This is ignoring the massive distinction between manned flight (where failure is not an option) and unmanned tests. NASA and SpaceX both know this well.

Calling it a "SpaceX like approach" and connecting to Apollo 1 is a neat trick, but SpaceX wouldn't (and doesn't) adopt that risky approach during manned flights.

It's all about "the right risk for the job." You can't be risky with human safety, but you also don't want to be overly timid and failure-averse during safely managed R&D tests, or your R&D grinds to a halt.