Hacker News new | ask | show | jobs
by kristjansson 1488 days ago
Great! I have 132GB of photos, and would like to store them on the blockchain to ensure I always have access to them. How much will that cost me? Show your work, and round your answer to the nearest million.

Maybe too much snark, but there’s a real tension there. “Store it on the blockchain” means “store a copy on every single node, forever”. It’s understandable that incurs some costs, and that one has to store only the most essential information. Maybe, there are some paths forward, like sharded chains?

1 comments

No. The blockchain is not a data store. The blockchain is needed only when you need:

- global view of the state that is taken as truth by all participants (consensus)

- where anyone is allowed to make operations to change the state (permissionless)

- and all operations that change the state are agreed by all participants (trustless)

You don't need the blockchain to store your music collection. Any content-addressable filesystem can do that for you. But you do need a blockchain if you want to make a decentralized "top-40" list.

I actually agree with all of that. I don't agree that storing that sort of data on the blockchain supports statements like "the users are in control of their own data". You're make a distinction between assets/content/etc. and 'data' that I don't think most users would recognize as a difference. If I make a playlist on decentralized-Spotify, and then that service goes bust, it's weak comfort that I'm able to move a list of song-hashes unless I (or another service I can move to) have access to the corresponding content. I can only exert as much control over my data as I over the off-chain content store[1].

What categories of data do you see that (a) average users care about and (b) would want to control beyond the life of a service provider but (c) doesn't depend on the availability of assets/contents/etc. cannot be viably stored on-chain?

[1]: Sure, FileCoin and dependent services offer some control and guarantees over off-chain data I guess.

> What categories of data do you see

Again, it's not about the data. Data is cheap. It's the meta-data and what you can do with it that is interesting. It's about the new classes of applications that might come up when the underlying data is not stored by any middleman.

To illustrate: the "Pirate Bay catalog" is many times larger than anything offered by any of the streaming services, but the majority of people still prefer to pay for the services. Github is the destination of almost every open source project, and companies are paying a good amount of money not for the storage of data, but for what it can be done with it.