4 ms·
Yawn. OK, I gotta recycle this from John Markoff:[0] > "True Names" by Vernor Vinge (1981): "The basic premise of that was, you had to basically hide your tru
by mirimir 8y ago
Yawn.
OK, I gotta recycle this from John Markoff:[0]
> "True Names" by Vernor Vinge (1981): "The basic premise of that was, you had to basically hide your true name at all costs. It was an insight into the world we’re living in today … We have to figure it out. I think we have to go to pseudonymity or something. You’re gonna participate in this networked existence, you have to be connected to meatspace in some way."
Indeed. At least pseudonymity.
We need to compartmentalize. In meatspace, we must be totally unremarkable. We must make only routine communications, for entirely utilitarian purposes.
Everything even vaguely sensitive -- that is, everything that we want kept private -- must use secure channels. And if it's truly sensitive, it must be ~anonymous. With the degree of anonymity proportional to the level of sensitivity.
I see no other option.
0) http://renaissancechambara.jp/2017/06/14/oprah-time-true-names-by-vernor-vinge/ http://renaissancechambara.jp/2017/06/14/oprah-time-true-nam...
- bostik 8y ago> Everything even vaguely sensitive -- that is, everything that we want kept private -- must use secure channels. I see this in a different way. The starting point should be that everything is secured. If you use the secure messaging only for non-mundane stuff, it stands out. So you want to add as much noise into the use profile of the secure channels as possible, to make it that much harder to detect when anything of value is being transmitted. This, however, I completely agree with: > And if it's truly sensitive, it must be ~anonymous In the context of secure channels, that adds that the use of ~anonymous methods must not be any different from outside the channels. Preferably not even from the inside, because you have to assume that a portion of such a system is compromised at all times.
- mirimir 8y ago> The starting point should be that everything is secured. If you use the secure messaging only for non-mundane stuff, it stands out. That's a good point. But consider, how do I use "secure messaging" with my meatspace bank? Or with my wife, or my meatspace friends? Sure, I use HTTPS. But trying to get any of those to use truly secure channels would be pointless. And even worse, it would attract attention. For noise on secure channels, I torrent random stuff. That all goes through a VPN service. But among the torrent traffic, there are indirect VPN connections.
- bostik 8y agoVery good questions, and something I believe we as a society AND as a profession are only starting to realise. The truth is, people everywhere say they want security. But that's a white lie. What people really want is convenience. That is also the wedge we as a developer community can work with. I truly believe that if you can make the secure option also the more convenient option, you won't be able to keep people away.[ß] As far as usability goes, secure and/or hardened software is still atrocious. The point about meatspace bank is a really good one. Well, as it happens, WhatsApp based banking is a real thing. I found one startup doing that[0], but even before this I remember that somewhere (Philippines? Indonesia? Kenya?) WA was already used for bank customer support. Now, is that somehow more secure than a well done, well maintained, properly hardened banking app from a dedicated big player? Certainly not. But it's hell of a lot better than a MVP from a company that might be barely in their second year to a basic banking license. It has the benefit of adding useful business traffic to a reasonably secure transport. Obviously it still depends on a known-malicious third party for routing, but then again, for a truly secure communications system you have to assume that you're dealing with compromised and/or actively malicious infrastructure. Which makes it a very good first step. For any real progress, we will need at least another decade of research into usable, latency-sensitive, information-theoretically noisy communication systems. Secure and ~anonymous message routing is going to be a damn hard problem. (Without the latency part, and assuming very cheap bandwidth, something like a probabilistic K-anonymity based batched multicast might be a reasonable starting point.) Even then, the usability has to be baked in and considered from the start. > For noise on secure channels, I torrent random stuff. That all goes through a VPN service. Now that's a really nice countermeasure. Well done! ß: Global censorship and blanket bans with aggressive enforcement notwithstanding. 0: Company is apparently called "Absa".
- taejo 8y ago> 0: Company is apparently called "Absa". FYI, Absa is one of South Africa's largest banks, formed in 1991 by the merger of several large, long-established banks, and by now owns all of Barclays' African operations. It's not a startup.
- stakhanov 8y ago30 years ago, when anonymity was the default on the internet because of the low standard of trust that was universally presumed, this worked remarkably well. I'm thinking about IRC's golden age now. There is a small number of special-interest communication media (in software development, for example) where this still works today, like hackernews or freenode. But for the vast majority of people and types of communication acts, anonymity is basically no longer an option: The presence of non-anonymous media like Facebook has displaced innocuous communication from media like IRC that do allow anonymity. People aren't going to go looking for kiddie porn on Facebook but IRC still enables that kind of information-seeking behaviour. People doing foodie-posts are not going to do that on IRC but go on Facebook instead. -- This means that IRC has become a really scary place to be if there's a possibility you might get deanonymized, and generally not a place where you will find like-minded individuals unless you are one screwed-up individual. On the flip-side of the coin, at Facebook etc, you are going to be the victim of surveillance capitalism. So: Compartmentalizing doesn't work. Unless anonymity is the default for everyone and used for both sensitive and non-sensitive communication, it doesn't actually provide a solution to the problems of surveillance capitalism. And the majority of people are always going to be too dumb not to be tricked into deanonymizing themselves most of the time.
- mirimir 8y ago> So: Compartmentalizing doesn't work. Unless anonymity is the default for everyone and used for both sensitive and non-sensitive communication, it doesn't actually provide a solution to the problems of surveillance capitalism. And the majority of people are always going to be too dumb not to be tricked into deanonymizing themselves most of the time. Well, it "provide[s] a solution to the problems of surveillance capitalism" for me. And for anyone else who cares. Everyone else will do what they want, and get what they get. Other than offering helpful suggestions from time to time, it's just not my problem. I used to think that it was, but I was a fool.
- pdimitar 8y agoThe more people opt into anonymizing tech the more noise the adversary organizations have to sift through. If there are not many of us, we are making it easy for them. I see your point but IMO your parent commenter has a good point as well.
- Izkata 8y ago...y'know, that phrasing reminded me of something from various fantasy books: To know someone's true name is to have power over them. It's like a twisted version of Clarke's Third Law.
- mirimir 8y agoThat's exactly how it is in True Names. Utilities and daemons are actually like demons in the fantasy genre. Or like knowing the true name of a genie. It's like casting spells. Hannu Rajaniemi does something very similar with invoking utility fog stuff.