Fun fact: the core technologies for a decentralized, interoperable, format-agnostic, multi-device, interactive, programmable web were relatively mature by the beginning of the 21st century, when they were used for a brief and ultimately unsuccessful engineering experiment called the "World Wide Web." I am being cynical and sarcastic, but it's also accurate to say that most of the baroque technical layers being pushed to somehow build a better Internet on top of the existing one are necessary only because so much of the existing Internet breaks already existing capabilities. We had, and have, full-featured standards for complex distributed hypertext, multi-format content negotiation, and a host of other "innovations" that are hard to do now because they have to be built on top of browsers, libraries, infrastructures, and even devices designed with an instinctive, almost unconscious horror about what a true Hypertext looks and feels like (which I would argue is a version of the somewhat more conscious horror of, and war against, general-purpose computing).
The canonical link to what we didn't build: As We May Think. And, note, this isn't a description of a product; yes, you can buy or built Evernote or whatever. But the Web was built for, and still has very much every technical piece necessary to be, a distributed knowledge engine under the total control of every edge node. The simple fact that you can't sort whatever "feed" of content you're looking at — that you have an incredibly powerful device connected to an unfathomable planetary network and you can't even do that &madsh; is a loss so deep that I'm tempted to think we, as a civilization, looked at computers, while we were still dazzled we looked at the Internet... and we recoiled at what it meant, a collective intellectual vertigo that was to the benefit of, and pushed at by, some of the largest companies in the world, but that wouldn't have been possible if we hadn't been complicit.
None