To answer the thread title, it's nowhere near fully reversible. We're maybe 1% of the way there, and medical technology is not progressing exponentially.
Of course, if you apply strong enough AI (and note my words there: strong AI, but not necessarily general AI), 1% could be half-way there. A lot of this is a puzzle. In some cases, we don't have all the pieces, so we'd need AI to help find those pieces in the first place (and thus we'd need AI to improve our tools for us since our tools are likely not sufficient enough, like looking for puzzle pieces in the dark in a 20,000 square foot mansion equipped only with a dying flashlight). But if we have the pieces or at least enough of them, we'd just need the AI to piece them together— no pun intended. Again, it'd be like putting together that same puzzle in that same mansion in the dark (in the earlier stages, it's indistinguishable from finding the pieces). But the AI would be as if you have a smarthouse that doesn't need ambient light and uses little drones and robots to put it all together, and it has a full-fledged image of what it's supposed to look like at that, if without the puzzle pieces themselves.