For example, authentication/identity using blockchain technology seems interesting to me.
[1] https://blog.ipfs.io/decentralizing-the-internet-s-root/
Identity on blockchain beyond resolving zooko’s triangle (I.e. stably mapping pseudonyms to pubkeys) typically does not offer an advantage over (partially) centralized identity providers. E.g. putting anything related to government ID on a blockchain is pointless because you might as well have the ID issuer run a central or federated service.
Bitcoin (and it's successors) are a decentralized solution to the double spend problem. If "decentralized" is a requirement, then a cryptocurrency can be an excellent solution. Other known solutions (like "just trust everyone") don't scale well (an understatement!).
If "decentralized" is not a requirement then there are other solutions to the double-spend problem.
1. How much security do we need for this transaction?
2. For this block of transactions?
3. Can we have varying levels of security?
What we are ultimately trying to know is, is it profitable to double spend? If the answer is yes, then no one should be transacting, and anything they do transact is a gamble. If the answer is no, there's much less risk, and the farther the answer is no, the safer the transaction.
But we likely don't need as much security as we do now. In fact, it's very easy to handle at a user level.
A. Put the transaction on the blockchain.
B. Wait for additional blocks.
C. If your transaction is huge, wait for more additional blocks. Wait a week, a month. This is an important transaction. When the blockchain has not reverted for such a long time, finally do the transaction in the real world.
If you are transacting with yourself, you don't need any security at all. If you are transacting with a bar and the bar knows you, much like a tab, you don't need much security. If you are transacting with an anonymous person, you want high security.
So security is currently handled on a blockchain by blockchain basis, where it should be a transaction by transaction basis. What if you can't wait a month? Then you should put heavy fees on that transaction and it should be put in with other higher fee transactions.
4. How does security change with transaction size? Does it make sense to have limits?
5. How does security change with total transaction block size? Does it make sense to have limits?
6. When the coin price goes up, how does that change these factors?
7. When the mining reward in coins drops, how does that affect security?
8. When Bitcoin, in 100 years, moves to 100% transaction fee system, what happens to security? What would the transactions fees look like?
9. In 50 years when the block reward has been halved 5 times, either the price of Bitcoin moves to $1 million, or the security drops. Is that okay?
If someone hacks into my account, my bank will likely reimburse me for any losses. There are controls in place to stop them from stealing it al. If someone just transferred it to their own account, I could use the courts to get it back. If I use a credit card to pay a scam, I can do a charge-back.
With cryptocurrencies, it's just "bye money". They're not going to fork for you. There's no logic to stop people in Russia transferring my entire account. It's too hard to find criminals to get any relief. It's just your fault for using cryptocurrencies and not having 100% perfect security.
Not to mention, the chance of my bank stealing all my money is nil. They'd be sued and arrested, and the FDIC would reimburse me anyway.
What if it's taken thorough wrongful civil forfeiture? Will you be able to afford to navigate the courts to get your assets back after they've been seized?
Cryptocurrencies don't have to replace all existing financial instruments to be a useful hedge. They can just be one tool of many to diversify your risk profile.
Especially consider that for people living outside the developed world, the risk of having your assets stolen through legal channels is going to be much higher.
You had to cover the "too big to fail" system.
On btc no.
Funny - I have never seen someone in the traditional financing system do that and not get caught. Closest things were ponzi schemes like Madoff (and he did get caught)
I have, however, heard of dozens of exactly this incident happening with cryptocoin exchanges (wallets, services, the most general term for all combined slipped my mind)
If you want to trust a conglomerate of folks who have a vested interest in keeping you small and poor to solve your problems, then there is no advantage. If you do not, then the advantage is exactly that it is not centralised.
Imagine a world tmr where bitcoin is everywhere. You ll have mining conglomerates in the cheapest electricity providers working as cartels to help each other and protect their interests (so can envision centralizing a blacklist of addresses they dislike, coordination on software version etc), de facto cartels of whales (like todays "billionaires" people describe as a collaborative force), and large lenders who concentrate capital and therefore still can buy out whatever they want (today called banks).
How do you think the blockchain solve any of that?
It's not rocket science. When you keep your wealth in monetary terms, you don't get to have it in physical terms. Money represents how much real wealth you could have if you spent it. It's deferred consumption. There is no real wealth in something that didn't happen.
Essentially yes :)
There are some stateful blockchains which allow changing the passphrase but of course there is no 'forgot my password' feature; a malicious actor who manages to steal your passphrase could potentially use it to change your passphrase and lock you out of your account but at least you don't have to live in doubt if you suspect that your passphrase was compromised.
The idea behind the blockchain auth approach is that you would only need one passphrase for all services but you would have to secure it carefully since losing it would cause you to be locked out of all services associated with that account.
So you're gambling bigger: less likely to have your identity stolen, but complete loss of everything to do with that identity, with no recourse, if it is stolen.
I'm not sure I like those odds.
This is only marginally better than SSO with Google/Apple/Facebook/whatever. It dodges the "Google closed my account" problem, which is a meaningful improvement. But I'm not fully certain why you need the public key to be part of a blockchain for this to work. Couldn't you just publish your public key identity on any number of considerably less wasteful systems? As far as I can tell, the only advantage here is the mapping between the name "UncleMeat" and my public key would be able to be reused across services... but in so many cases I don't want my identity to be consistent across services.
- It allows you to use the same passphrase for all services without compromising security since your passphrase never needs to leave your own machine; it is never sent over the wire. You just send signatures to different services.
- The cost associated with purchasing tokens needed to initialize an account on the blockchain would serve as a spam prevention mechanism; an alternative to SIM cards which most centralized services rely on today as a cost barrier to limit the creation of spam accounts.
- Superior integration potential between different services/systems provided by different companies/groups since they can all refer to the exact same account on the same blockchain and provide new ways to unify data between the different services.
I'd wager that most people don't care about this much, though it is real. Still, this can be achieved with signature-based authentication that doesn't require the tremendous waste of BTC. Just... tell the service your public key and sign a message when you sign up. Your key pair doesn't need to be associated with some distributed system for that to work. This has been around for decades and not exactly caught on.
> It allows you to use the same passphrase for all services without compromising security since your passphrase never needs to leave your own machine; it is never sent over the wire. You just send signatures to different services.
This is true for all signature-based authentication, which does not necessitate a blockchain.
> The cost associated with purchasing tokens needed to initialize an account on the blockchain would serve as a spam prevention mechanism; an alternative to SIM cards which most centralized services rely on today to limit the creation of spam accounts.
This sounds like an anti-feature to me. Especially since it'd be foolish for a service to only support authentication via a blockchain. Imagine telling my grandmother that she needed to buy some btc in order to sign up for the service that hosted my wedding photos.
This is different because your grandma would have to buy the tokens once and signup once and she will have access not only to the 'wedding photo service' but a large number of other services. It's more like paying a small amount of money to get access to an ecosystem of services. Like how people pay money to buy an iPhone and this gives them access to the Apple App Store and all the apps therein.
There will also be network effects associated with the blockchain price going up with adoption. People who signed up early to the right blockchains will have priority access to certain software ecosystems and exclusive services which are only accessible to the richest among them (for example).
This seems even more like an anti-feature TBH. Imagine not being allowed to use the future equivalent of basic services because you were not "early enough". Sorry grandma, no wedding photos for you because you didn't buy these specific three coins out of the thousands started every month.
I cannot possibly imagine a world where services will choose to only permit authentication via some expensive blockchain such that they deny themselves access to markets outside of the global rich.
And telling my grandma "don't worry, the coins you don't understand will go up in price" is not going to help.
This is the core question: why is signature-based authentication using public keys associated with BTC wallets superior to signature-based authentication using public keys not associated with BTC wallets? As far as I can tell, the only benefit here is that now my identity on my photo sharing service can be linked to my identity on my wine rating service. Why do I want that?
I'm convinced that decentralized communities will eventually corrupt and destroy all centralized services and agencies by manipulating their employees and members... Because they can. People will always find a way to create new problems to make themselves necessary. In these political games, decentralized anonymous communities have the upper hand. Unlike centralized projects, you can't corrupt decentralized projects. What can be corrupted, will be corrupted.
To understand why it's the future, you need to look at the big picture. You're assuming that centralized auth providers are secure and trustworthy. This is not going to be true in the future. Trust in institutions and organizations has been eroding and will continue to erode.
“Services will make more money because blockchains are trendy” isn’t a meaningful difference.
Doesn't seem like that to me. If I get banned I can just move my cash to another account and start again, and assuming multiple services does this I can amortize the cost over all of them.
The rest seems to me like something mTLS could solve better.
You don’t need blockchains for this. Ordinary websites can implement this. They usually don’t, but, they could?