Computers have always been, and will always be, tools for automating away work. The specific nature of the work the user needs automated is in a constant state of flux. One particular kind of work that we've wanted computers to do for us since the 20th century is the sharing of written and visual information.
Ostensibly, this is what the Web was written for, and though it wasn't perfect, it could have evolved into a simply, highly efficient tool for this purpose. Instead, it can be seen evolving into a sandboxed application runtime, a workload with vastly different properties.
Good implementations of software for these two workloads would share little structure outside of that required for inevitable hardware access. Evolving the Web into an application runtime is possible, and well underway, but doing so still leaves the niche of software for an unfettered global commons of knowledge unfilled.
By using the Web we're building for the latter purpose, we are met with drawbacks which include (as I mentioned):
- lack of privacy
- rendering efficiency
- security bugs and malware
- transport performance
- uncacheability
- accessibility
We can attempt to address each of these piecemeal, and with sustained effort we're likely to make good progress on most of them, in due time. But a lot of these are really hard problems for an application runtime to solve.
Conversely, these problems are a lot easier (where they exist at all) on a software platform designed from the ground up for hypertext-based, archivable communication, rather than as an application runtime.
No, we're complaining about those inventing new ways to make people's life much worse in exchance for the off chance that someone leaves them more money.