The early version of Facebook platform was essentially meant to be a social OS. For various reasons, FB/Twitter/LinkedIn all locked down and strictly limited their APIs. That’s one of the most compelling reasons for new crypto social networks. Devs can’t be deplatformed.
santi
 @santisiri
25 Jul 2019
Replying to @balajis
i’m beginning to think twitter as the os lately.
7
13
1
64
Doesn't have to be crypto (although that's fine too) - eg, ActivityPub or Indieweb networks have the same properties. Developer access was one of the underlying driving forces for building them.
1
1
Yeah. But here’s an interesting factor I think about frequently. Basically, getting global consensus on one integer (user count) in an implicitly adversarial distributed system like Mastodon was a challenge. Crypto solves that problem. ansuz.sooke.bc.ca/entry/335
2
2
Why is a global consensus on user count a required (or even an important) feature? That doesn't even model real-world relationships, and plenty of systems work completely fine without that.
1
Great q. I’d argue that many social apps make use of global consensus on at least *some* values. Could be upvotes vs user counts. If you can get global consensus on one byte, you can run it again to get global consensus on N bytes (possibly with some upper limit on N).
1
3
Replying to @balajis @benwerd
And I'd argue that global consensus is unrealistic and unnecessary, and pursuing it comes at the cost of too many other useful properties. Local consensus is more useful and easier to achieve, and still results in perfectly functional systems.

Jul 25, 2019 · 6:55 PM UTC

1
Replying to @aaronpk @benwerd
Yeah. I think the right model may be to use Postgres or traditional DBs for *most* of the system, and a blockchain for the parts people will fight over. So things like upvotes, likes, follower counts — and of course money. We haven’t seen systems consciously built this way yet.
1
1
What I'm saying is I actually don't think a global number of likes on a post is even a good idea. What's more interesting is how many ppl within my own network have liked a post. Otherwise we're just mirroring Twitter along with all the negative behaviors that come with it.
1
1
more replies