They also show how the problem is hidden on my lower-gamut MacBook Air, and revealed only on the UP2414Q.
Yup, that would do it. Instances where a user would see no difference in the test images under Firefox or Chrome:
1. You're on an iPad or iPhone. Both of these devices default to sRGB for everything.
2. You're using sRGB as your monitor profile. (On a wide gamut monitor, they'll match, but both will be over saturated, as will everything else.)
3. You're using an older or worn out monitor (most likely not a wide gamut display to start with). You may see some difference, but not necessarily.
As near as I can tell, all web sites are affected
Yes, that would be the case for Firefox and Chrome since no matter what web site you're on, all untagged images and all other color elements will display according to your monitor's gamut and color range (uncontrolled), rather than as the author may have intended. Another way to say it, what the person who built the page saw on their screen has little chance of being seen the same way by others.
This is the case anyway since monitors and the color they are able to reproduce are all over the map. Millions of monitors are either; old, worn out, not properly calibrated and profiled, have never been properly calibrated and profiled at all, are set to use canned monitor profiles which don't represent the device in front of you in any way (such as using Adobe RGB, ColorMatch RGB or some other provided profile as a monitor profile).
That is at least Safari's biggest advantage. It applies some sort of control to everything. It's not perfect either to apply sRGB by default to all uncontrolled and untagged elements, but it's still better than no control at all. Especially images since any competent web designer will save web images in the sRGB color space, whether they have a profile embedded or not. So an untagged image is still being displayed as intended when Safari applies sRGB to it by default.