There's countless things that people tried for decades and nobody succeeded in that are now ubiquitous.
It's easy to justify being dismissive of people trying things that have yet to succeed by this argument. But it's also, clearly not valid.
For instance, with the rise of cloud storage and web-applications, we are doing, essentially, thin client computing now - something that people said would never come.
Self-driving cars were another constant future-device (e.g., http://ThisWillTake.Me/youtube/Johnny_cab_clips_from_total_r...). And now it looks like we'll get them by the end of the decade. That's pretty impressive.
From a software standpoint, there are now 2 categories:
1) Software as a service 2) Software as an end product
"1)" is more honest, because your customers only care about the service and you have nothing to gain by churning and regurgitating new-old-new-again ideas every 10 years in order to keep scoring new license deals.
"2)", well, I said it already: As a software vendor, you have no financial interest in selling anything that won't need to be replaced at least once a year. Hence, the constant renaming of recurring concepts over and over and over...in order to keep license revenues coming in.
Though I use many SaaS offerings, I still prefer installable software as much as possible because once I own it, I own it forever. For example, Windows XP may reach EOL next year, but existing, activated copies aren't magically going to go away at that time.