4 ms·

I honestly don't get it. With the number of IT billionaires out there, why isn't there just one who wants to solve this problem? We already know how, in princi

I honestly don't get it. With the number of IT billionaires out there, why isn't there just one who wants to solve this problem? We already know how, in principle - capability-based, formally verified OS kernels like EROS and CapROS have shown us the way. The next step is building a real world OS and user environment that carry these principles through. Yes, if would take 10 years at a minimum. But it wouldn't even be a charity, even if you went the OSS route! There would be massive commercial opportunities arising out of the product. If you have the money, and it's profitable and world changing in 15-20 years, why not do it?

hiddencost 3y ago
The billionaires aren't coming to save us.
Veserv 3y ago
Because you can just lie and claim you make secure systems like Microsoft, Apple, and Google have been doing. Much easier and you get to have a billion features in your "secure" product so it supports more uses than actual secure systems. There was a small pile of systems sufficiently secure for actual high security work during the regime of the TCSEC Orange Book and the early era of the Common Criteria. But once Microsoft and the other consumer IT companies lobbied to get serious security certification requirements removed or reduced since their security was inadequate to bid on new projects and they found that unfair, they sucked all the air out of the room for actual high security commercial systems. And, even if they did decide to pursue it their every instinct is completely incorrect. They have all made their fortunes on entertainment and consumer software and principles like "move fast and break things" that are antithetical to the development of secure and reliable systems; they do not have the foggiest clue how to solve problems in security or even how to compose secure components into secure systems. That is not to say that nobody is doing so, you can just look to what is actually being deployed in domains that actually demand high security and high reliability systems and have the testing, auditing, and verification to establish conformance like aerospace and certification requirements like the Common Criteria Separation Kernel Protection Profile (SKPP) which required formal specifications, formal proofs of security, and a clean NSA penetration test (i.e. the literal NSA could not find any vulnerabilities). However, the market for things that actually work is not as large as you might think given how much people crow about security because again, you can just lie. The chickens have not come home to roost yet because cyberattacks are only just starting to be a serious problem. It takes a while for 18 year old kids to bootstrap worldwide criminal enterprises able to attack millions of companies. I mean, even Zuckerberg had VC funding and it still took them over a decade to saturate the world. You have to cut the cybercriminals some slack for taking a few years to become a actual crisis. Give it another 10 years.

hiddencost 3y ago

The billionaires aren't coming to save us.

Veserv 3y ago

Because you can just lie and claim you make secure systems like Microsoft, Apple, and Google have been doing. Much easier and you get to have a billion features in your "secure" product so it supports more uses than actual secure systems. There was a small pile of systems sufficiently secure for actual high security work during the regime of the TCSEC Orange Book and the early era of the Common Criteria. But once Microsoft and the other consumer IT companies lobbied to get serious security certification requirements removed or reduced since their security was inadequate to bid on new projects and they found that unfair, they sucked all the air out of the room for actual high security commercial systems. And, even if they did decide to pursue it their every instinct is completely incorrect. They have all made their fortunes on entertainment and consumer software and principles like "move fast and break things" that are antithetical to the development of secure and reliable systems; they do not have the foggiest clue how to solve problems in security or even how to compose secure components into secure systems. That is not to say that nobody is doing so, you can just look to what is actually being deployed in domains that actually demand high security and high reliability systems and have the testing, auditing, and verification to establish conformance like aerospace and certification requirements like the Common Criteria Separation Kernel Protection Profile (SKPP) which required formal specifications, formal proofs of security, and a clean NSA penetration test (i.e. the literal NSA could not find any vulnerabilities). However, the market for things that actually work is not as large as you might think given how much people crow about security because again, you can just lie. The chickens have not come home to roost yet because cyberattacks are only just starting to be a serious problem. It takes a while for 18 year old kids to bootstrap worldwide criminal enterprises able to attack millions of companies. I mean, even Zuckerberg had VC funding and it still took them over a decade to saturate the world. You have to cut the cybercriminals some slack for taking a few years to become a actual crisis. Give it another 10 years.