Well, placebo effect is a powerful thing, but let's start with my original take before I'd watched the video.
I'd certainly heard the claims before, and I've been around enough that I've also seen the disc polishing and 'disc fixer' stuff over the years, and had some idea of the theory behind the claims.
From a purely technical standpoint, the Red Book standard (the official spec for classical audio CDs, later updated to the Scarlet Book to encompass SACD) talks a lot about its layers of error correction; you have the EFM (eight to fourteen modulation for how it physically gets encoded on the surface platter), and then you have oversampling so that's absolutely a thing, and if you're not running into the oversampling you will get the 'original' rather than a slightly reconstituted version.
So I can buy the theory that if your disc is cleaner, you don't have to listen to the error correction or rely on the oversampling.
And I can buy from a scientific perspective that the light refracts through the clear outer coating and that the black cover would cut that out - again, the physical properties of how Red Book audio defines what a CD 'should do' makes sense to me.
And yes, I can buy the notion that filing the edge off for balance prevents physical vibrations that influence the electronic noise that the DAC is handling down the line.
Now, I will give you that his test was smart enough to eliminate the likes of the DAC from it. What I will also give you is that I think his equipment is superior to what people were using back at the time. This makes a difference.
CD players of the era only really did 1x playback - anything else was a fudge - which means you're getting into reading the data stream at the correct speed and accepting any faults as they are because you have no time to read ahead and buffer. Newer devices had the time and capacity to read ahead and attempt to hit all the data and error correction because you can do it more than once in the time you have, meaning that you have a much better chance of hitting the data as it was originally encoded, and not reliant on synthesising it through the error correction.
Such measures - if they work - would have been for the gear of the time, trying to fix the media to make playback better then. The reason the company no longer sells them is partly because buying CDs is hardly the market it used to be, the mixing/mastering of CDs now makes such error correction utterly unnecessary and by extension such fixes also unnecessary.
I would, by way of other studies, refer you to a review where someone tries to discern discs from the individual measures rather than the holistic view:
https://www.enjoythemusic.com/magazine/equipment/0603/audiodesk.htm - this is more contemporary to the device when players were still more susceptible for it.
As to his comments about it applying to DVD that actually does make more sense than it does for CD. DVD data (especially prior to Blu-Ray) was encoded on a disc in a compressed fashion, so any loss will *absolutely* be taken care of by error correction logic because the entire MPEG standard is about reconstituting missing data from inference (and it doesn't use RGB, but starts with luminesence as the primary driver, meaning that more data being able to be saved does mean crisper blacks and lighter lights)
Normally I'd get in with the debunking crowd, but this one is genuinely a little more complicated than 'the data's digital, pfft' because there's a mass of electronics that operates very much in the analogue that sits next to it.
I should add, I have some not-so-fond memories of trying different techniques to rescue damaged CDs from back in the day that people thought were weird but have practical science attached to them. At one point I had a kit with several different kinds of brush, a couple of different buffing pads and a small tube of supermarket own brand toothpaste as a mixer. Rescued a number of mates' scratched CDs with that little kit - but of course I haven't needed it in years.
And as for gold plated HDMI, yeah, no thanks. The data comes out of one end already digital, goes in the other already digital, the gold plating, it does nothing!