preferences as a placeholder.

I know for a fact that what's on my Discover feed (surrealist art, UX/UI stuff, Newfie dog videos) is unique to me.

I don't think likes are actually a big signal, it's mostly based on watch time in my experience.

>> one thing not considered is the location of the person using TikTok.

I run into this daily. I'm in the military and so my internet connection at work is mingled with a thousand other military people. Ok, it is accurate to say we are predominantly male and younger, but every time I see an ad does it have to an an anime character? Some of us recently gathered to watch the DC riot footage on youtube. The footage started with a non-skippable ad for some pervy anime game. Try holding a strait face in front of the general as a busty cartoon character in a negligee is bouncing up and down on a ten-foot screen.

I don't know, there's something reaaaaally dystopian and weird and fun but concerning about the chain of events that led to that situation.

I personally use ublock to correct in real time my world lenses.

Try holding a strait face in front of the general as a busty cartoon character in a negligee is bouncing up and down on a ten-foot screen.

Classic! I love real world tidbits like this. I would suggest in the future use youtube-dl https://ytdl-org.github.io/youtube-dl/index.html to download the video to your machine to avoid situations like this.

There are all sort of ways to do this, but they are difficult in the military context. We gathered in one room because youtube/facebook is blocked on our normal office machines. This was an unrestricted podium computer used for totally unclassified stuff like streaming live news feeds. Like most corporate networks, we aren't allowed to just install software, even simple browser extensions have to go through an approval process.

Can you talk a bit about the ambiance and the reactions in the room as the events unfolded ? If you can, that is.

There was no reaction. This is normal. The unfiltered internet is full of inappropriately sexual ads.

I mean.. reactions to the events that took place at the Capitol.

I love how even the DoD won't pay for YouTube Premium.

Note that you can also hold on the video then click ‘not interested’.

I hadn't found this piece of UI yet, will try and do that too, thanks!

> I live in a big city and most of the feed is highly sexualised, highly political, or highly prank-related.

More anecdata: I also live in a big city - London - and I don't see this. I joined TikTok just a few weeks ago and I was surprised at how overwhelmingly positive and fun it all is.

> People say things like this in an attempt at shaming the original poster but one thing not considered is the location of the person using TikTok.

That's a good point. I was wondering about the initial seed of the feed.

It wasn't my intention to shame though.

Until tiktok has something like channels or theme/tag subscription it will remain a problem.

The content your parent comment refers to was (partly) what I saw when I installed it and scrolled through the default suggestions before it had the chance to learn anything about me. Sure there was a lot of harmless and just silly stuff, but the amount of animal abuse like https://twitter.com/noellesdesigns/status/122443946827220992... that scrolled by made me stop using the app.

The OP did say they used it daily for six months though, that really ought to be enough interactions for the app to personalise to the users choice.

I don't think it tells you anything about what the commenter likes. When I created a TikTok account it showed me nothing that interested me. I marked videos as "not interesting" for half an hour before giving up.

I got a few animal abuse videos in an hour or so of swiping. It shows you these random videos in case you are interested. A few of these was enough to uninstall. I tried again some months later, and after awhile the recommendations get a little better, but there is still the occasional piece of awful content that gets thrown in there. In all cases I swiped past as quickly as possible.

Would you trust the algorithms to tailor most content appropriately for a kid?

I think OP’s point is that there is plenty of questionable media on Tik-Tok.

Your point about what they like is irrelevant to that point.

There is just as much questionable media on YouTube. The difference is TikTok's algorithm works a lot harder to show you more of the content it thinks you will spend more time viewing, even if you are not actively searching for it.

So if OP is seeing "loads" of sexualized content, it's because they spent enough time watching sexualized content before swiping for TikTok to think that's what they want to see, and ensure they see more of it.

> So if OP is seeing "loads" of sexualized content, it's because they spent enough time watching sexualized content before swiping for TikTok to think that's what they want to see, and ensure they see more of it.

Even if the picture you are painting of the previous poster is accurate, being horny once in a while is by no means justification for an endless stream of sexualized content.

That's not a neutral quality of the system, but a deliberate feedback loop establishing an addictive behavior. We already know the shortcomings of the human condition and should ask for less footguns, not more.

I've heard people describe a situation in ML (could be hill climbing) where an optimising algorithm will get stuck in a valley. At that point it will try various things in order to get unstuck, often involving a good dose of randomness. What's to say this didn't happen to op or that you really won't start seeing recommendations for Indian train death videos?

> What's to say this didn't happen to op or that you really won't start seeing recommendations for Indian train death videos?

We can never be sure of anything. What if HN start pushing train death videos on the front page and the admin can't tweak the algo because a glitch in the stack prevent them from accessing the backend ?

Mine is full of fitness instructors showing of their fitness challenges,dancers showing of their synchronized dance routines, and oddball humour. I had a moment where some BLM stuff trended but on the whole it seemed to show both sides (albeit a bit the extreme versions of both sides).

If the content is harmful it's harmful...

> It tells more about what kind of content you are liking and watching than anything else.

This sounds very much like victim-blaming, I don't think it's appropriate to side with TikTok on this.

It's not victim blaming. It's widely known that the "magic" of TikTok is that it carefully scans your interactions with content as you consume it, and seamlessly adapts your stream to what it believes you'd like to see. Algorithmic suggestions taken up to 11, and generally as close to mind reading as we've came as a species.

It seamlessly adapts your stream to what it believes will increase “engagement“. That's a different thing to “what it believes you'd like to see”.

Have you ever used TikTok?

TikTok has seemed to present me with shock/outrage content in the name of engagement less than any other social media - far less.

I almost always see content that is only fun.

That has made it much better for long-term engagement and positive for my mental health, rather than other social media's short-term outrage engagement that is wholly negative.

That means the TikTok algorithm designers are half-way competent.

I intended this in contrast to your point above.

It could mean the algorithm designers have decided that optimising for fun is more effective at increasing engagement than optimising for shock value and conflict.

But, this is a different choice than other social media designers have made. It seems like there is some more wholesome motivation.

I don't think the other social media designers deliberately optimised for shock value and conflict. I think that was a side-effect of a naïve optimising algorithm: short term (i.e. until people just quit the social media platform entirely), conflict and doomscrolling are two of the most engaging behaviours.

This may be true for people with short attention span. For me personally sure, I may click on a clickbait article or watch an attractive lady, but it doesn't mean that I want those suggestions even shown.

It's like putting a lot of chocolate with sugar in my apartment: the wise choice is to not even buy anything that contains sugar, as I can't stop myself from eating it if it's in front of me all the time.

Maybe reptilian mind reading. While we might intuitively be very attracted to sexual and gore content, it doesn't mean that's what we want.

I think there's a problem with this though: people aren't always in the mood for the same stuff. The easiest distinction here is 'inappropriate' content from the rest. Sometimes the user is interested in that content, but not every time they open the app. It might even be that most of the time they don't want to see that content.

I experience this with politics videos on YouTube. Sometimes I'm in the mood for that. One evening I'll watch a few politics videos. The next day, my recommendations are filled with politics. Over time my recommendations get back to normal. Then I'm struck with the politics bug again, but my recommendations don't give me an easy way to find more politics content. They're recommending the type of stuff I'd normally watch. The day after that I'm annoyed again about politics being all over my recommended lists.

It feels like I should have separate profiles based on the topics I'm interested in. The platforms don't seem to really support that though. Even different users seem to influence each others suggestions.

Reply on news.ycombinator.com