[jMyles]

I'm head-down in writing NuCypher as we speak, but it's obvious to me how to do what we're doing without a distributed consensus mechanism, and I do think that the real world use cases will be pretty rad.

Our system allows an actor (Alice) to select any number of recipients (Bob) in a Policy. Alice can disappear from the network forever, and subsequently, any DataSource can encrypt data, using Alice's public key, which can then be decrypted by all of the Bobs.

That's pretty cool to me. I do think that medical devices / IoT are an obvious use case. I also hope that our tech is used to build selected consortiums of journalists, whom whistle-blowers can then encrypt for only by knowing the policy key.

Another interesting use case is for distributed ops: if you have a number of streams of operational data that you want to share only with a certain number of watchers, presently you need to trust a centralized service to do that.

I'll admit: I'm not really the use case guy. But I am waist-deep in the python over here, and I can tell you we have a good thing going.

[wpietri]

So I'm sure you mean well this is an exact example of my problem with the cryptocurrency/blockchain space.

For years and years I've said, "Yes, that is a pretty cool technology, but what real-world value is it currently providing?" One common answer is, "But it's a really cool technology!" No argument, but that seems to miss the point. Another is, "I'm sure it will be amazing!" Which again, misses the point. A third is, "It might be great for X," but without any real proof that people doing X want the technology, without demonstration that the current alternatives are inadequate, and without apparent recognition that a future hypothetical does not in any way satisfy somebody looking for traction.

Plenty of technologists think they have a good thing going. Right up until the investor money runs out and customers have failed to show up.

As an example, look at 3D movies and TV. 3D has been about to change the way we see things since the 1950s. There is no denying the technology is very neat to technologists. Early adopters even get excited! And then it turns out once again that customers don't really care. This pattern goes at least as far back as the Brewster stereoscope in the 1850s.

So please, don't be shocked that people are tired of blockchain/cryptocurrency hype. That you find the technology interesting does not mean that anybody else will find the (lack of) actual deployed use interesting.

[acjohnson55]

> A third is, "It might be great for X," but without any real proof that people doing X want the technology, [...]

It often seems to me that the mathematical purity of many crypto techs are a poor match for the fuzzy real world requirements. The result ends up being a big pile of abstractions with poor usability and major holes. After all this time, this still applies to basic payments for tangible goods.

[wpietri]

100% agreed. Somebody elsewhere in this thread just did the "blame the user" thing again just now:

https://news.ycombinator.com/item?id=16854568

Blaming the user only works (for some value of "works") in a situation where a power relationship constrains the user. E.g., we've all seen customer service agents dealing with shitty in-house software. They can't easily quit, so they will just accept being told they're "doing it wrong".

But that doesn't fly when the user can easily make other choices. People who get blamed for "not doing enough research" when they have trouble using Bitcoin will probably not work harder. They'll just go back to using credit cards and Paypal and Venmo, which a) work much better, and b) have people who are trying hard to make that work well for them. People whose Bitcoins get stolen mostly aren't going to go and become security experts. They're going to use existing methods, which they generally understand how to secure, and which often have security and anti-theft measures built in.

Sure, they may pay a little more in transaction fees. (Although those fees are often hidden, so they may not notice.) But in effect, those fees are buying insurance. They're buying security teams. They're buying user interface designers who work hard to make things easy. For many people, that's worth it.

[jMyles]

Hmm - it seems, though, that you haven't responded to the use cases I've outlined above. What say you?

[wpietri]

They are hypothetical. I asked for "some practical use where they're sufficiently better than existing technologies that they are displacing existing businesses," which is pretty clearly not hypothetical.

Cryptocurrency and blockchain proponents always have hypothetical use cases. The original Bitcoin paper [1] gives a hypothetical use case, a "purely peer-to-peer version of electronic cash would allow online payments to be sent directly from one party to another without going through a financial institution." This still remains basically hypothetical; even prominent Bitcoin boosters have given up on that vision. [2]

I'm done taking hypotheticals seriously in this space; I've seen too much hype and approximately nothing in the way of results. Maybe someday you'll be proved right about your use case. Maybe you'll have customers who not only buy it but keep using it and come back to buy more. But until then, you should expect people to be skeptical. Previous blockchain/cryptocurrency promoters have, for me and many others, used up all the reasonable benefit of the doubt and more.

[1] https://bitcoin.org/bitcoin.pdf

[2] https://avc.com/2017/08/store-of-value-vs-payment-system/

[querulous]

how is this better than pgp? what does it do that keybase can't already do?

[jMyles]

With PGP (for that matter, PKE generally), you always need the original recipient to stay alive and use her private key to decrypt.

With proxy re-encryption, you don't. So any use cases that involve Alice disappearing while others continue to be able to encrypt for Bob (even without knowing who Bob is or what his public key is) are good ones for NuCypher.