"You'd almost be able to pull it off if you went into overdrive on the compression."
To a first approximation, a Mbps is a Mbps is a Mbps when it comes to image quality. Just because you push the 4K button on your compressor doesn't make it a high quality image if you don't spot it enough bitrate to actually paint distinct pixels. You'd be far better off measuring image quality in Mbps than resolution. (Except in pathological cases where the video is overshrunk, but I'll let you know when I encounter the video stream that is just too jam packed with quality. Hasn't happened yet.)
My camera claims to take 1080P video, but it's a dirty rotten liar, in as much as I'm able to take the video coming out, cut the resolution in both directions by 2x and cut the bit rate per pixel when I reencode it, and produce a video that is indistinguishable from the original when zoomed full screen. (If you are thinking to yourself "why, that means the original video must not so much have pixels as blotches of color when looked at at the pixel level"... correct. It's what gave me this idea in the first place, and it turns 1.5GB videos that are even today somewhat annoying to store into 200MB videos that are much easier.) And I only stopped at 15% of the original size because I don't care about the last few percent and it's better to still be a bit fluffy than start cutting into real quality, but I suspect I could get it to fully 10% of the original size with only minimal quality loss.