The more I explore the space, the more it's becoming clear to me that the origional vision of crypto probably isn't going to happen. But that the potential is even greater than the origional vision. Any time I've tried to design systems that tie the crypto world to a physical object that's not easy to verify in a distributed way (so verifying an item is delivered as opposed to what the weather is) it's clear the system is fragile... and the user experience will always be worse than a trusted system.
But digitally native stuff works great (on alternative chains, I haven't had good experiences on ETH... kind of expensive, kind of slow).
The blockchain isn't about digitalizing the real world economy. I think it's about creating a new purely digital economy. Trading digital assets for digital money. Things that can be easily digitalized can fit in this model well. SaSS, Finance etc. things like eCommerce can be partially put on-chain, but the experience is better in a hybrid way. In my vision of the future, a robot that delivers packages might be controlled via an NFT, but a package is not going to be tracked via an NFT.
Also to explore why it might be useful to do some stuff on-chain, it's kind of a chicken and an egg problem. It's better to do it on-chain, because something else on-chain needs it. For example, on-chain cloud hosting is not necessary superior in our current world (but it does exist) but if you're trying to do something else on-chain (like say run an autonomous organization that manages a web service) running the web hosting on-chain is just going to be eaiser to do business.