nostr:nprofile1qqsvnvvlln2ru6jlyweayugxecv7ftfdlzd6zqca63sh7x6ercggjegpp3mhxue69uhkyunz9e5k7qg4waehxw309ajkgetw9ehx7um5wghxcctwvszvprka nostr:nprofile1qqs0m40g76hqmwqhhc9hrk3qfxxpsp5k3k9xgk24nsjf7v305u6xffcppemhxue69uhkummn9ekx7mp0qythwumn8ghj7un9d3shjtnswf5k6ctv9ehx2ap0d2j30f nostr:nprofile1qqs8fl79rnpsz5x00xmvkvtd8g2u7ve2k2dr3lkfadyy4v24r4k3s4spz9mhxue69uhkummnw3ezuamfdejj7qg7waehxw309ahx7um5wgkhqatz9emk2mrvdaexgetj9ehx2ap0ege3s8 nostr:nprofile1qqszrqlfgavys8g0zf8mmy79dn92ghn723wwawx49py0nqjn7jtmjagppemhxue69uhkummn9ekx7mp0qy08wumn8ghj7mn0wd68yttsw43zuam9d3kx7unyv4ezumn9wshsud80st nostr:nprofile1qqs8k0mcqd6sw3h524qn5gslszt9am9knmes3uh268dgnnpv3yfwj6qpr9mhxue69uhhqun9d45h2mfwwpexjmtpdshxuet59uq3camnwvaz7tmjv4kxz7fwdp5kw6rvd9nksar9wghxxmmd9u5jtddh and others.
What about solving op_return with nested PoW? In this case all tx would be mini blocks, with difficulty scaled down from the network difficulty and then multiplied by the bytes in op_return. It should effectively scale the cost of larger arbitrary data in op_return by making miners have to choose if the fee is worth the PoW for the tx and scale the cost of larger blobs.