Of the five salient points they make here, I would argue that they are a lot less strong than the authors believe them to be.
Point 1 can also likely be made for the original Sars-CoV, and probably is irrelevant. If they show CoV-2 is more human like than the earlier, then they can begin to make the point.
Point 2 is a gigantic leap of faith - I don’t understand how they connect amino acid inserts into the sequence to gain of function experiments. It makes no sense!
Points 3 & 4 are related: The extra charge on the RBD. If you look at the previous virus, it has a much smaller positively charged patch, and this expands in CoV-2, which makes an evolutionary drift in this direction entirely reasonable. It’s not a wholesale integration of a brand new feature. This charged patch is actually important for binding to heparan sulfate, which gets involved with a bunch of viral entry events (eg AAV, Chinkunguya virus, some other coronaviruses I think). There’s a few recent preprints that discuss the HS relationship with Cov-2, and can go a long way to explain the tropism of the virus. The HS can overcome lower expression of ACE2, by possibly supporting the attachment.
Finally, point 5 is where they throw out DC-SIGN as a receptor. While there are high mannose glycans on the virus particle, I don’t know of anyone proposing this as a receptor for the virus. There is no biochemical/experimental evidence that they show to support this idea.