I downvoted this comment (as well as your comment below) for strongly pushing misinformation. As others have noted, the CRISPR/Cas9 system has evolved in bacteria precisely to target viral genomes — “CRISPR is not able to target viruses at all” is simply false. ”...and also does not destroy the things it targets” is also false, in a sense; a well-targeted Cas9-induced double-stranded break in the DNA/RNA of a viral genome can certainly disable a crucial viral gene and reduce viral replication, even if you don’t consider this “destruction” of the genome.
That’s not to say that the CRISPR/Cas9 system is quite ready for antiviral therapy in vivo. One problem is that you could rapidly generate viral escape mutants. Not only do you create selection pressure for the virus to mutate such that your bespoke CRISPR/Cas9 system can’t target it anymore, the Cas9 cutting itself guides this process along more rapidly, since double-stranded breaks are often accompanied by random insertions and deletions at the cut site (incorporated during attempted cellular repair of the break). This could potentially be addressed by targeting important, conserved regions of the viral genome and/or by multiplexed editing (i.e., targeting multiple sites simultaneously).
Perhaps a bigger challenge is delivery. Systemic delivery (i.e., throughout the body) is risky, since you can get off-target edits in cells that aren’t even infected with virus, which could result in increased risk of cancer or other maladies. Targeted delivery to only a certain class of cells of interest is sometimes possible but difficult. There’s also the perennial question of whether or not your looks-good-on-paper molecular mechanism of action translates to real clinical benefits, something that can only really be definitively answered in clinical trials. For example, you might see efficient viral genome cutting in vitro but see no clinical benefit in a patient, because maybe your Cas9 protein doesn’t stick around long enough in cells, or maybe you can’t get it into enough cells to matter, or it’s detrimentally immunogenic, or a host of other hard-to-evaluate-in-advance reasons.
All that being said, this is a real direction of interest and many are looking into it — the OP is not “completely on the wrong track” and this idea is not “nowhere close to the sort of thing that would work”. The fact that CRISPR/Cas9 is so programmable makes its potential use as an antiviral therapy exciting and at least worth exploring more, in my view. Here’s a nice review if anyone would like to learn more (warning: paywalled): https://www.cell.com/trends/microbiology/fulltext/S0966-842X(17)30093-8
Hmm, can you think of a plausible biological mechanism by which a virus could evolve to not cause fever, or to cause fever later than usual? My initial reaction is to be skeptical that fever screening would result in the effects you suggest, mainly because whether or not you get a fever is mostly a function of your innate immune system kicking in and not a function of the virus. Whether or not you get a fever is mostly out of the virus’s control, so to speak. The virus could perhaps evolve methods of evading innate immunity, but other examples I’ve seen of viral adaptation to innate immunity seem like they involve complex mechanisms, which I would guess would not evolve on timescales as short as we’re concerned with here (although here I’d welcome correction from someone with more experience in these matters).
But even if there were potential evolutionary solutions close at hand for a virus to evolve evasion to host innate immune responses, I’m not sure that fever screening would really accelerate the discovery of those solutions, given that the virus is already under such extreme selection pressure to evade host immunity. After all, the virus has to face host immune systems in literally every host, whereas fever screening only applies to a tiny fraction of them.