49 ms·

In my opinion, this is a tarpit idea [0]. It's not that it's impossible, it's that the user expectations are high, the marginal utility of the new features is n

In my opinion, this is a tarpit idea [0]. It's not that it's impossible, it's that the user expectations are high, the marginal utility of the new features is negligible and, even if you had a vastly superior product, you would still need to somehow convince some critical mass of people to migrate over and use it. I think there's a reason new social networks succeed from a fresh set of users rather than somehow convincing people who are already established on FB/Twitter/Whatever to migrate over. In addition, any new service will be fighting against the tendency to centralize, create a walled garden and prioritize tools for advertisers and user management rather than enabling users. There was a recent quote of Jack Dorsey to that effect [1]: """ The biggest mistake I made was continuing to invest in building tools for us to manage the public conversation, versus building tools for the people using Twitter to easily manage it for themselves. """ That being said, I don't think it's impossible but it's almost going to be a "worst is better" solution and one that at least has some traction now, like Mastodon [2] or Scuttlebutt [3] (I'm on Mastodon but haven't used Scuttlebutt). Maybe IPFS thrown in there somewhere [4]? There might even be a way to bootstrap a social web through some web3 solution, though that's pretty speculative at this point (or maybe overlaps with IPFS?). All of the FOSS/libre alternatives have major problems, not the least of which is that they're not tested at scale, have growing pains or just don't have the critical mass that any of the other platforms do (though maybe Mastodon is getting more popular?). I go where the people are. The major feature, in my opinion, is Metcalfe's law. Any other feature might be necessary for long term survival but for to even get started, the value added from Metcalfe's law dwarfs any new feature that people might think up. I'm willing to put up with some pain for a libre/free/FOSS solution but even then, it needs to be decentralized in some way or else we're most likely to be doomed by a repeat of something like Reddit open sourcing their stack, it being a mess and them just close sourcing it again, without any major benefit to the public or users. Put it this way: Why is some idea going to work when so many other people have tried and failed? This is not a rhetorical question. There's bound to be a social network that takes over FB/TWitter/Whatever. What are the conditions for this new social network to take over? A new breed of users? Cheaper compute? Web3 adoption? Cheaper storage leading to easier data distribution? Continually cheaper communication costs? Sometimes it's good to throw stuff at the wall and see what sticks, that's certainly how a lot of startups succeed, but that's not something I'm going to invest a lot of emotional energy into. [0] https://www.youtube.com/watch?v=GMIawSAygO4 https://www.youtube.com/watch?v=GMIawSAygO4 [1] https://twitter.com/BillyM2k/status/1603135805039382529 https://twitter.com/BillyM2k/status/1603135805039382529 [2] https://mastodon.social https://mastodon.social [3] https://scuttlebutt.nz/ https://scuttlebutt.nz/ [4] https://ipfs.tech/ https://ipfs.tech/

abetusk 4y ago
Sorry to reply to my own thread but here's something that actually might change this equation: AI. How can AI help create the social networking site we actually want to see? Language models to filter out hateful content? Boost content we want to see? Generate content we want? Google's search engine initially worked because it was piggy backing off of our focus on quality content by having multiple links to content. There might be an argument to make that Google came at a time when the hardware to do "whole internet" analysis became affordable. Twitter is a kind of "web2" site that piggy backed off of the ubiquity of the cell phone (in my opinion, if other people have differing opinions on Twitters success, I would love to hear it). Now we have (potentially some rudimentary) AI, like language models or stable diffusion. How can we leverage these to drive engagement and add actual value to the user base? What's the next generation of AI models that can be used for this purpose?

abetusk 4y ago

Sorry to reply to my own thread but here's something that actually might change this equation: AI. How can AI help create the social networking site we actually want to see? Language models to filter out hateful content? Boost content we want to see? Generate content we want? Google's search engine initially worked because it was piggy backing off of our focus on quality content by having multiple links to content. There might be an argument to make that Google came at a time when the hardware to do "whole internet" analysis became affordable. Twitter is a kind of "web2" site that piggy backed off of the ubiquity of the cell phone (in my opinion, if other people have differing opinions on Twitters success, I would love to hear it). Now we have (potentially some rudimentary) AI, like language models or stable diffusion. How can we leverage these to drive engagement and add actual value to the user base? What's the next generation of AI models that can be used for this purpose?