Comment by ralston3 | Hacker News Reader

Comment by ralston3 | Hacker News Reader

Parent Full thread

I feel like a lot of people use the term "web3" which in and of itself, doesn't convey too much information. I feel like this question should be asked within the context of web3's components:

- The ability to login to many platforms without needing to create an account on that specific platform (universal login as mentioned by others)

- Being able to use the decentralized ledger as a database -- this is currently only done with art NFTs - but could literally be done with SO much else

- Being able to create dApps that run on their own and are maintained by a community -- i.e., apps that don't pit the user incentives against the incentives for the company owning the apps (dating apps are a good example of this)

- Decentralized storage. Not keeping everything in S3 -- not saying this is better for in-production, but just mentioning with web3 - people are finally open to this type of thing as opposed to just saying "let's just host it on S3".

- Community. As corny and overused as this word is, I know. But specifically finding a project, hopping on their Discord, contributing, asking questions, etc.

Web3 is as much an ethos as it is a "specific thing". It's a rejection of the walled-garden, data hoarding techniques of past platforms (Instagram, Facebook, Google, etc) and a recognition that there is enough pie for a lot of smaller, more niche teams to have a piece.

I agree, Web3 is a buzzword with little meaning. The features I am most looking forward to are:

1. Self-owned storage. I would love it if I could make and use apps where the users owned the storage (in a convenient way). This means that running apps is basically free and you don't have the legal pain of controlling user data. It also means that users can export their data (the just need to figure out how to interpret it).

2. Decentralized hosting. I think it would be amazing if you could keep using sites even if the original host went offline. IPFS is making good strides here. For example if you like my first player chooser you can pin /ipfs/QmWBUag1ynHaZekVHyr1VfMgF6fh64PPP46oiNYCgCPCQK and use it forever. This combined with 1 means that you can keep using apps without any creator involvement.

These two feel so close that they could be used today. There are efforts, but at least for 1 not much is really sticking.

The above 2 are all about single-user data though. Combining data for multiple users like many of your other points reference would be super cool but is a much harder problem, particularly around security and privacy. I think that will be full "Web3" but I see us as a long way off.

> - The ability to login to many platforms without needing to create an account on that specific platform (universal login as mentioned by others)

Dude, how come no one is talking about this instead of cryptocurrency and NFT FOMO bullshit? That's actually useful. When do we get that?

Is it actually useful?

It saves people from needing to create passwords for each service if they don't use a password manager while also eliminating SSO as a data collection mechanism for major internet companies. That's not nothing, but the entirety of this is achieved by signature based authentication, which has existed for years and years but has achieved basically zero adoption.

It is obviously useful, and indeed it could be implemented by readily available methods. If "Web3" was really about getting that done, making the adoption actually happen, I think I'd be a lot less inclined to think the whole thing is just hype-based scams.

Like IndieLogin/IndieAuth?

Isn’t this Google Login, which is already present in many online services?

Do you really want the same key to every door?

If the alternative is a different key for every door, yes. There's no reason it has to be strictly one key for everybody, but I don't need separate keys for most of the things I do online. I'm ok having my bank take one key and all my streaming services take a different one.

Reply on news.ycombinator.com