It should be obvious to everyone now that centralized platforms (like Apple's App Store) might be generous and foster innovation during their growth stages, but then become extractive and limit innovation in later stages. After all, they have total control of the platforms. This is why decentralization matters.[1]
The problem is that "decentralized" protocols often end up becoming de facto centralized because they lack key economic features.[2] And centralized platforms fill the gaps by offering those economic features — at the cost of giving them total control of the network.[3]
This is why the most innovative and most interesting web applications occur in the browser, which is open and permissionless — and where web domains are mostly decentralized in nature. The only way to ensure a platform retains something close to its original economic properties is to ensure it remains decentralized, but it's hard to resist centralization.
I know HN hates it, but there is a class of protocols that use cryptography and Byzantine fault-tolerant consensus mechanisms and are designed to resist centralization... and they have explicit and verifiable economic properties. They're known as crypto, and involve blockchains. And this appreciation for decentralized-by-design systems is colloquially called web3.
I don't see hate for protocols or cryptography on HN. But I do see
(a) skepticism about the actual decentralized (vs distributed) nature in practice (especially where the topic is financialized tech or even an end run around all the learning that's gone into the social codebase about finance).
(b) realize that decentralization presented by things like the web are essentially "good enough" that capabilities are strong, personal or entrepreneurial latitude is pretty wide, and obstacles are few.
EDIT: Thanks for this comment — I edited the original for clarity.
I'm talking about crypto as in Ethereum, which are designed to resist centralization. HN wrongly thinks of crypto as a financial system (which it can do, for good and bad reasons thanks to its permissionlessness), instead of a family of methods for making composable, decentralized protocols with built-in economics (blockchains).
But why do these protocols need built-in economics? No user is going to want to have a wallet that slowly drains as they use the web -- it's just not convenient and will be thoroughly rejected by the mainstream.
Compare this to something like torrenting - people give, people take. Sure, indexing is centralized, but once you have the index, it's all p2p. That's the direction decentralization should be taking, instead of the next generation of microtransactions where they're required for operation.
You now too have a wallet that slowly drains while you use the internet. Web3 wallets are just better integrated into it.
Also, there is a misconception that the cryptocurrency based web is like dial up. It isn't. What it is is:
1. a way to decouple the ownership of the digital goods from the service that gave them to you
2. a way to actually do privacy preserving computing
3. an open layer to build added value things that can interact with assets and services others have created (think web2.0 in the bookmarklets era but your bookmarklets can talk with any site)
Which of my wallets is draining as watch YouTube/TikTok/Twitch or read HN/Reddit? I have guesses on what you could mean (data, attention, etc), but some clarification would be nice.
1. There's no question that ownership is going to be a hot-button issue going forward, but I don't see how decentralization solves that problem (i.e NFTs being on ten different chains).
2. How does it preserve privacy? If I share data with a third party, what's to stop them from copying it over and ignoring my request to cut it off?
3. What is that added value? If you're talking about data portability that's certainly an argument in favor of decentralization, but that has a mess of competing & conflicting standards to wade through (see #1)
I actually meant your normal wallet or bank account. They have a tendency to go towards zero even if we don't do anything because we already are paying monthly to be connected to the modern world. But yeah, attention and data are also good example of ones value draining away while using the web.
Re 1: most L1 blockchains have bridges to each other now. This means that I can move assets from one ecosystem to another quite easily. If one discounts the ethereum network, moving across chains is actually quite economic as well.
Re 2: in web2.0 sharing your data with another party was the only way to play. In the decentralized internet it is not. There is a lot of interesting work going on in this domain actually.
Re 3: I'll use an example from the gaming world but the concepts are transferable to non gaming 1 to 1. If Activision decides to create digital collectibles in their games today, I as an indie Dev cannot create a game where a user can log in and have their Activision stuff available in my game. In web3 these assets are encoded as state on an open platform so anybody can integrate them in their applications. In the case where NFTs are 1st class chain state, like in Cardano, you are even guaranteed that you can trade your NFT on whatever platform you choose, be that an indie game or the originating marketplace. Is it clearer now what the added value is?
Built-in economics doesn't mean users pay per use. A popular design path is to try to build systems where data that goes on the blockchain is very limited[1] and easily subsidized. It's also very easy to design systems where data writes and computation are subsidized by clients or the system itself.[2],[3],[4]
The reality is that in current networks, which are totally financial but only implicitly and in a way that is opaque to most users, computation and data costs are still subsidized. Crypto networks open a huge design space that simply makes the economic flows explicit — and not necessarily borne by users.
> But why do these protocols need built-in economics?
I believe, in this time of humanity where economy matters, that Blockchains have built-in economics for two reasons:
1. They all started copying Bitcoin, which by nature is an economic system (this is the goal of the Bitcoin's paper).
2. Reward by money is a common way to attract a sustainable amount of active users that keep the network alive without evil intention appart from making "quick bucks". In contract to volunteer networks such as Tor, where a lot (even a majority) of nodes are run by gov. agencies with the goal of identifying users, popular blockchains are still run properly since users are focusing on money and not destroying the network.
>No user is going to want to have a wallet that slowly drains as they use the web -- it's just not convenient and will be thoroughly rejected by the mainstream.
You don't need the mainstream, you just need the cool kids. The mainstream will follow---or not.
As for why anyone would want this, there is a very compelling reason---no ads.
The existing internet is sufficiently permissionless for the vast majority of potential activities. Where it isn't, you start to reach into the boundaries past which there are often reasons for permissions to exist.
> HN wrongly thinks of crypto as a financial system... instead of a family of methods for making composable, decentralized protocols with built-in economics
A system with "built-in economics" is a financial system.
> HN wrongly thinks of crypto as a financial system (which it can do, for good and bad reasons thanks to its permissionlessness),
That's how it's implemented in real life, so that's what people's takeaway is going to be. What was the last dApp anybody used that wasn't a pay-to-play game, or crypto marketplace?
You may as well say communism is theoretically sound, but people will continue to get hung up on the fact that's it's real-world implementations have always led to despotic leadership.
I'd be fine with maximally rights granting patents, all but sovereign corporate charters, and other final boss economic and political arrangements, provided that it all gets burned down every 7 years.
The water bowl a good oxygenation for the gold fish to thrive.
It's called dweb. Sometimes people do get it confused with 'web3' but that was just the name of the massive 2-billion-dollar shit a16z took on everything.
Dwebcamp just extended ticket sales if you want to go talk to real dweb people: https://dwebcamp.org/
I'm familiar with dweb, but it seems that they're just preparing to relearn the lessons of the early web (federation doesn't work; you need mechanisms to maintain Byzantine fault tolerant consensus; explicit economic incentives are required if you don't want the system to develop out-of-protocol centralizing economics; etc.).
So I'm just working on a CSS thing that barely qualifies as dweb. But the people I talk to that are into building solutions are well aware of the problems and tradeoffs. I don't disagree with your assessment. But the relearning is like our way of mining for genuinely new lessons. We're quixotic but totally self-aware about it.
The problem is that "decentralized" protocols often end up becoming de facto centralized because they lack key economic features.[2] And centralized platforms fill the gaps by offering those economic features — at the cost of giving them total control of the network.[3]
This is why the most innovative and most interesting web applications occur in the browser, which is open and permissionless — and where web domains are mostly decentralized in nature. The only way to ensure a platform retains something close to its original economic properties is to ensure it remains decentralized, but it's hard to resist centralization.
I know HN hates it, but there is a class of protocols that use cryptography and Byzantine fault-tolerant consensus mechanisms and are designed to resist centralization... and they have explicit and verifiable economic properties. They're known as crypto, and involve blockchains. And this appreciation for decentralized-by-design systems is colloquially called web3.
[1]: https://cdixon.org/2018/02/18/why-decentralization-matters
[2]: https://www.youtube.com/watch?v=WGfS6pPJ5jo
[3]: https://knightcolumbia.org/content/protocols-not-platforms-a...