Re 2: on this software engineering forum, the following example will help.

If you have core dependency goes rogue, and you have to switch to an alternate library with similar features, is that a free switch? Think of how many thousands of hours of work are often needed? How many businesses have gone under because of such issues?

Growing a particular variety requires a lot of knowledge gained by each individual farmer from experience. You can't just go back to an old variety for free. It may take several years for yields to go back to previous levels and by then the farmer may have gone under.

Farmers change seeds all the time. One I know tells me that a great variety will terrible in 3 more years, though I'm not clear why. In any case they all are planting several varities ever year - four different ones in a field isn't uncommon - with harvest data to track the difference (different soils need different seeds). Test plots where they do many different side by side are somewhat common. they are always trying different options to see what works to do more. Plus predictions on weather mean different seeds.

Ok but going back to the library analogy, GMO bans are like the government banning react.js because they're convinced angularjs (or jquery) is good enough and facebook might go rogue. Shouldn't it be up to individual farmers to decide?

The difference I think, is that the libraries are open source, and you don't have to pay Facebook yearly to use React.

Countries can and do ban closed source paid products when they don't trust the provider (e.g. Huawei)

When you buy hardware that goes rogue you're forced to throw it away and install something new, which costs money for hardware and labor.

When a GMO goes rogue you can just buy a different seed the next planting season, which you were going to do anyway.