I'm not so sure Alysa Liu's jump technique was fixed, as much as it was marginally improved. And I wonder if part of her retirement is a realization that it isn't going to last and/or will take a significant effort to maintain.
Liu was still quite slow in and out of some of her jumps, especially the triple axel, at Worlds. If she tried any triple axels in warmup at Worlds, I completely missed them because she was skating so slowly into her jumps. Even in the program, I had no idea she was attempting a triple by the takeoff. Like many others, she also is relying on edge pulls between jumps in triple-triple combinations and the hopped v. jumped eulers in order to get the combinations around. Her underrotation issues seemed to be getting worse, not better, until a sudden (but not complete) reversal at Olympics and Worlds. I'm not sure if that's technique improvement or just training. I tend to think it's the latter. This is taking nothing away from Liu, who skated her absolute best when it counted. But I do think her technique and her overall speed/ice coverage would have continued to be issues.
As I posted in another thread, Levito has cleaned up some of her technique - her toe loop is looking excellent and drawing in big GOE. Her double axel has been very nice for awhile. These jumps are big and floaty, unlike the likes of Jenny Kirk or Tara Lipinski, who always whipped their jumps. I'm cautiously optimistic.
For both Levito and Thorngren, I think (J)GP results next season are irrelevant. What's relevant is whether they can work triple axels and/or quads into their program content. This is the point in the cycle to take risks.