I think you're generally right about the OP, but I don't see your alternative method as a tonic to the problem the author is discussing. I agree that treating reading as the only way to ingest information is probably wrong, and that the original post tends towards hierarchy in an almost reactionary way, yet at the same time I also see that many people tend to ingest information from other forms of media less critically than they might a book or written piece. I think this dichotomy derives from the different levels of focus we apply to different activities. Reading, even for leisure, demands a level of effort that forces you to focus and, in a way, analyze the information you're ingesting. By removing that barrier, you might fail to fully grasp the ideas that the primary source expressed.

In your use case, the LLM can be a great tool to get a start, but it's much easier to accept that output uncritically, which is especially problematic because it bears the biases of the training corpus. Sure, you can mitigate that issue if you're driven by the intellectual curiosity to dig deeper, which it seems you are, but at that point aren't you doing the kind of leisure research that the original post is talking about anyway? You're just using the initial steps to discover the "foundational texts", as the original post puts it, of the field. The LLM step becomes less material to the follow-up reading you've done after the fact.

Many, perhaps even most people will not make it to the second step. I see the lack of intellectual curiosity as a problem of focus more than anything else, and in my experience the overly rich media landscape of the internet does not concentrate that focus in most people.