haha... i have a project coming up for these guys

it's still only in testnet and i have yet to hear or read a credible model for actually immutable permanent storage without an absurd price due to the clear actuarial risk of it

i think they need to bulid a system that opportunistically cashes and replicates frequently requested data and gradually trims back the replication rate as data gets less attention

the very exact same problem exists with Internet Computer Protocol which is an earlier, and similar sort of blockchain project, and my current project is using IC as an event store (has a basic filter search functionality bulit into it)

i dunno what is in the mind of the guys who build these things exactly, i've done zoom meetings with several of them and honestly they seem spaced out to me

and i get into frequent debates with my colleague over issues of architecture and engineering even though i'm 8 years into actually working in this field and he's fresh out of a B.Sc in CS

don't count on that data sticking around forever

i think a dynamic, automatic archival storage economizing protocol is coming in the future but arweave is not it, not yet 

 Yes I agree.  I think Arweave's true weakness is that the gateways are socialist.  So to store the data you get paid, but there's no motivation to deliver it like you said.  So the bulk of the gateways are cloudflare