Hmm. Putting 1 gigabyte in S3 for 500 years at today's most expensive price ($0.023 per gigabyte month) would be $138. Prices are certain to decrease over time. This says something about the cost of using the Ethereum blockchain for storage.

Would S3 exist in 500 years? Bezos would be long gone. Our tech would be long gone. I'd think about the only thing we could do the same is walk and open our mouth's.

And who knows what the unexpected looks like...Wars, political shifts, etc. Hell, until 2019, I thought toilet paper was unlimited. Once again, toilet paper is starting to look like the currency of the new-world.

:) That's a joke BTW. But is it? I don't want to go back to using squirrels as toilet paper!

Add 1 download per day for 500 years and you will get 21k.

S3 and blockchain have very different use cases right?

Blockchain being for small quantities of data made immutable, used in a p2p manner with as many copies as there are peers.

S3 generally being purpose built for short term (relative to blockchain) storage and price.

Yes, they have different uses cases. But one can easily see an arbitrage opportunity here for building an immortal database atop S3 (and other cloud services) for a lot less money. For $12,000 USD, I could store the same data in S3 for (at the very least) 445,217 years. (Using the example above.)

That makes the value proposition of the Ethereum blockchain as a data store a lot less attractive.

This is so theoretical, I don't think it has any value. Meanwhile, consider the effort involved in destroying all data stored in S3 vs destroying the entire ETH blockchain. One is expensive, possibly only achievable by a national superpower, the other is virtually impossible without destroying the planet.

I fully realize the theoretical advantages of blockchains. I can still store multiple copies in multiple clouds across multiple availability zones cheaper. The original question did not imply the destruction of S3.

Better would be an ENS name pointing to an IPFS address. Then just hope that IPFS sticks around.

EDIT: ENS expiry would bite you. Maybe just IPFS then.

You need to make sure someone continues pinning the data for it to exist on the network. You need to create the right financial incentives for that to happen, which is what Filecoin is supposed to solve.