I think this is because at the end of the day, normal people dont care about decentralization or protocols.
Look at email and http. I can use any client I want to access email/web. Outlook, Thunderbird, Chrome, Firefox. We already had this sort of utopia. Anybody could spin up an smtp server or web server. But that involved a little too much friction. So Google and Microsoft offered ready to use email. And Facebook and Wordpress offered up ready to use web sites. Google and Microsoft could probably wall off their garden and people would just jump in, instead of run. Facebook went from jabber to proprietary. RSS has died and been replaced by the Twitter follow. Find me one non geek that cares about Mastodon/GNUSocial/OStatus/ActivityPub at a philosophical level.
When people talk about blockchains they either too narrowly focus on currency application or substitute it for the word database. Quite often on HN you see people comment "what does this really get you that a normal database doesnt."
Blockchains should be used when the goal is to distribute trust, allow non-oligopoly institutional verified participation, and minimize fraud. But we live in a world where _most people are perfectly comfortable just trusting the easiest to use Institution._
One application of blockchain potential that repeatedly comes to my mind is cab hailing. Instead of the Uber/Lyft duopoly, ideally there would be a "hail a cab" protocol, just like smtp/http. Something that works between any client that chooses to be compatible. Many many competing institutions could run their copy of the ledger: a ledger that holds driver, passenger, and client app trust ratings, route information, load trends, predictive driver placement, event schedules, all while facilitates payment. Any client could choose to ignore any part of the distributed database it wants, like Mastodon. Why is this better than the current situation? Because I could use Lyft to call an Uber driver. Cross app compatibility, standard hailing protocols, COMPATIBLE TRUST. Innovation in the client side features would skyrocket. When the protocol updates, everyone benefits. And because ride hailing doesnt depend on a social graph, unlike communication tools, lock in to one particular institution is much much lower. I hail a stranger, not a friend. I can swap my driving dispatcher app or cab browser at any time, just like I am free to flow freely between Chrome and Firefox. And unlike how I cant browse the facebook/uber database with anything besides facebook/uber made apps.
But no one cares. The majority of the world is ok just trusting and paying 2-5 institutions to handle trust and money transfer and data security in any one particular space (banking, communication, ride hailing.) We have already seen email and sms give way to fbmessenger, whatsapp, snapchat. IRC gave way to Slack. We have gone from a world where clients are inter-operable, to silod. We seem to want this. And until the benefit of cross compatibility and distributedness reemerge and permeate into general culture, Non-cryptocurrency uses of blockchain will be few and far between. Because the big institutions can offer something thats already "good enough." When something is good enough, people dont go looking for better. Blockchain and mobile app developers need to give the general public a UNIQUE benefit, and a reason to care, switch, and stick with open systems.
(Another interesting application of blockchains could be to the yelp/foursquare/tripadvisor/opentable/dining-rewards/loyality space. One protocol to handle menus, reviews, and reward points, instead of a separate incompatible reward app and shitty non mobile website for each damn restaurant. Another, and different component of trust in this ecosystem would be verification of authenticity vs spam.)
(Another application of blockchains, shared accounting systems between cooperating firms. General contractors and sub contractors all working off one distributed database for all ERP: general ledger, finance, forecasting, budgeting, cost management, change management, orders, invoices, contracts, ach, settlement, project schedules, purchasing, inventory, time and attendance, manpower, deliverable, shipping, materials, manufacturing, drawings, BIM (revit) etc, union tracking, payroll, training, safety tracking, SHARED analytics! In this scenario, the threshold for trust can be lowered ala https://azure.microsoft.com/en-us/blog/announcing-microsoft-... As of right now, if the general uses primavera or plangrid, you use primavera or plagrid. The subcontractors end up using a litany of incompatible tools across isolated projects. Each ERP and accounting system uses dual entry and every firms duplicate each others efforts. Massive redundancy, coupled with very foggy visibility into whats actually going on. Interoperable ERP with shared triple entry accounting would be a worldwide gamechanger. But construction isn't sexy enough for most of Silicon Valley fad chasing, and the scale of Enterprise ERP is too massive for anyone but the big dogs.
I dont see SAP, Oracle, and Microsoft opening up their products to each other anytime soon.)