The problem is that KA (and as you point out, many others) believe that lessons are "delivered". This assumes what Paoulo Freire described as the "banking model" of education and it is fundamentally wrong. People do not learn through delivery and people or organizations perpetuating a banking model are misguided and should be challenged.
Unfortunately all evidence so far seems to indicate that online approaches to learning, on average, do not work very well. To be clear, meta-analyses continue to reveal that schooling, as it exists now, works as well or better than tech centered or online education. I do not find this surprising, if you do not change the philosophy and pedagogy, no repackaging will overcome the serious inherent flaws in the approach. This is particularly pronounced in math Ed and I encourage anyone to dig into the research on what good teaching and learning of math is characterized by.
You wrote "...online approaches to learning, on average, do not work very well..."
Huh? Listen my good man, we aren't talking about the average online approach to learning, we're talking specifically about Khan Academy approach. And I know Khan Academy works for some kids because I've seen some kids using, and liking it -- both kids ahead of their peers and kids behind them.
Whether it works for all kids, on average, is totally irrelevant.
Sadly the majority of Khan videos are not process focused but rather focused on memorizing procedures. This is a very big problem in math education traditional and otherwise. Hard to fix, but KA makes no acknowledgement that incorporating good mathematics pedagogy is even important. This is short sighted and could be remedied.
Part of school is playing the game. But if you reduce education to a game, you are losing something of incredible value. School and education are not always the same thing. And education is not a game. It's purpose is freedom. And it is fundamentally important to the human experience in a way that points and rewards simply aren't.
No. It implies that in general having fewer students makes it easier to be a good teacher. And that being in a smaller class makes it easier to be a successful student. The importance of class size is well-supported by systematic reviews and meta-analyses.
The US does not spend anywhere near tour nations in professional development nor build in the same collaboration time. It does not have as much coaching and other "frippery" . There is plenty of great research on this and while you are right that many pay lip service to improving these areas of US education, little is ever done to meaningfully invest in said improvements.
Also the article did not bring these up as a panacea. These kinds of complicated, cultural, and meaningful improvements will not be made easily, in a short period of time, etc.