We’re clearly missing some huge part of the story.
Obviously, faster load times should improve engagement. So if engagement went down with lighter embeds, that implies that some feature or functionality had to be sacrificed. Yet this blog post claims nothing is sacrificed.
Something is missing from this story.
I wonder if it’s a similar thing here. Maybe the lighter embed is small enough that a huge swath of people is actually able to load a video now, but they bounce because playback is too slow on their device/connection. It would shows as a higher percentage of “person stopped watching video” instead of (nothing, because they never fully loaded the player).
I don't think it's this clear-cut. People can have different levels of expectation and tolerance for loading times of different things, and the page (post/article/whatever) is a different thing to the video embedded in it.
The load time of the rest of the page is also not necessarily limited by the loading time of the embed (depends what you're looking for). Arguably, the overall page load time will mask the content (embed) load time.
Finally, and perhaps most importantly, the loading time of the embed is a different thing to the loading time of the video. The metric to look at may be click-to-playback lag.