-
ohshit, this looks great @palashkaria/1482329615259803654
-
so exciting to see a peer-to-peer website like this - provides a useful service, but doesn't actually send data through their servers.
-
which is great because privacy, but also great because it means you can have a useful service like this, but it is cheap enough to run that it can be done on a volunteer basis, and doesn't need to coalesce a business model around it
-
unfortunately decentralization is now a buzzword used to attract money to terrible crypto projects, because it is actually cool when it works (and it is increasingly possible for it to work)
-
I should learn more about WebRTC. feels like that's the hidden force making a lot of this stuff newly possible (at least on the web)
-
also, while we're talking about free decentralized personal file sharing services: @v21/1473352666264780806?s=20
-
the big problem with services like this is NAT punchthrough, which is the result of Internet providers not having enough addresses for you to be able to directly communicate with arbitrary other computers.
-
the fix, depressingly enough, is to have a centralised service that both computers can talk to in order to discover where the other one is. or, for the entire Internet to move to using IPv6, and giving every device a unique address.
-
anyway, here's a good read about the history of that (failed) transition, and the way that a commodity market in IP addresses has formed @apenwarr/1482993783285649408?t=3zhIJaOuA2hg3L1i2OsASA&s=19
-
here is a blog post which more eloquently says a lot of the things i was saying in this thread robinsloan.com/lab/bad-hosts/ (tl;dr - decentralized services are hard because there's no-one to do the NAT punchthrough)
-
kind of makes me want to get together with some friends & set up a Tailscale network and just... send packets at each other. share files using the file sharing services that operating systems still ship with. host websites that are only up when our laptops are open.