I equally enjoyed our conversation, that’s why I summoned you back .
As to the differences between our attitudes toward the trust - debate is often not about finding a common ground, but about understanding the other party. I could bring the same doubt as you - you seem to present a very conservative view on a role of trust in communities, which I equally regard as evidently false. However, I’d love to learn more about reasons that make you perceive the same subject so differently from myself.
As to disdain/contempt - your assumption is incorrect. I express my thought that way in other places as well, it’s easy to check. It’s nothing personal. I just don’t find a diplomatic and overly correct style honest enough. While I’m a big fan of Mr. de Talleyrand-Perigord for his wit and humor there must’ve been a reason why his boss called him la merde dans un bas de soie .
Your example with a car is fundamentally wrong, because there are many things that can go… right, can go wrong.
For instance, while we know each other since a decade with that friend, we are in a closer contact only for last few years. And she just “forgot” to tell me that 5 years ago she had an accident and the car was almost totaled. She went unscathed and never mentioned it for some reason. Fortunately, I looked into a VIN database and discovered that fact.
Going along with things that might go wrong. I know that my friend has a heavy foot. Apparently she never learned that there is something between a full throttle and a full stop. While both cars are comparable I assume that most people drive in a more relaxed way. The way she drives certainly destroys car’s components faster than usual…
To sum the above - my friend can act in good will, or act in a way justifiable for her (she desperately needs money, she does nothing wrong hiding some minute details), but regardless of her intentions accepting her offer doesn’t have to be beneficial to me.
While I know her personally since many years I have reasons to doubt in her ability to assess car related matters. Because your story doesn’t state that she’s a car mechanic. I may trust in her judgement when it comes to situations that happened many times before and said judgement was mostly identical to mine. But only if I don’t have time to check it by myself or those matters are of negligible importance.
I might end up buying the car from her not because of trust, but because it’s easier, because I care about her and know that she won’t get a good price from someone else, or because I don’t care about that car and how long it lasts. Or all together.
Let’s now dive into the absurd that you build further in your post. My friend, whom I trust 99% (because she has a life long experience with cars, is a car mechanic, never had an accident, drives flawlessly) has a friend (of course also so professional) that she trusts absolutely (99%). I posit here that 100% is always a fiction. It might happen only with the same person and even then it’s not so obvious. So, 0.99 multiplied by 0.99 gives us 0.9801. Perfect. But perfect situations happen only in a perfect world. Which is not the case.
So, let’s bring some more realistic factors - say 80%. Still high and probably only possible when a relation is really firm and long. We then have 0.8 * 0.8 = 0.64. Still not bad, but the degradation of cumulative trust is already clearly visible. Let’s add another, third, actor with the same trust level. We get 0.64 * 0.8 = 0.512. Just a little over 50/50 chance. From that point on it starts to be more reliable to toss a coin.
Just for illustration how fast it degrades.
I over-simplify above, but you started it.
The fact that the ebay system is not perfect, and that existing reputation systems are “prone to manipulation,” as you say, is not a fatal flaw but rather a design challenge.
It’s neither a design flaw, nor a design challenge. Such systems are designed for two purposes. First and foremost they act as pacifiers. Satisfied clients, assured about their choices, buy more. Second - they help in building average profiles that can be later projected on specific buyers with a fair degree of certainty to target them with other products.
I believe that Holochain provides the tools to overcome those design challenges.
Holochain is the only one I’m aware of that is agent-centric, putting as much emphasis on data provenance as on data integrity, meaning you don’t have to rely on the system to give you “objective” truth about trustworthiness, you can judge for yourself based on who is making what claims.
Blockchains, Holochain including, are just databases. Often smart, immutable, distributed, but databases. It doesn’t matter what built in mechanisms they have while they have no impact on the data that is fed to them. If you put a monkey in front of the best text editor in existence you won’t get any meaningful document.
It’s a problem of an interface (or an oracle). Humans are terribly complex creatures. Our brains, apart from planning shopping, must deal with a plethora of different stimuli that affect the decision-making process. Extracting, using outside tools, any consistent intent or wish from that bunch of neurons is at least problematic if not impossible (due to stochastic factors). The only thing that makes rating systems deliver something more than a white noise is our natural coherence when we act as a group. That coherence however has no guaranteed strength.
Another issue related to the interface is that the electronic system cannot verify who provides an input. The very recent example - by ICMP. It shows fairly well how easy is to manipulate a crowd support.