this post was submitted on 06 Feb 2024
208 points (96.8% liked)

Asklemmy

43908 readers
1045 users here now

A loosely moderated place to ask open-ended questions

Search asklemmy ๐Ÿ”

If your post meets the following criteria, it's welcome here!

  1. Open-ended question
  2. Not offensive: at this point, we do not have the bandwidth to moderate overtly political discussions. Assume best intent and be excellent to each other.
  3. Not regarding using or support for Lemmy: context, see the list of support communities and tools for finding communities below
  4. Not ad nauseam inducing: please make sure it is a question that would be new to most members
  5. An actual topic of discussion

Looking for support?

Looking for a community?

~Icon~ ~by~ ~@Double_A@discuss.tchncs.de~

founded 5 years ago
MODERATORS
 

Pretty much in the title. Maybe you wouldn't even use it, but would like to simply see it exist for the sake of having a federated alternative.

For me, it'd be the following:

  • LinkedIn
  • Meetup
  • Tiktok

I am on the first two, but would prefer a federated alternative. I'm not on Tiktok, but would like to see a federated alternative.

I'll admit these might not be a good idea. But as a thought experiment, I'd be curious about the community weigh in on what you all think this might look like.

you are viewing a single comment's thread
view the rest of the comments
[โ€“] jeremias@social.jears.at 14 points 9 months ago* (last edited 9 months ago) (2 children)

Also tiktok really only makes sense with a big algorithm knowing what users want to see. Even if you were to follow many people, with the average video being only about 30 seconds long you won't have much content to enjoy. The whole short form video thing is kinda built on knowing what your user likes and doesn't. I don't know how you could design such a platform without some privacy concerns.

[โ€“] sxan@midwest.social 12 points 9 months ago* (last edited 9 months ago) (2 children)

I don't know how you could design such a platform without some privacy concerns.

Yes, yes you could.

Companies like Google have successfully brainwashed us into believing that algorithms like this can only work on their server farms. The only reason those werver farms are necessary is becauwe they're processing data for millions of people.

We forget that in each of our hands we hold a device that is 5,000 x more powerful than a 1985 CRAY-2, at the time the world's fastest supercomputer. And let's not forget our home desktops and laptops, which are several times more powerful that that.

We each have devices with persistent internet connections that could be at work scanning, categorizing, and filtering personalized content for each of us, without giving any privacy away. It's only because we've been conditioned to be dependent on having our data centrally processed that we believe that's the only way.

Note, it is more efficient to process content centrally, where the data is stored. However, generalized categorization and content tagging with robust metadata and standardized APIs would address the efficiency. Given companies are unlikely to do this and scupper their own surveillance revenue, the next best thing is local, privacy-respecting, smart content filtering assistants.

[โ€“] 1984@lemmy.today 6 points 9 months ago

Are you Richard from Silicon Valley TV show? :)

[โ€“] Zak@lemmy.world 5 points 9 months ago (2 children)

Those sound like good ideas in theory, but your phone's battery would last about 2 hours if you did this.

The heavy lifting, like tagging the content of millions of videos probably needs to be done somewhere other than the end-user's mobile device. Some sorting and filtering of text-based metadata on the user's device to pick what videos to see next is viable though.

[โ€“] sxan@midwest.social 1 points 9 months ago

True, although it would probably not be so bad for the textual content. CPU load for indexing would be relatively low, and the average phone is dumping tons of data over the network to Google, Apple, and whomever else for these same end-result "benefits" already.

But, regardless, ideally, -ou don't do it on your phone. You pay $10/m for a VPS that does it, and delivers it to your phone via push notification + fetch -- same way it's done now, but without the middle man.

It's not a solution available to the average Joanne, although it'd be easy enough to achieve. The problem is that there's no incentive for anyone to make these appliances: most people don't understand what they're sacrificing, or don't care. And while it's a relatively small amount of work, it's a large effort for a few OSS devs to take on, and it'd require at least some support infrastructure, apps, and so on to be truly turn-key for The Public. And so, instead, we have TikTok.

I'm fine with requiring users to tag their own content if they want it to be discoverable. Like if you want to tell people "hey I'm talking about pixel art over here!" just add #pixelArt to your thing.

If you don't want to shout it loud for all to hear that's fine too. Not everything needs to be indexed, cached, and highly available to all who might potentially, possibly want to see it.

Algorithm doesn't have to be a secret engagement sauce. It can just be based on an editable list of the user's preferred tags and keywords with associated weights.

No need to get more complicated than that because you're not trying to juice their "engagement" since their are no ads to show them.

Although I'm not even sure if infinite shorts make sense without a company pulling the strings for their own motive. But maybe it's just not my thing