In CS, a lot of the knowledge that is very valuable today was discovered years ago and is taught well in Master's programs. As a concrete example, I know some great database developers making >1 million a year because of their expertise. They tell me the vast majority of knowledge the apply at work was discovered in the 80s. They're up to date on newer stuff, but it's just not what gets used.
Could that all be invalidated by some new hype technology? Possibly, but even then it would be decades before these older systems are ripped out.
I think you're imagining a much more rapid change in technology than usually happens. Heck, I knew someone who's masters was focused on vacuum tubes and graduated the year the semi-conductor was invented. This was one of the more rapid seismic technical shifts in history and it still allowed him to have a long lucrative career with legacy systems. I wouldn't worry about CS knowledge being outdated anytime soon.
Academia will always try to teach the best way around a problem. So, you won't see academia, in an updated program, trying to teach these old ways, eventhough there's still going to be a market for legacy technology. They will usually be highly profitable for that same reason.