I'm honestly just really annoyed about this "society and environment"-spin on advise that would have an otherwise niche, but perfectly valid reason behind it (TFA: slow satellite network on the high seas).
This might sound harsh and I don't mean it personally, but making your website smaller and "being vocal about it" (whatever you mean by that) doesn't make an iota of difference. It also only works if your site is basically just text. If your website uses other resources (images, videos, 3D models, audio, etc.), the impact of first load is just noise anyway.
You can have a bigger impact by telling 100,000 people to drive an hour less each month and if just 1% of your hypothetical audience actually does that, you'd achieve orders of magnitude more in terms of environmental and societal impact.
Now, it is true that it didn't save much because probably many people were uploaded 8K videos at the time, so drop in the bucket. But personally, I found it quite inspiring and his decision was instrumental in my deciding to never upload 4K. And in general, I will say that people like that do inspire me and keep me going to be as minimal as possible when I use energy in all domains.
For me at least, trying to optimize for using as little energy as possible isn't an engineering problem. It's a challenge to do it uniformly as much as possible, so it can't be subdivided. And I do think every little bit counts, and if I can spend time making my website smaller, I'll do that in case one person gets inspired by that. It's not like I'm a machine and my only goal is time efficiency....
Depending on the type of video this may not matter, but it often does. For example, my FPS gaming and dashcam footage gets utterly destroyed if uploaded to youtube at 1080p. Youtube's 4K seems roughly equivalent to my high bitrate 1080p recordings.
And no, a million small sites won't "become a trend in society".
Plus, trying to signal your way to societal change can have unintended downsides. It makes you feel you are doing something when you are actually not making any real impact. It attracts the kind of people who care more about signaling the right signals than doing the right thing into your camp.
If we want to really fix places with bigger impact we need to change this approach in a first place.
This is micro-optimisation for a valid use case (slow connections in bandwidth-starved situations), but in the real world, a single hi-res image, short video clip, or audio sample would negate all your text-squeezing, HTTP header optimisation games, and struggle for minimalism.
So for the vast majority of use cases it's simply irrelevant. And no, your website is likely not going to get 1,000,000 unique visitors per hour so you'd have a hard time even measuring the impact whereas simply NOT ordering pizza and having a home made salad instead would have a measurable impact orders of magnitude greater.
Estimating the overall impact of your actions and non-actions is hard, but it's easier and more practical to optimise your assets, remove bloat (no megabytes of JS frameworks), and think about whether you really need that annoying full-screen video background. THOSE are low-hanging fruit with lots of impact. Trying to trim down a functional site to <14kB is NOT.