I am also in the course(invited). One of the big ones to me personally, as I have been following fastai for awhile now, is that this is the first time the library is going to go into “stable” releases. I don’t know if companies will adopt it, as I think it is more of a “hacking/research/kaggle” library, but I think a lot of the general high level training regimes are decided at this point. The new library is also very flexible, so you can override it or use callbacks fairly easily to add any functionality you want without overwriting nearly as much code, which was a major issue with previous versions of fastai.
I think overall you can learn with the 2019 version, and quickly go through the 2020 version when it is released to get caught up. No reason to delay your deep learning journey because the course isn’t released yet.
I of course still do recommend the paid course. Having such an engaged community when you are learning deep learning is really invaluable to how quickly you pick up material. The social connections are also very valuable, to your development. Getting to know people who are also interested in deep learning, who are reading about research and sharing their experiences is how you become part of the “community” and keep up with the every advancing field of deep learning. You can still do this without paying, but I am always surprised at how motivated some of the members of the fastai community are in continuing their own development.