[Cryptography] Anonymous rendezvous (was Business opportunities in crypto)

[Natanael]

Den fre 7 maj 2021 22:48jrzx via cryptography <cryptography at metzdowd.com>
skrev:

>
> NewEgg has meta reputation for good curation of reviews, and reviews on
> NewEgg have reputation by being on NewEgg
>
>
>
> So NewEgg is a trusted third party:
>
>
> It acquired a reputation for hosting good reviews by hosting good reviews.
>
> It does not have a reputation for hosting good reviews because a
> Cerificate authority says it is NewEgg.  Thus, no trusted third party
> required.
>
> If they had been hosting good reviews under a bitmessage or Zooko type
> key, or an onion service key would have worked the same way, without the
> people being able to confiscate their good name, and without people having
> the capability to intercept their messages.
>
> What NewEgg accompliehed, can be accomplished on a system that s strongly
> distributed - no connections between participants except those explicitly
> desired; and no trusted third parties
>
> Bitmessage is what you say cannot be done, and it is doing it.
>

How to beat a grandmaster at chess;

Start playing chess under two different pseudonyms with two grandmasters
and copy each player's moves against the other, keeping the two chess games
mirrored.

When one of them wins against the other, you are in control of a pseudonym
which has won against a grandmaster.

In other words, what you're missing is that being Newegg and doing a good
job of curating reviews isn't sufficient to establish *identity* because
their actions and the data which are used to describe them are not self
authenticating. You connect to an untrusted network and somebody presents
this data to you associated with an arbitrary key. That which you attribute
the reputation of a given actor to may not be under their control *even in
the system you propose*.

Establishing that broad consensus of knowledge exists across many entities
without centralization is literally the main innovation of Bitcoin via
proof of work. But like PHB mentioned for his Mesh system, knowing that you
share a common association between a name and a public key isn't enough to
establish an identity, as you still have no trust anchor (although it
simplifies the task). In addition to his Mesh, there also trusted
timestamping and transparency logs (keybase.io uses the latter) which shows
alternative ways to implement this capability through federated entities
(using distributed knowledge as rollback prevention/detection).

Stuff like checking for low entropy names and pseudonym age in a system
like this is only a heuristic at best for detecting a MITM or Sybil attack.
Trust anchors are important.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://www.metzdowd.com/pipermail/cryptography/attachments/20210519/c152d2f9/attachment.htm>