The hard part of making new breeds is getting one that is resistant and would probably be accepted at market while not having an overwhelming number of seeds in the fruit (those last two bits overlap a bit).
The problem is finding the cultivars and doing the whole experimental process with specific plants and documenting it... Given that bananas grow fairly quickly it is probably relatively do-able. Just need some land and some volunteers... And flying under the radar of the banana companies :)
Sure you can. Only 15% of bananas grown for food are the sterile cavendish type. There are many many other types.
Few things on earth are as weak and helpless as a cloned monocrop.