Speaking generally, I think you should strongly consider adding a new answer, especially if the existing answers are solely about the state of support for something at a particular time. I think the ability to add new answers to old question is one of the things that makes our site suitable for the long haul. If you cannot write the answer yourself, perhaps someone else can. You can point out the need in a comment and/or put a bounty on the question. If you don't have the rep to place a bounty, there's a good chance a higher rep user will help out.
However, if there's an answer that gives a timeline, or otherwise catalogs how something has developed, it may well be appropriate to edit that answer. This is most obviously the case for answers that contain relatively short lists of information for various releases (e.g., an answer showing where to get
mini.iso for each supported release). But I don't think it is limited to things like that. I think that, at least if an answer is clearly written with the intention of explaining what something was like before and also what it is like now, then you can feel free to go ahead and edit it to make it continue doing that.
Sometimes when I write an answer that I anticipate will be especially subject to change, I make it community wiki to encourage editing, but this is certainly not required and I think there's no reason to avoid editing non-CW answers with updated information.
I'm tempted to say, just follow the usual guidelines for editing: respect the author, don't change the fundamental meaning of the post, and so forth. But of course, when it comes to underlying subject matter that has changed, it's possible to interpret those guidelines to pull hard in the direction of editing... or not editing.
So I think the best course depends on the circumstances. Sometimes it's apparent what courses of action are reasonable. When it isn't, or if you're not sure, you can always post here on meta to get people's views on it. In particular, if you are thinking of a particular such question right now and that's why you've asked this question, then you might want to expand your meta question by editing to explain the situation.
This topic has been touched on a couple times before; perhaps some of the ideas people had then will be relevant.
How do I get involved with testing and QA with Ubuntu? motivated the meta question What is the most appropriate way to bring an old/outdated question up to date?, which you may want to take a look at. It's similar to this meta question, except that this one seems like it's specifically about advancing technologies (such as device support), while that question is asked in a way that makes it rather more general.
That QA question didn't need any radical action taken to keep it useful. The accepted answer remained useful, the meta question's OP posted an answer, and an additional valuable answer was later posted, filling some of the gaps in the other answers. That later answer received a bounty.
The meta question What should we do with an invalid accepted answer? was about a more extreme situation. The recommended way to install Oracle's proprietary JDK changed, relatively rapidly, and, as I understand it, there was concern that the old advice about obtaining it through the partner repository would be useless and distracting or, at worst, actively harmful. (It's important to keep up to date with security fixes for Java.)
The eventual resolution was to delete both questions mentioned in that meta question. (They had been closed as duplicates of another question that is itself now a duplicate of How can I install Sun/Oracle's proprietary Java JDK 6/7/8 or JRE?, with all answers merged into that one "canonical" question. Many highly upvoted answers were deleted along with those two questions. We don't usually do this, and I'm unsure if it was even the right action for that situation (though I'm not arguing we should go undelete them now, unless some new reason to think the content is useful presents itself).
I bring this up to make clear that deletion--even of highly upvoted questions with multiple highly upvoted answers that have actually helped many people--is something we can do in the more extreme cases of old information having a combination of very low future usefulness and high potential to cause confusion or lead to harm.
This is something of a last resort, though (and, when necessary, is still best done with community involvement). Posting new answers, editing existing answers, downvoting unhelpful posts and upvoting helpful ones, placing bounties to get new answers or highlight existing answers as valuable, and so forth are usually sufficient to solve the problem of outdated information.