4 ms·
"trust" is one of those words which has contextual meaning. I actually do "trust" signal, for low risk things. I don't routinely do high risk things and so I ca
by ggm 3y ago
"trust" is one of those words which has contextual meaning. I actually do "trust" signal, for low risk things. I don't routinely do high risk things and so I can make a cost/benefit decision facing who might exploit what I am doing and the consequences for me.
Other people live in the margins of what either their own, or some aggressor society wants and information from/about them is key. I can understand somebody having a heightened sense of what should be the foundation of "trust" in these circumstances. But that said: simply using signal is probably putting a giant red flag over you in some circumstances. Or, PGP. Any self-installed cryptographic technology you are associated with, puts a marker against you for some problems.
I don't "trust" closed source. I prefer that people have review, and that objective groups of people I trust in a personal sense (because I know them, no matter how vaguely, from online lists, and IETF wg and the like) tell me from their position of contextual knowledge, and I admit some authority, where trust can be well placed. Some of them are American. Some of them work for agencies related to the state, like the FCC and in some cases the NSA. I again have to re-frame "trust" into the sense of what it means when somebody who works for an agency like those tell you something. (others of them are not american, do not work for these agencies, or sister agencies in five eyes or other economies. But I observe, if you are competent to speak about cryptography you almost certainly DO engage with people who work for state enterprises like NIST and the NSA, even if in other economies. Its just the way the world IS: academics who work on cryptography have a circuit of relationships which includes the mathematicians inside these agencies)
I don't "trust" Moxie quite as much as I did, because some of the actions around Signal confused and worried me. And, I saw little to no desire to engage. Well, he doesn't owe me anything, financial or moral. I have no direct relationship, and in the end if he decided trying to assuage a million edge users like me (edge user not edge lord) was a loosing game, I can't fault his logic but the other side of that logic is I don't "trust" him as much as I used to.
I do trust Meredith. I have met her (once or twice) and she seems to me to behave in an open, direct manner, and when she says things I see them reflected in what she does. And, I trust the kinds of statements she has made about signal and the kinds of things I see emerging around governance of something like signal.
In due course I do hope she and the board can say something about source code, and future development, and how we can re-build or build from scratch some reasoned trust in the code, and its cryptography and the trust in systems behind the code.
- tptacek 3y agoYour trust in Moxie doesn't much matter anymore, because he's not involved with the project. It is absolutely not the case that cryptography engineers generally interact with NIST employees, and certainly not NSA employees.
- ggm 3y agoIndeed. But, it's foundational for why people have reserved some judgement about the software and the service. I didn't know Phil Zimmerman but I "trust" PGP because of the amount of eyes on the code. That was the substantive point here about Moxie. Unless and until somebody audits and reviews, I would believe code he wrote remains in the application suite and head-end services? Should I "trust" that implicitly? > It is absolutely not the case that cryptography engineers generally interact with NIST employees, and certainly not NSA employees. Day to day? no. Communications are confined to matters of substance between them. But I see NIST people at IETF standards meetings, Mailing to lists. I see engineers who submit algorithms to the NIST beauty contests talking online about who they discussed things with. I knew researchers in Australia who were engaged in NSA related study and it was simply understood no technology entered the building with them, to go and do face to face but the fact they met was not secret, and there are lots of people in bodies like IETF who meet these people on an as-needs basis. Around the time of IETF being held in Beijing a number of people wound up being asked or told not to take their normal laptop with them into that economy. A lot more than I would have expected simply because of commercial in confidence reasons of risk. "generally" might be too sweeping. It's not forbidden. It might be unusual.
- tptacek 3y agoPhil Zimmerman also has very little to do with the design of modern PGP, and the influence he did have was malign (if not deliberately so). Signal is so well studied it won the Levchin Prize at RWC. I don't understand why you believe cryptography engineers generally participate in standards bodies. A comparable percentage of cryptography engineers do that as do software engineers with software standards bodies.