This is a special post for quick takes by Mckiev. Only they can create top-level comments. Comments here also appear on the Quick Takes page and All Posts page.
Zero Knowledge Proofs allow proving that you are a part of a certain group without revealing who exactly. This can allow credible whistleblowing. For example, if some ML company issues a list of it's employees linked to ethereum wallets (or any other cryptographic key pair of choice), then any of it's employees would have an option to whistleblow on unsafe practices, while being 100% sure it's impossible to know who made this post.
This seem awesome to me and I wonder if anyone else shares my fascination and think this could be useful
A couple people, including me, do share your interest. See ring signatures for a mechanism.
I'm not working on this. I know someone who might be. If you're interested in working on cryptographic-mechanisms-for-whistleblowing or you have nonobvious ideas, let me know and I'll let my friend know.
sismo.io has an opensource implementation based on zero knowledge proofs and ethereum wallets. I think ethereum wallets are a more convenient asymmetric cryptography implementation then e.g. pgp, hence it would probably be easier to get people to use them.
I have an interested-user-level knowledge of internals, but I'd be very keen to dive deep in case it could be useful.
Do you know of a group that is actively interested in such a system? I doubt it's possible to convince people to do it otherwise, since key management remains painful if done safely.
This is an unusual use-case in terms of most ZK software, but luckily it's easier. (most of the difficulty of wrangling predicates/circuits is capturing business logic in a way that's consistent with traditional compilers, whereas the whistleblower user story would only need very simple data structures that are fixed in time).
I can imagine cases where diffpriv is a bit more to the point than ZKPs, namely with a sufficiently large number of employees you might want to say "the whistleblower was somewhere in this table, and we have a bunch of metadata that ordinarily we'd be able to sleuth over to pinpoint which individual, but we can't".
sismo.io has a working opensource implementation that allows proving your inclusion into groups of choice without revealing anything else. (I'm not associated with them to be clear) User knows exactly what he proves and can easily determine the list of people who could prove the same statement. (this is my dilettante understanding)
Probably it can be implemented with a more efficient algorithm, but I'm not sure whether this optimization is worth the time to implement.
I find tax codes in the US (and basically every other country) very frustrating, not because the tax is big, but for all the bureaucratic burden
I thought I could circumvent the whole problem by putting most of my hard-to-deal-with assets into a charitable fund, and then not worry about taxes there at all. However, DAFs don't quite solve the issue since they are very limited in where I can invest. There are also options such as "Charitable Trusts" and "Private Foundations," but they seem to create lots of overhead.
Is there an option that I'm missing? In an ideal case, I would like to have fully functioning brokerage and crypto exchange accounts where I can make any transactions tax-free, given that all of the money I earn must be donated in the end.
I don't think you are really missing anything. Your best bet might be to find (or found) a DAF with similarly-minded people . . . but that DAF is still going to have to deal (in the US at least) with state-law prudent investment laws for charities. I don't know which states are more open to adventurous investments.
What you're describing is closest to a private foundation, and Congress put tighter rules on private foundations because it perceived certain downsides.
What if anyone could buy a certified speed camera that can be installed on a building a or a moving car?
Wouldn't it instantly solve all traffic violations and so massively reduce death toll, if the hardware is cheap enough, which is probably the case in 2023?
Zero Knowledge Proofs allow proving that you are a part of a certain group without revealing who exactly. This can allow credible whistleblowing. For example, if some ML company issues a list of it's employees linked to ethereum wallets (or any other cryptographic key pair of choice), then any of it's employees would have an option to whistleblow on unsafe practices, while being 100% sure it's impossible to know who made this post.
This seem awesome to me and I wonder if anyone else shares my fascination and think this could be useful
A couple people, including me, do share your interest. See ring signatures for a mechanism.
I'm not working on this. I know someone who might be. If you're interested in working on cryptographic-mechanisms-for-whistleblowing or you have nonobvious ideas, let me know and I'll let my friend know.
sismo.io has an opensource implementation based on zero knowledge proofs and ethereum wallets. I think ethereum wallets are a more convenient asymmetric cryptography implementation then e.g. pgp, hence it would probably be easier to get people to use them.
I have an interested-user-level knowledge of internals, but I'd be very keen to dive deep in case it could be useful.
Do you know of a group that is actively interested in such a system? I doubt it's possible to convince people to do it otherwise, since key management remains painful if done safely.
Thanks!
No.
This is an unusual use-case in terms of most ZK software, but luckily it's easier. (most of the difficulty of wrangling predicates/circuits is capturing business logic in a way that's consistent with traditional compilers, whereas the whistleblower user story would only need very simple data structures that are fixed in time).
I can imagine cases where diffpriv is a bit more to the point than ZKPs, namely with a sufficiently large number of employees you might want to say "the whistleblower was somewhere in this table, and we have a bunch of metadata that ordinarily we'd be able to sleuth over to pinpoint which individual, but we can't".
sismo.io has a working opensource implementation that allows proving your inclusion into groups of choice without revealing anything else. (I'm not associated with them to be clear) User knows exactly what he proves and can easily determine the list of people who could prove the same statement. (this is my dilettante understanding)
Probably it can be implemented with a more efficient algorithm, but I'm not sure whether this optimization is worth the time to implement.
Do you know about groups who would be interested?
I find tax codes in the US (and basically every other country) very frustrating, not because the tax is big, but for all the bureaucratic burden
I thought I could circumvent the whole problem by putting most of my hard-to-deal-with assets into a charitable fund, and then not worry about taxes there at all. However, DAFs don't quite solve the issue since they are very limited in where I can invest. There are also options such as "Charitable Trusts" and "Private Foundations," but they seem to create lots of overhead.
Is there an option that I'm missing? In an ideal case, I would like to have fully functioning brokerage and crypto exchange accounts where I can make any transactions tax-free, given that all of the money I earn must be donated in the end.
I don't think you are really missing anything. Your best bet might be to find (or found) a DAF with similarly-minded people . . . but that DAF is still going to have to deal (in the US at least) with state-law prudent investment laws for charities. I don't know which states are more open to adventurous investments.
What you're describing is closest to a private foundation, and Congress put tighter rules on private foundations because it perceived certain downsides.
Decentralizing speeding cameras
What if anyone could buy a certified speed camera that can be installed on a building a or a moving car?
Wouldn't it instantly solve all traffic violations and so massively reduce death toll, if the hardware is cheap enough, which is probably the case in 2023?