I was talking to Moxie's cofounder, Joshua Goldbard, the other month about this. They strongly depend upon secure enclaves to achieve it.
It seems like a rather clever use of DMCA related technology, potentially backfiring in the DMCA's face if it is extended to other use cases.
However, using secure enclaves seems like chasing a fading rainbow. The tech doesn't work without trusting the enclave, yet if we trust it, we're re-centralized our trust in chip makers like Intel.
Or worse, if it works extraordinarily well, I'm afraid DMCA related gangs will figure out how to lawyer those use cases out of existence in the same way they lawyer-ed in the tech to begin with.
I'm glad Moxie & co. are trying though, somebody has to! I have been previously very skeptical about the claims made by Signal, though, but as long as the enclave works, then they seem to hold true.
For what it's worth, we also trust the NSA to have written and verified our most important cryptographic hash functions, leading to the first rule of crypto -- never roll your own novel hash function.
I see this as the start of a community effort to verify Intel SGX. I imagine that the process of rolling out a consumer facing high stakes product on SGX involves external verification of SGX beyond what it's gone through before.
Comments
I was talking to Moxie's cofounder, Joshua Goldbard, the other month about this. They strongly depend upon secure enclaves to achieve it.
It seems like a rather clever use of DMCA related technology, potentially backfiring in the DMCA's face if it is extended to other use cases.
However, using secure enclaves seems like chasing a fading rainbow. The tech doesn't work without trusting the enclave, yet if we trust it, we're re-centralized our trust in chip makers like Intel.
Or worse, if it works extraordinarily well, I'm afraid DMCA related gangs will figure out how to lawyer those use cases out of existence in the same way they lawyer-ed in the tech to begin with.
I'm glad Moxie & co. are trying though, somebody has to! I have been previously very skeptical about the claims made by Signal, though, but as long as the enclave works, then they seem to hold true.
For what it's worth, we also trust the NSA to have written and verified our most important cryptographic hash functions, leading to the first rule of crypto -- never roll your own novel hash function.
I see this as the start of a community effort to verify Intel SGX. I imagine that the process of rolling out a consumer facing high stakes product on SGX involves external verification of SGX beyond what it's gone through before.
Care to share what claims you were skeptical about?