CrypTech: Making the Internet a Little Bit Safer
cryptech.is
cryptech.is
I mean, people, let's be real here. Unless you're a hardcore mathematician with programming background in devising crypto algorithms AND you have strong knowledge in signal processing on top of that, how do you know that anything you download/buy in terms of "more private" hardware/software, is actually indeed "more private"?
We're at the mercy of people who keep saying "we care about your privacy and security". Everybody is screaming "we are better than the rest!" and all of us are simply left scratching our heads.
For all we know, even GNU's guidelines on which crypto algorithms to use might be flawed due to an ingenious and impossible to spot subtle mathematical flaw in the said algorithms.
What actual hope do we have?
From a psychological standpoint and my own experience, as you grow older, it becomes harder and harder to resist bribes or being bought out to inject vulnerabilities in a software/hardware you work on. Idealism lasts 5 (maximum 10) years of heavy work during which you squander the seemingly infinite energy of your youth, then you become very tired, then you make it worse with coffee, energy drinks and all-nighters, then you destroy several possible relationships and friendships, you become grumpy and cold, and one day you find that you don't care about what's good for the world anymore.
You also find yourself staring at your ceiling and wondering WTF are you doing with your life. And unless you're one of the rarer human beings who is actually decisive and has the courage to make a U-turn and change your life to exactly what you want from it, then you'll be stuck for life.
Now imagine we fast-forward lots of years later. At 35-45 you no longer care about most (or all) of your youth ideals. All it takes is one phone call or email offering you a 5-digit figure and you're bought out. And if the sum is 6-7 digits then forget about it, you'd sell the entire humankind. We can all pretend to be on high moral thrones but I am sure most people wouldn't think for more than 2 minutes after they receive the said phone call or an email.
So at this point, I simply can't believe anyone when they say they really want privacy and security to be enforced and to try and beat NSA & Co. in their own game. I know I want to, if only out of spite and to show them they're not so awesome as they think, but meh, I am not an electric or electronic engineer. I can't reverse-engineer a chip. I can't devise a new chip. I can't isolate equipment in a room and try and check if it emits signals I don't want it to emit. I can't hack the kernel of my Android phone and kill its services who always drain my battery.
Basically, I can't do almost anything about it. So I am thinking about it on a human level. Humans are easily bought. If not, they are even easier to threaten.
I'm even more at loss trying to imagine how a specialized device can help "securing email, web, DNSsec, PKIs". Well, for DNSsec I can imagine something, but not the others.
Is it to be used as a SSL terminating proxy in front of your services? How could it secure email, for example? How is that goal compatible with not getting "drawn into re-implementation of application protocol layers"?
https://markgamache.blogspot.com/2011/05/why-we-use-hardware...
I'll add that many have tamper-resistance mechanisms of varying security for physical attacks. The better ones also have a trusted path, or dedicated device, for entering PIN's or passwords. The main drawbacks, as link notes, are usability/complexity, closed-source components must be trusted, and very-high cost (eg four, five digits). Projects like this intend to reduce or eliminate effect of those last three.
One might prefer a Pi with a suitable smartcard attached...
It's what my pal "Thoth" is essentially working on minus the Pi part. The EAL5-6 rated smartcards have significant effort put into SW protections and tamper-resistance. Not stopping high-end attackers with physical access in all likelihood but might keep out riff raff. He's putting mail and/or file encryption on them to protect keys against host compromise. Tells me the Ledger Blue is the best of the new, more-open ones as they know what Red-Black separation is, implement some of it physically, have trusted path, and act on some feedback. Check it out.
EDIT: Given this is open source, I forgot to mention multiple smartcards & SW TCB's can be combined in voter configuration with trusted hardware & software to reduce risk of vendor subversion. Unlikely that all the major vendors would work together or be compromised by same group unless threat profile is too big for most to survive already. :)
I wonder: Do they mean private sector cryptographers ?
Jokes aside, I'm assuming they either mean people that build custom encryption into their applications (such as the many messaging apps that we've seen broken over the past few years), or they've worded it somewhat poorly and mean to say 'people that implement cryptography in the private sector'.
Not all cryptographers who work in the public sector do so in secret. Not everyone who works in the private sector shares research openly.