Quote from: AnonyMint on August 21, 2014, 12:35:17 PM

Quote from: kelsey on August 21, 2014, 06:10:06 AM

Quote from: rpietila on August 21, 2014, 05:32:40 AM

Money cannot have an issuer that must be trusted.

However could a genuine trustless system even be possible?

Based on designs I've explored, I believe yes. Except for adding new features, you need to trust a developer to take control. Once you are willing to lock the features into stone, it can run on autopilot if the features are resilient enough. When it eventually fails, you design a new currency learning from the past mistakes.

If we agree that the semantics of trust are "confidence placed in an agent by making that agent the nominal owner of property to be held or used for the benefit of one or more others", then I entirely agree with AnonyMint that software agents can be trustless.

We do not need to have confidence that a software agent will do the right thing if we can inspect the source code and its peers can verify each other peer's correct behavior.

This is how Satoshi's Bitcoin operates, indeed that is how any enterprise financial system works. Years ago while working at a bank, my code was routinely checked by a data security officer. Daily runs of the bank's software had a staff of people in the Controls department that verified the accounting inputs and outputs of the system. The bank did not need to trust that I would perform my job correctly, rather others would verify my work before it entered production, and others would verify the results of the calculations.

Customers, on the other hand, trusted the bank because we were subject to audit by the board of directors, by the State government and by the Federal government. And if the bank failed despite the audits, then customer deposits were insured to a certain limit by Federal insurance.