Re: [bitcoin-dev] trust

[Gregory Maxwell <gmaxwell@gmail.com>]

On Sat, Aug 8, 2015 at 6:10 AM, Thomas Zander via bitcoin-dev
<bitcoin-dev@lists.linuxfoundation.org> wrote:
> The idea that Bitcoins very reason for existence is to avoid trusting anyone
> but yourself is something I've heard before, and I have to comment because it
> is a destructive thought. It is very much untrue because we don't live in a
> black/white world.
[...]
> The point was NOT to trust no-one, the point was to trust everyone, but keep
> everyone honest by keeping the ledger open and publicly available.


I think you are drawing a disction that does not exist; rather it's a
disagreement about terminology.

If a system exists that provides high confidence of faithful behavior
even from malicious parties, then it's one this community would call
"trustless" (assuming the security properties are strong enough).
The result is a system where it's possible to trust everyone.

Trust in this case has multiple meanings.   In one meaning trust is an
(well founded) expectation of faithful behavior.  In another it is a
blind reliance.   When "trust everyone" is used, it's speaking to the
first definition, when the trustless definition is used it's referring
to the second defintion-- without blind faith.

A trustless (second def) system allows its users to trust (first def)
everyone, even the inherently untrustworthy (second def).  In doing
so, the considerable cost and inequality created by the maintaince of
trust (second definition) relations is mitigated, and the availablity
of faithful performance increased.  Doing so is a prerequsite to
having a strongly decenteralized system, because otherwise trust
requiremets drive the enviroment towards natural monopolies (as it's
cheaper and more effective for more people to trust (second def) a
smaller number of parties.

Less philosophically, if you're willing to have systems defined by
trust (first definition) (e.g. you do not believe that what I descibed
above coveys value, or hope that witl a small number of very trusted
parties external factors will transform blind faith into a rational
expectation of faithful performance) then there are _much_ more
technically superior ways to structure a system than Bitcoin does
today-- ones that acheive much greater performance, flexibility,
reliablity, and better security (ignoring any insecurity that arises
as a result of the trusted parties strong assumptions).

If one wants to layer trust based things on top of a trust mitigating
framework, they can do so-- and enjoy efficiencies from doing so.
Doing the converse doesn't appear really possible.