I know that this has been brought up in a previous thread, but it was definitely brushed off too quickly. After doing some digging, I've noticed a significant difference in load times between runners, and it simply isn't fair for somebody to have such a distinct advantage just because they could afford to build a faster PC.
For example, according to the leaderboard the WR is 28:49 and my time is 31:56. However, when I retimed these videos myself to discount load screens, it turned out that Armork's run had 152 seconds of total load time throughout the run, averaging about 10 seconds per load screen, and my run had 308 seconds of total load time, averaging more than 20 seconds per load screen. That's a huge difference that has absolutely nothing to do with gameplay. It still puts him ahead, but only by 31 seconds as opposed to several minutes.
In the sake of fairness, I'm happy to take it upon myself to retime every run on the leaderboard. Timing stops when the white load markers appear on screen and continues when the player regains control of the character. It's objective, and it's fair.
Going forward, there are ways to make it so that runs submitted would need to be timed according to this standard. The Crash Bandicoot NST community, for example, has a program built into LiveSplit that stops the timer automatically whenever a loading screen is detected. This tech exists.
I cut the time after confirming throwing away the cards, however the seconds can easily be counted if it's decided that the timing extends through the dialogue. There was a brief delay in the timer stopping (my phone had logged out) but if you pause the video when I say "time" it's 6:25:22.
Thanks, guys! This run was exhausting, but I'm glad I did it.