Even with a “new type of blockchain” (whatever that means), “permanent” isn’t something I’d trust a startup with. Even “lifetime” deals from profitable established companies tend to run into horrible problems a few decades later [1], when unexpected trends appear. Average company lifespan is now in a low number of decades [2], and startups tend to be at the lower end of that curve just by the nature of being more experimental.
This is less about trusting the company and more about trusting the viability of the blockchain should the company go away.
If the blockchain continues to exist and it is economically viable, then the content will remain.
What I don't understand is how users are economically incentivized to store old content when no one is paying for it any more. I haven't read the yellow paper yet, but I assume that new users end up paying for the costs of hosting the legacy content.
It doesn’t store content “forever”. The node operators vote on which files to keep, and if the majority of node owners vote to delete your file, then your file gets deleted.
Basically, it is not permanent in the sense that the bitcoin ledger is permanent. It is permanent as long as nobody cares to send a takedown request.
It’s hard for me to think of a case where this is preferable to AWS.
A takedown request can't easily trigger the majority to delete something, similar to how a hard fork request can't easily trigger the removal of something from the Bitcoin ledger. A globally illicit content can be naturally chosen to be rejected by the majority, but there should be no single coordination point by design.
Somewhat less importantly, if the file is from a block deep enough in the chain (so that a reorg is not possible) and the majority decides to delete it, it only gets deleted by the majority - the minority still keeps it because it remains incentivized by the protocol.
1 - https://thehustle.co/aairpass-american-airlines-250k-lifetim...
2 - https://www.cnbc.com/2017/08/24/technology-killing-off-corpo...