> I think the main class of problems it solves is it lets multiple applications, which may be developed by completely different entities at different times, share data with very strong guarantees.

It doesn't, though. That's bullshit. Sharing data between web applications is not some sort of grand challenge that required new technology to solve.

Not bullshit, I've seen it happening already. One team set up http://hicetnunc.xyz/ to host niche NFT-s on the Tezos blockchain. Later they abandoned the whole project and meanwhile another team started to build a new platform on the same dataset. https://objkt.com/

This was pretty cool because all the Hic en nunc generated content over their existence is still accessible through Objkt.

Sharing data is of course possible since stone tablets, but doing it without any form of collaboration or approval is probably unique to crypto.

> Sharing data is of course possible since stone tablets, but doing it without any form of collaboration or approval is probably unique to crypto.

To the extent that crypto doesn’t require collaboration or approval, neither does Wikipedia. Or libraries.

Or, indeed, stone tablets — Pharaoh 𓃹𓈖𓇋𓋴 didn’t give anyone permission to go raid his tomb, and yet https://en.wikipedia.org/wiki/Unas

I'm not trying to swindle you. I'm trying to have an intelligent conversation. You don't need to call all my points bullshit :)

It depends on the level of guarantees you want for that data. If we're communicating, and you're okay with me changing my APIs, or even changing the data itself in ways we didn't previously agree upon, then you 100% don't need a blockchain. And if you don't have a lot of entities that need to coordinate on the data, then there might be an easier solution.

And introducing a blockchain to the equation changes nothing about that! It doesn't magically make applications interoperate with each other, or make them consume each other's data, or force them to maintain API compatibility. At most, it provides a shared, public location for them to store data -- and even that has some very awkward limitations associated with it.

A lot of the blockchain boosterism I see online comes from a place of assuming that the blockchain represents a solution to a problem, and trying to work backwards from there. Avoid that.

Maybe we're just talking past each other, but this is the entire point of a smart contract platform. I deploy code to it, and you now have a guarantee forever on the interface of that code, as well as its implementation logic. If I'm running the same code on my server and you're hitting my API, then I can change the interface, change the logic, or mutate the data in ways not specified by that code.

So really when I say "applications" I'm referring specifically to other smart contracts.

Smart contracts that cost money to run on a virtual machine with less power than a Raspberry Pi. In order to access a DApp someone needs to buy into some cryptocurrency, providing some early holder the ability to cash out. The DApp user also has to trust that the smart contract doesn't have a stupid bug that will drain their wallet. Oh and thanks to the low transaction rate of a blockchain that DApp might never execute or have a flash crash type spike in cost.

It's a bunch of rent seeking bullshit on top of what could just be a web API. Storage on a blockchain is stupid expensive so there's oracles everywhere pointing to off-chain resources with zero guarantee of availability.

Blockchains aren't adding any value. They're only extracting rent and adding friction.