Trivially untrue. Inability to learn except by doing is a defect, which renders you extremely vulnerable to critical errors - you won't be able to learn about these errors since making them (which you find you need to do in order to learn) is fatal.
That doesn't make learning by doing a bad idea, or even necessarily a poor first choice, but if it's the only way for you to learn that's a problem.
I don't think they were suggesting that at all. It is not an on/off feature flag, people learn in different ways. People learn in different measures of many ways, even in the one person.
School typically only caters to one type of learning, and it actually wouldn't matter which type since only focusing on one always leaves out the other.
Lastly, if you can only learn one way, and it is a defect, what do you expect them to do? Genetically modify themselves? Chemically correct themselves? They're kids, they need to be catered for, they can't do it for themselves.
So your claim is that one can learn how to do something simply by being told, and not doing any practice problems, thought exercises, reviewing solutions, etc?
No, your claim is even stronger - that anyone who doesn't learn that way has a learning disability?
I think either version is far too strong a statement.
Regardless I think it's trivially true that one learns by doing primarily or possibly exclusively. When I think of all the practice problems or "think through implications" that I have to do before being competent enough to claim i know something... Let alone my first attempt at applying the knowledge. That's all "doing".
Saying that a whole style of learning for neurodivergent folks like Temple Grandin is kinda misguided. They don't need to learn by doing by everything, and I'm sure you knew that.
It's even more nuanced than that, there's the type of task, the person's learning style, the makeup of the specific task, the manner of feedback from performing it, whether there's flexibility in performing it quickly or slowly. Only chronic take-havers and bloggers will reduce it to one or two variables.
Based on my several decades of life experience, I would say the point is there’s hardly anything useful that you can learn without doing. And there are many things that you can’t do without leaving the traditional classroom environment. That’s why schools have chem labs, wood shops, kitchens, orchestras…oh wait, do they have those anymore?
Another point is that you can’t really learn a skill unless there are stakes - a real goal you need to accomplish, real customers, real coworkers. Grades aren’t real stakes; at least I didn’t regard them as such.
I’ve seen this over and over through the years as new college grads arrive who know a lot about things but have no idea how to do those things. Unless they went to a school with a good co-op program.
That doesn't make learning by doing a bad idea, or even necessarily a poor first choice, but if it's the only way for you to learn that's a problem.