I don't think you necessarily need blockchain. Can't you just prove that you are who you say by signing something and sending it to the service? You can just use the protocol.
It's important to remember that blockchains are just public-key cryptography where you have a private key that can sign things and, importantly, everyone knows everyone else's verified public keys. That's it. It solves the key distribution and verification problem that PGP and TLS etc have and this enables a lot of use cases such as universal private communication channels and authentication.
Signing the message is key for this yes but knowing that a certain key is connected to a specific user and that user having the ability to use it to sign verified messages everyone in the world can trust is the real utility here and what makes this universal SSO system work well.
On HN I am Sargos. You know this because I am replying to you and only I can do that with this account. I can also tell you that I'm @JamesCarnley on Twitter but there's no way for you to verify that. If I were using my public key to log into HN and Twitter you would know those are both my accounts and thus my persona is verified across multiple applications. If I were to link my public key to my government's identity database then you'd also be able to verify I am really James in real life as well.
Ethereum provides a robust, secure, and increasingly usable key storage and usage system to everyone which makes "just signing a message" a simple task and not a 10 step process probably involving a CLI. It's worth considering the utility of this and the possibilities everyone having a person public/private key pair allows. My fellow software developers among us likely have their mouths watering at the use cases this unlocks. Here's a pretty good thread about the implications: https://twitter.com/BrantlyMillegan/status/13892701158840975...
Unless you were to upload each and every chat to a blockchain - which is prohibitively expensive - I don’t see the killer advantage over the previous alternatives. Also, many people here are also programmers, developers, and - surprise - hackers, so I am sure we would be interested in the mouth-watering use cases you’re thinking of (I looked at the Tweet thread you linked but it was just an explanation of public key cryptography in general.)
[^1]: Apple also refused to backdoor a terrorist’s iPhone at the demand of the FBI. OpenSea intervened when someone stole assets from a collector (https://blockzeit.com/opensea-nft-marketplace-stops-hacker-f...)
You can improve the UX of key management tools without a global network of redundant computers.
I never said this.
>You can improve the UX of key management tools without a global network of redundant computers
Yet nobody has ever done it until now.
Keybase.io is a quite elegant solution and didn't "fail" due to any fault of its own (the team was acquired by Zoom).
I as a person have accounts on lots of apps but no real way to prove I own all of them. When you use a public key as your identifier then everyone can verify that the entity that owns Sargos on HN also owns Blah on Reddit if I want them to. Essentially you can trust that the digital entity you are interacting with is the digital entity you knew and trusted on the rest of the web in the past.
If you are using a web3 app and see vitalik.eth then you know for a fact that it's Vitalik Buterin. Unfortunately we only know this for sure because he said that is his address in public but there are many identity protocols trying to solve this problem and if you were to tie your public key to your government's identity database then you would be able to prove real world provenance.
2. Your possession of the private key “verifies” your public key, if someone takes it they are now you.
3. Depends on the consensus mechanism but in the best case, “everyone” and in the worst case “coinbase.”
4. You don’t trust them, the system is supposed to be trustworthy with untrustworthy participants, and when that’s not true you will just have to trust the architects of the hard fork.
5. Magical off-chain oracle!
Also GPG doesn’t have a key distribution problem. You can spin up a keyserver or use a popular existing one.
Login/verification doesn't require a transaction though, so is relatively quick. Blockchain in this context can be thought of as a collection of (public) keys.
Even if the technological ideal comes to fruition in a few years (sharded modular proof-of-stake consensus blockchains with zero-knowledge rollups and dedicated data availability layers), it will still eternally remain enmeshed with speculation and scamming. I think there's a narrow time and place for the speculative assets but wouldn't want that interwoven throughout the fabric of everything online.
An idea will come along that enough of us can get behind, that idea will attract money and solve real problems for a while and when they're no longer problematic enough to warrant spending money on the system will collapse back into speculation hell until the next idea-that-we-can-get-behind comes along.
Sometimes it's all about being in the right place, at the right time, with the right amount of hype. Inferior technologies win out all the time.
That being said, if (major if) auth through web3 did take off, I wouldn't be surprised if over time it slowly creeped back toward a solution that doesn't use blockchain since a non-blockchain solution would probably be simpler, cheaper, and faster.