Hi synx, On 04/03/2010 08:23, synx wrote:
> Ian G wrote: ... >> It is possible to categorise ... but that doesn't mean it is useful to >> do. Libraries catalogue books, but that tells us where to find a book, >> not how good it is. > > Nevertheless, you wouldn't go to the science fiction section expecting > to find a good book on how to cook the best chili in the world (Serve it > Forth by Anne McCaffery). What I'm saying is that you trust that books > in a certain category will fit properly with that category. Otherwise > you wouldn't be able to find anything. Right, but what you are talking about is not trust, or it is a trivial variant of trust. Trust is a complicated risk-based analysis where we rely on others to give us information, etc, and we know that when it goes wrong, there are losses. When you're searching the library, nothing really goes wrong, more than, you don't find the book you're looking for. As well as the downside potential, there is also a certainty aspect. As in, "we know how to do it" versus "we don't know how to do it" debate. We know how car brakes work, and we "trust" they will work. But actually, it is something else, brakes just work, these days, and a failure of a "utility" mechanism like brakes or electricity is not quite a failure of trust. We use the same word, but the semantics differ. We "believe" they work all the time. We don't apply risk decisions to brakes, we've moved beyond that: we know they will work. When they stop working, it isn't so much a failure of trust, it's something worse: a failure of belief. Whereas when we enter the market place for unique, individual goods, like houses, spouses, children, degrees, wars, health etc, we don't know how to predict the perfect outcome, we are in the hands of many and various intermediaries, and the resultant process of leans heavily on trust. Who do I trust, how much do I trust, etc. > For instance the "graphic novel" shelf in my local library is > unfortunately filled by someone who likes those really depressing > nihilist drama comics. So that reduces my trust in their ability to pick > out light hearted friendly comics, and also reduces my trust that I'm > going to find anything worth reading in that section. But it increases > my trust that I'll find yet another book by R. Crumb on the shelves. > Similarly I know the library has an excellent section on native plants, > so I have a very large trust in that library to provide me with > information in that category. In that way I'm forced to categorize trust > not just according to the person, or organization who I'm trusting, but > also in the type of trust I'm ascribing them. Another way of looking at this is that if there is value in the system, and trust is important, it will be gamed. And, if we shift from "trust analysis" to "belief" then we lose. E.g., the financial ratings system in the USA (Moodys, S&P) failed or collapsed because over time they were gamed by the sellers of toxic derivatives. Or, Internet security systems were totally secure until they were worth money, and then they were worthless. >> Also, there is a sort of top-ten winners effect as soon as you succeed > > I call that the "clique" effect. The problem is once you get a group of > people you trust, you stand to benefit from including other members in > your group and expanding, but you also stand to lose from attempting to > include members who are untrustworthy. So there's a point of diminishing > returns, when reaching out to new members doesn't help you more than the > pain of being stiffed by betrayers. As a consequence "cliques" form, > since people become willing to exclude other worthy members from their > group, since that way they can also lock out the trolls. So, the winners seek to exclude the losers, but they are doing it on loose signals such as "don't know this person, therefore bad". So ... the system doesn't scale, and trust is perverted. > What I'm trying to do is reduce the amount of damage betrayal has on any > network of friends. The way I concieve it, members of a clique could > serve as bridges to other cliques, or "cells" if you like. Individual > members put themselves at risk, but upon being compromised the group > overall distributes that loss. So with 10 people you'd only have 1/10 > the risk, but still 100% of the gain from acquiring new members. Which then is easy to pervert, because if you only lose 1/10 of the loss, I can offer you 2/10s to give me the 100% loss, and you're still in profit. Incentives... > In > /that/ way, cliquish groups are motivated to grow larger, thus reducing > the consequences of betrayal, thus offering further motivation to grow > larger. Even in an environment where betrayal is highly likely, cliques > can and should form, but can still have bridges into other cliques, and > in doing so hold society together as best as possible under the storm of > abuse they happen to be suffering from. > >> the ones on the top ten sell disproportionately to the ones off the top >> ten list. > > If 10 people have more of a certain resource than everyone else, then > it's arguable that they should give it disproportionately ot the ones > who do not have that resource. One "top ten" group would exchange > something with another "top ten" group, such that in their exchange both > ended up getting something they wanted. Most groups I've come across would not take the next step and want to exchange anything with another group. > The resources exchanged can > include IOUs from the Federal Reserve, but I really wouldn't recommend > it. That in itself has consequences that would pose an incontrovertible > risk to both groups. (The form of settlement of the trade is really another subject.) >> Right, the PGP Web of Trust is a network in name, but trust isn't quite >> what it delivers. More it delivers a sense of "who met who" and >> therefore likely similar interests. But that isn't trust, it is more >> like loose community. > > Who met who I guess. You're supposed to verify people's identity even > having met them. But that gets me to my other point... Supposed to, sure. But there are no standards and/or quality. So how does one trust? > For the record I do not have a penguin t-shirt. would you buy a t-shirt from this nym? :) >> catalogue what trust meant, instead we created a vector, a message, that >> can be used for anything ... but carries weight. > > So basically what you're saying is... you created money. lol... not really, not yet. In a sense we've created a basis for credit money, or claims to circulate, but we're not pursuing that angle. Although you make a very amusing observation on the structure of money, and who's doing what. I've often wondered about doing p2p money. Technically, it is doable (I would like to say easy, but that would be slightly deceptive). But it raises so many non-tech questions that one would have to try it to see. iang _______________________________________________ p2p-hackers mailing list [email protected] http://lists.zooko.com/mailman/listinfo/p2p-hackers
