If it's okay, I think we should set the non-vertabrates aside for the discussion since their optical systems are so different. As to the safety of UV-A, I don'

If it's okay, I think we should set the non-vertabrates aside for the discussion since their optical systems are so different. As to the safety of UV-A, I don't think you're correct. A common element of all the comparative species selected is that their life span in the wild is far shorter than our own, and there isn't going to be selection pressure happening before they would accumulate UV damage. UV damage to the eye is kind of like yellowing of plastics, it accumulates over time. As is pointed out in your first linked article, aphakia patients are advised to wear blue-blocking glasses to protect their vision. My first thought as to how bad could it be was retinoblastoma—I am not aware of any studies looking at the incidence rate of Rb in aphakia patients and PubMed doesn't show anything either, but I don't think it's an unreasonable thought. Aside from the ionization problem, UV light also has a much higher propensity towards scatter. Floaters are bad enough for some in visible light! Of course, we're talking futuristic gene augmentation, so it's not unreasonable to suggest we'd vacuum the eyes out every now and then, so maybe floaters will not be a problem :) We still have a problem with the much higher dispersion of UV light relative to visible, given the materials we have to work with in the eye; that, plus scatter, means we're probably going to have very fuzzy UV vision. Replace the whole lens system and maybe you get better UV-vis vision, but we still haven't overcome the safety problem. Maybe instead of vacuuming out the eyes, we just swap them out for a new autologous gene-tweaked set every couple years. NIR would seem to be more realistic than UV (certainly considering safety), but here, the quantum efficiency of the pigments would seem to be a limiting factor. As you point out, we can already actually see a little into NIR (IIRC, in scotopic with something like 10-6 QE at 750nm relative to 555nm). At first glance, the chemical structure of human receptor pigments doesn't look to be super favorable for modification to getting something that would have a good absorption cross-section in NIR. Compare the structure of vitamin A or retinal with a strongly NIR-active chromophore, there's a lot of steps bridging the two and the structure as a whole is hostile towards selective modifications, not to speak of the rings where the modifications need to happen. As to thermal IR, I think we'd need a metamaterial for something that would work in vis and in LWIR, and that's a handwavey statement on my part. It would definitely be of interest to the military! Overall, I think it's more plausible to augment our vision by wiring a camera directly into our visual cortex. Biology is too hard.