[murbard2]

> Digital Pokemón Cards. Basically the ability to create digital collectors items.

Benefits from being distributed, but can use a simple, trusted timestamping service given the low stakes.

> E-books that could gain value. You could sell ebooks at a premium and allow people to re-sell them. Because the history of their ownership is recorded you could even see them gaining value if they had been owned by a celebrity.

Celebrities are, by definition, not anonymous, double signing is not a problem. Doesn't need a public ledger.

> Private but public healthcare records for research and usage. Store the health-records publicly as personas but allow individuals people to link it with their identity. This would help with research in completely new ways.

That's a distributed database with signed edits. Since it doesn't require to maintain a mutable singleton, it doesn't need a ledger at all.

> Voting made public but anonymous Make it impossible to fake voting results by making the results available for everyone.

Potentially a good use of a decentralized ledger.

> Artificial Intelligence Using the protocol to create automated consensus models.

Also potentially a good idea. Those two ideas seem to be related to this proposal, http://tezos.com

> Companies without owners Build a company with a political purpose without any owner.

Potentially a good idea.

[ThomPete]

Don't you think it's fair to say that there are enough potential ideas there that are ledger based?

We can always discuss the specifics all of these can be centralized but they benefit from being decentralized.

[murbard2]

Centralized vs decentralized is not a box to check, there are major costs associated with it.

Imagine for instance someone suggesting that a supermarket offers coupons on a distributed ledger. Sounds like a good idea? No, it's dumb. It's dumb because you trust the supermarket to honor the coupon anyway... there's no credible threat model here.

OK so we've figured our supermarket coupon system can be run on a centralized database. Well, you suggest, sure we can, but why not just have it decentralized if we can?

Because it's expensive. It's expensive in terms of engineering to interface with the blockchain, it's expensive to wait for blockchain confirmations, it's expensive to rely on the non-malevolence of a few mining pools. And did I mention there is about no benefit?

However, this is not the most offending use of blockchain technology. You know what, maybe that supermarket gets to federate their coupons to other stores and you can start to see an engineering case. And maybe you're running this in Argentina and there is a credible threat that the government might want to meddle in you coupon scheme, maybe for price control. OK, then maybe you have a case. No, the worse use of blockchain technology is when there is no need for a global mutable singleton.

The whole idea of a cryptocurrency is that you need to have an irreversible consensus. This is the crux of the double spending problem, you need to dissipate heat to be irreversible (yes actual heat). This heat could come from "overwriting" your key, but that's hard to prove, so instead Bitcoin uses the heat from the ASIC. I'm simplifying a bit, but this can be made rigorous.

However, I've heard many proposals where there is zero need to maintain such a consensus. You mentioned medical records, that's one. People have been talking about running webforums or websites on top of a distributed ledger. Most of those proposals are braindead.

[ThomPete]

That really depends on what you see the technology solving.

For instance it can solve the problem of very slow moving governments and allow people to build solutions that might be necessary for the survival of that society then thats actually quite useful to be able to build your own rules that even if they try can' be shut down by a central unit.