The context of the video is to demonstrate to technical speedcubers that he's achieved the feat he claims. By showing his solves on video, speedcubers can tell that he's using the algorithms he claims and not more common techniques to obfuscate the achievement. The length of the video is to provide even further evidence of the claim.
Edit: The achievement is that he's been able to minimize the number of turns he needs to complete the cube by memorizing ~4000 edge cases and the specific turns needed to solve the cube from those configurations, as opposed to generalized algorithms that require memorizing less edge cases at the expense of more turns.
This has been a known possibility since 2011, but this is the first documentation of someone demonstrating mastery of it.
Hmm I see, thank you. How do we know that he hasn't just learned the specific cube he's solving in the video, though? Is there some verifiable source of randomness?
I'm not a member of the speedcubing community (only adjacent to it through acquaintances), but as far as I know, the website he's using is standard in the community and those in the know who have viewed the video acknowledge it as legitimate.
Edit: The achievement is that he's been able to minimize the number of turns he needs to complete the cube by memorizing ~4000 edge cases and the specific turns needed to solve the cube from those configurations, as opposed to generalized algorithms that require memorizing less edge cases at the expense of more turns.
This has been a known possibility since 2011, but this is the first documentation of someone demonstrating mastery of it.