I feel like the CEO of Blackrock pumping #Bitcoin is a concrete sign that it's failed to exist as a useful currency to advance liberty. Some would argue it was never intended for this, but even now there are people who will sloganeer Bitcoin fixes this in the face of censorship and financial repression.

#MoneroOrBust? More and more I start to side with those that argue that currency is the wrong battle in the first place when it comes to attempting to hold the powerful to account. (pun intended)

@gabriel Huh? All currencies and securities are susceptible to manipulation.

If you want to avoid that, use the ones that have the least amount of attention.

@realcaseyrollins@social.teci.world
To be fair, there are many who erroneously believe that bitcoin can't be manipulated. While it is possible to maintain sole custody of your own digital assets, you naturally can't prevent others from playing shell-games and other schemes.

My overall point is largely in response to much of the web3 & nostr world that acts like in 2023 Bitcoin is still in early-days and ready to revolutionize liberty online any day now. Examples like this to me are concrete proof that this couldn't be further from the truth.

Much of the "blockchain tech" companies seem identical to big tech in culture and approaches. They just use the simplest impression of being different to sucker people in. This is why all else being equal, I think the fediverse is objectively superior to nostr and other web3 attempts at the same thing.

Doubling down, I agree with you but arguably everything is on the radar these days.

@gabriel @realcaseyrollins

1. blackrock could just be signaling loss of faith in the dollar and other reserve currencies, in which case their support is _reinforcing_ their belief that bitcoin will work as advertised
2. web3 is pure marketing there is almost nothing there, the few things that are there that are worthwhile (like ENS) are damaged and not enhanced by association with the label
3. the fediverse isn't better than anything until it gets out from under centralized dns and client-server

@Moon@shitposter.club
I'd argue client-server isn't inherently bad, unless you're talking about user auth in which case I get your meaning.
With the client server model you cansmooth out the limitations between different devices. Like having your "big box" store all your media and use thin jellyfin/kodi clients on other devices.
@realcaseyrollins@social.teci.world

@gabriel @realcaseyrollins allow me to clarify, I mean that right now activitypub IDs are URLs which point to individual accounts on a server (of which there may be only one, which is slightly better.) this doesn't need to be eliminated completely, just AP implementations need to be able to handle pure P2P users.

My other point is a little more provocative, but I think that technology that has a hard requirement on DNS at any part of the process is not viable freedom-securing technology at all.

@gabriel @realcaseyrollins Work toward this end is already in progress. Mastodon probably won't support it, or not for a very long time but people that need it just won't use Mastodon.

@Moon @gabriel @realcaseyrollins

the dns requirement is because of https, and the problem with dns is not necessarily dns itself but rather the reassignment problem. there is a more fundamental issue with https uris in that they are dependent on the authority and path components never changing. you can add a layer of indirection for the path component, so you can put the actual file wherever you want and then the HTTP server does the translation/routing from path to file.

but you can't generally add a layer of indirection for the authority component without centralizing *somewhere*. at minimum you are just punting it to some proxy that you trust to stay alive longer than yourself, and that proxy becomes the new point of centralization. think PURL services, or the ATP "Placeholder" Server. this proxy is a name server that acts as a second lookup table that redirects/routes to the current data server, which acts as a lookup table that translates/routes to the file or backend.

you could have a third layer by externalizing the resolver, and then an https bridge could be used as an entrypoint. think IPFS gateways. this gateway bridges/routes to the resolver which fetches the resource from some data store. but then you need to dedupe by the external identifier, or devise yet another layer to tell you that two https uris are the same resource.

if you wanted to stay within the https network, you could have such a dedupe server that kept track of all known locations... or have some distributed data store. most likely you will end up with webfinger playing the same role as dns, and allowing you to dedupe resources by their subject/aliases, and pulling this information from that data store. but that data needs to be authoritative, which means you end up with root servers just like DNS, and you've essentially just recreated DNS on top of HTTPS on top of DNS on top of TCP/IP. except it's more of RNS, a Resource Name System than a Domain Name System.

arguably you could eliminate a lot of this complexity if only the IANA would also assign you an authority *number*, which would be assigned once and would never lapse. think DOI on a research article. you would then *alias* all domain names to this canonical authority number. and your browser or user-agent would keep track of canonical identifiers within the authority of this canonical authority number.

but going back to what's most realistic or feasible... maybe we end up with centralized name servers for minting/resolving https: uris. maybe they have some kind of DHT backing them. i can't really see much more happening than that.

@a @gabriel @realcaseyrollins completely removing control choke points is extremely hard and maybe impossible but minimizing them is very possible. the reason I didn't say dns had to be gotten rid of for AP is that is impossible or close to impossible while retaining http compatibility, completely impossible if you want phone apps because phone vendors don't let you incorporate tech that bypasses these things. things like bitcoin or ethereum are decentralized in many more places but still centralized in practice on the development teams and cults of authority surrounding people like spooky money skeleton.

@a @gabriel @realcaseyrollins dns is a threat today because registrars now can and do regularly pull names for stupid reasons or government pressure. it is a win to move from systems like dns to systems where people can pack up and leave en masse, as hard as that is, if possible.

@Moon@shitposter.club
Your post inspired me to refresh my understanding of opennic. Which could be a drop-in solution to DNS woes.
Setting up a certificate authority that plays nice with it seems like an interesting idea. Because apparently letsencrypt cannot

@a@pl.nulled.red @realcaseyrollins@social.teci.world

@gabriel @realcaseyrollins @a my position is only replace dns with something decentralized, replacing it with just a different authoritative root just causes problems

@Moon @gabriel @a @realcaseyrollins Just get rid of DNS. If people can memorize phone numbers, they can memorize IP addresses.

@xianc78 What about vhosts? One IP-address can host a plethora of things and the only way to tell them apart if the domain name they use 🤷
@Moon @gabriel @a @realcaseyrollins

@m0xee@social.librem.one
Wanted more services? For a laugh? They were called PORTS! 😆
@xianc78@gameliberty.club @Moon@shitposter.club @a@pl.nulled.red @realcaseyrollins@social.teci.world

@gabriel Good point! Probably won't work so well with things we have now. I'm not sure if it would be possible to run a Fedi instance on non-web port with any software we have today — they guess how to get to the instance the user came from by the domain name. You can probaby put the port name in the .well-known hosted at standard web port, but I'm not sure how that would work.
@realcaseyrollins @Moon @a @xianc78

@m0xee @gabriel @realcaseyrollins @a @xianc78 if you can't run on a standard port then you probably can't use .well-known but you can put a nonstandard port in the URL and it typically works just fine. there are servers in my database running on nonstandard ports.

@Moon
Sure, but still, in theory, how would servers communicate if we don't have domain names, would we have IDs like @ name@195.34.88.21:32742?
Would that even work? I'm not knowledgeable enough about ActivityPub, can't tell 🤔
@xianc78 @gabriel @a @realcaseyrollins

@m0xee @Moon @xianc78 @gabriel @realcaseyrollins anything you could http get and post (or similar semantics) would work

@a
Sure, URLs have a standardized way to specify the port, but IDs aren't URLs — if I try to look up your ID from my instance, it still connects to known hard-coded ports, if I put a non-standard port into the ID, would Pleroma (or other software) know what to do with it? I think not 🤷
@xianc78 @gabriel @Moon @realcaseyrollins