While you are undoubtedly correct to some extent, there's also the entire academic side of decentralised systems which is more about solving the our (plural) sy

While you are undoubtedly correct to some extent, there's also the entire academic side of decentralised systems which is more about solving the our (plural) system vs their (singular) system problem. The same tech that makes these complex decentralised services work also often functions just as well if not better (performance wise) in federated/centralised contexts. These decentralised systems are a natural abstraction for so many different types of problems and systems. If you leverage them you can still get high performance systems in a centralised manner but the hard work making those systems resilient, localised, and highly available has already been solved. If you design your systems on top of this technical layer you can leverage the benefits without the costs of centralisation for the most part. The other academic argument to be made is that this is basically the only industry that is meaningfully pushing cryptography forwards (IMHO). Concepts like Functional Encryption, Fully Homomorphic Encryption, Secure Multi-Party Computation, and Zero Knowledge Proof Systems have been primarily funded by cryptocurrency research and towards the application in decentralised zero trust systems (which are basically the "ideal" environment for exploring cryptography and game theoretic technologies). Code doesn't fix people problems but it can augment existing systems and letting those systems evolve to their logical conclusions is a worthwhile endeavour.