I reject the notion that Amazon's hiring practices are neither implicitly nor explicitly comparative. The notion of hiring candidates that "raise the bar" is explicitly comparative to existing employees. That said, I do not know the details of Amazon's hiring philosophy. I do know that internal promotion within Amazon is quota constrained and therefore implicitly comparative.
You are absolutely right on that. The core question that needs to be answered at each interview debrief is "when the next review process roles around, do we believe this candidate will end up in the upper half of the stack rank for their role/level?" It's a little more complicated than that as Amazon's stack rank is a two-dimensional rank that looks both at contribution-in-level and perceived growth potential, but in the end it's still a judgment call of "is this candidate better than half the people currently in this role."
For Amazon, this is not a correct conclusion. Having done well into the low-hundreds of on-site interviews for Amazon, I do not recall a single instance of a debrief discussion comparing one candidate against another. At Amazon it is simply a matter of making the bar or not making the bar. While it may seem like you are interviewing for a specific position, in reality at Amazon you are interviewing for a given role and level. Generally the level is targeted based on the candidate's experience and if the candidate barely doesn't make the bar but the interview loop agrees they have growth potential they may get an offer at one level lower, though this is not particularly common. Amazon will extend an offer to any candidate who meets the bar regardless of how many good candidates are interviewing for a specific "opening". Between the company's growth rate and the high attrition Amazon needs more new hires than can be found that meet the bar.