• 2 Posts
  • 868 Comments
Joined 1 year ago
cake
Cake day: June 14th, 2023

help-circle









  • Do you think the same about lemmy?

    I think it depends on how the federated sites are administered going forward. We’ve already seen bigger sites - like Threads, for instance - try to integrate into the overall ecosystem. And I could see a future in which one of the larger instances - a .world or .sh.itjust.works - is too much for a handful of amateur admins to handle. Hand off the instance to a venture capital firm and you could see rapid enshitification.

    I just have a lot of trouble explaining how it works to people who aren’t tech savy…

    I’m reasonably tech savvy and even I’d struggle to tell you exactly how it works. How is .world hosted? Is it load-balanced or otherwise optimized? Who controls registration and which other instances does it integrate with? How do you find a list of active instances to federate against? Who do you even talk to in order to federate with another instance? What does the API look like and which instances allow you to crawl them? How do bots integrate with the environment and what can an admin do to limit them? No idea.

    There’s a bunch of things I think I should be able to do but I can’t. For instance, signing into .world but only surfing content that’s hosted on .sh.itjust.works.

    There’s also a lot of petty politics. Admins deciding on a whim who to block, whether it be individuals or whole instances. Waking up one day and suddenly not having access to a dozen of my favorite subs, because two admins are feuding, is not particularly fun. I never have a problem like that on BlueSky or Instagram.


  • https://en.wikipedia.org/wiki/Dot-com_bubble

    This goes all the way back to '98, when the original slew of start-ups gobbled up investments only to flop a few years later. Web2.0 had its own bubble burst starting in 2008, taking down a host of the early social media ecosystems (MySpace, Yahoo, and Geocities, most famously). Huge upfront investments with the promise of explosive ROI that took far longer to materialize (or simply never did).

    A great deal of the valuation in these firms was built on lies and bullshit - misreported user activity, overly optimistic monetization estimates, and outright accounting fraud.

    2020 gave us what looked like was going to be a third Crypto bust wave (FTX being the big industry leader leading the charge). But the pivot to AI appears to have bailed a lot of the bigger investors out. We’ll see how long that lasts.





  • Mastodon is more of a protocol than a single service. It succeeds/fails on those terms, in the same way the old Web1.0 protocols did. Which is to say, you can’t enshitify a thousand micro-sites at once like you can enshittify one big site that’s under central control. But you also can’t do things like navigate, search, and socialize efficiently.

    Mastodon is successful in large part because it isn’t. When you let a single cartel of corporate psychos run a Mastodon account like they would a Twitter or Facebook, you end up with Truth Social (literally just a Mastodon branch instance).


  • The key factor in Digg’s demise was a flawed design that was too easily abused by users. Digg had no controls over user verification, so individuals could game the system by creating multiple accounts to artificially inflate the number of votes for their own content. Because Digg displayed content in order of popularity, most visitors saw and voted only on content that was already popular. This system created a vicious cycle in which a small number of dedicated users could push their own content to the front page and thereby gain more followers, allowing them to more easily repeat the process. As Digg grew, so too did its problems related to power-hungry users cheating and gaining undue influence over content.

    Sounds like the same problem that every centralized social media ecosystem suffers from. The big difference between Digg and Reddit was that Reddit successfully monetized the “push me to the front of the queue” algorithm rather than engineering around it.