Reputations

[tcmay@got.net (Timothy C. May)]

At 8:09 PM 5/15/96, Steve Reid wrote:

>Another problem with reputations is that people are stuck until they get a
>good one. A bad rep can always start over. In fact, a person/company can
>look perfectly nice, but use a different identity for dirty work.

Sure. And they do this today (use cut-outs or subcontractors to limit
reputational damage or liability).

>Naturally, you'd only trust dealings that involve a "nice" identity.

Not necessarily. Many people play the numbers, implicitly trusting the
Mafia to pay off on bets they win. (Many don't think the Mafia has a very
"nice" identity, but your mileage may vary.)

In fact, nearly all of the alleged problems with anonymous systems,
especially the issues of defections, trust, expectation of payoff, etc.,
have parallels in other "extra-legal" situations. For example, the Mafia
and other extra-legal or criminal operations.

Do they sometimes defect (welsh)? Sure. Do they sometimes screw over the
little guy? Sure. Do people trust them just enough to keep dealing with
them voluntarily? Sure. (Before anyone mentions it, there are of course
cases where people are forced to deal with criminal gangs nonvoluntarily,
such as with shakedowns, hijackings of trucks, etc. But a large fraction of
the dealings with the Mafia, Jamaican gangs, Russian mob, etc., are for
market reasons, where a market need for drugs, girls, cheap cigarettes,
gambling, loans, etc., is being filled by players who are outside the
normal legal marketplace.)

>Problem is, young people who are just starting out have a no-reputation
>identity, and would be treated the same as the no-reputation identities
>that are used for screwing people over.

Sure, newcomers always have it rough. Whether the newcomer is Nancy the
Nym, or Eustace T. Collins, III, Esq., just starting out in a law firm, the
newcomer has little positive reputation. (It varies, and the degree may
mean something, just as Nancy the Nym may have some reputation capital to
show.)

I mentioned "postive reputation." There is a real cost in throwing out,
say, 30 years of accumulated reputation capital, and this will not be done
lightly by many. Thus, in a given transaction, a lot may be at stake.

(I don't mean for this brief article to be an essay on the many fascinating
issues surrounding reputation and reputation capital. Cf. my Cyphernomicon
for much more on this topic.)

>Reputations could be very hard to create, and very easy to destroy.

As in real life. People can destroy their reputations by being careless. (I
disagree with your implied point that the opinions of others can easily
destroy a reputation. In real life, it usually takes a lot more than just
bad-mouthing. Even in nym-space, the same is apparently true.)

>Well-known good reputations would be powerful and fairly hard to destroy,
>so it's possible that the big reputations might try to crush little
>reputations in an effort to gain some sort of reputation monopoly. Maybe.

Implausible as a central problem, but all things will likely happen. This
happens occasionally in the real world, as when a Big Reputation (e.g.,
Bill Clinton) belittles and marginalizes a Small Reputation (e.g., Paula
Jones).

There's a vast amount of stuff to think about with reputations. I applaud
Steve for doing some thinking, but I don't think he's yet uncovered
anything especially unique or worrisome.

As a last point, see especially the role of anonymous escrow agents.

A number of years ago the example I usually used was "Ace Escrow--You Slay,
We Pay," to illustrate that an anonymous escrow holder (holding untraceable
e-cash deposited by the purchaser of a murder contract) could pay off a
murderer who presented certain evidence, all without any of the parties
having any idea whatsover whom the other parties were. The problem is then
one of whether the escrow company will simply pocket the money and not pay
off.  First, it can be set up (I think) that the e-cash is uncashable by
the escrow company...but I'm not sure this is needed. A better solution is
to rely on the basic nature of escrow or bonding services: their reputation
capital is much more valuable to them than anything to be gained by
defecting and burning their clients. Except if they are about to retire
anyway...as with the bonded courier who defects to Rio de Janeiro with a
bag of diamonds....the trick is to spread the escrow money around to
multiple escrow agents, and to rely on "escrow testing services" which
periodically ping or test the services....

There are many issues here. I'm not advocating murder markets, just noting
that they provide an easy to understand and fairly "pure" example. If it
can be done with murders for hire, it can be done with nearly anything.

A few years ago, many valuable ideas were contributed in this area by Dave
Ross, Phil Salin, Dean Tribble, Hal Finney, and Robin Hanson.

You might search for their articles.

--Tim May

Boycott "Big Brother Inside" software!
We got computers, we're tapping phone lines, we know that that ain't allowed.
---------:---------:---------:---------:---------:---------:---------:----
Timothy C. May              | Crypto Anarchy: encryption, digital money,
tcmay@got.net  408-728-0152 | anonymous networks, digital pseudonyms, zero
W.A.S.T.E.: Corralitos, CA  | knowledge, reputations, information markets,
Licensed Ontologist         | black markets, collapse of governments.
"National borders aren't even speed bumps on the information superhighway."