[publiush]

Web 3.0 utilizes crypto in order to build something more than just the cryptocurrencies built in the past. A blockchain is useful in a decentralized system where a shared ledger is required. There are several different types of projects that are branding themselves Web3 -- and I think they are all correct; Web3 is a totally new web built by the people and controlled by the people.

> encompensate technologies like edge computing which are decentralising ownership of compute and data back to devices.

I've been working on just that [1], and I think many other projects are as well. If you stop focusing on the 'value' of 'coins' in the market and, instead, narrow in on the merits of the underlying technology, I think you'll quickly find that all the building blocks are there to get to this decentralized ownership model we are all seeking so eagerly.

[1] https://github.com/publiusfederalist/federalist

[giantrobot]

The decentralized nature is immaterial if actual useful data/computations can't be performed by the network.

Ethereum's chain is about a terabyte and the total compute power is less than a single raspberry pi. So you can't store a meaningful amount of data on it without insane fees nor can you do much in the way of processing. This means anything interesting exists off-chain.

If interesting things exist off-chain...there's not much utility in having a blockchain at all. It's just a slow expensive ledger with no off-chain authority. Even if a transaction points to some content addressing P2P resource (IPFS etc), that content only exists while someone is hosting it from some machine(s) somewhere. When it stops being pinned (or conceptual equivalent) the record on the blockchain is worthless.

[publiush]

If it stops being seeded, then the content is worthless.

[giantrobot]

Seeding focuses on recent or popular content with a particular preference for popular recent content. Seeding is a LRU model so does not reliably serve long tail content. The age or popularity of content is not necessarily a measure of its value.

[derbOac]

Not to suggest blockchains don't have a role in a decentralized internet, but on the other hand aren't there a lot of decentralized resources, made or to be made, that don't require crypto or blockchain at all?

I think this is where some of the confusion and skepticism comes from. Often in W3 discussions there seems to be a push to find a way to use crypto rather than an explanation of a resource, period. If a decentralized product includes blockchain or crypto, I'm immediately skeptical and expect a good explanation for why the product can't exist without it. Usually it seems like the product doesn't need it or could be built on some other option.

I'm all for decentralized and federated solutions, but they do tend to be slower and more resource-intensive. Blockchain most of all.

[publiush]

> but on the other hand aren't there a lot of decentralized resources, made or to be made, that don't require crypto or blockchain at all?

Absolutely. In the project I linked to, I'm using DHT and torrent protocols to deliver decentralization.

That said, we run into an issue in a decentralized system called Zooko's trilemma [1] when it comes to naming. Names need to be human readable, secure and decentralized. Using public keys, this isn't made possible.

> Often in W3 discussions there seems to be a push to find a way to use crypto rather than an explanation of a resource, period. If a decentralized product includes blockchain or crypto, I'm immediately skeptical and expect a good explanation for why the product can't exist without it. Usually it seems like the product doesn't need it or could be built on some other option.

In the case here, I opted to use the handshake naming protocol. Handshake's use case makes sense as decentralized networks are subject to sybil attack [1], so a free for all naming system doesn't make sense. Handshake is one of the blockchain projects I admire specifically because the blockchain project solves a problem that has limited the success of decentralized systems until now, and the purpose of the coin is to limit the ability for an attacker to take over all the names.

> I'm all for decentralized and federated solutions, but they do tend to be slower and more resource-intensive. Blockchain most of all.

I think this depends on the functionality. In the case of the handshake blockchain, lookups are faster since data is local. On the other hand, you're absolutely right as DHT is very laggy.

[1] https://en.wikipedia.org/wiki/Zooko%27s_triangle

[2] "A Sybil attack is defined as a small number of entities counterfeiting multiple peer identities so as to compromise a disproportionate share of the system."