Refreshing read. A lot of these posts people tend to try to justify their decision mostly after-the-fact based on usually very poor data or just some opinions/expectations. The author however completely acknowledges that their reason for doing so was just because they wanted to have some fun, and got some benefits out of it too.
The LOC change is fairly dramatic but it does make me wonder how much potential dead code was in the old codebase. Especially if it's been around for a while it'll likely have accumulated baggage that the newly implemented version just doesn't have.
Rewrites are always very hard to quantify. I read on twitter recently programming is 95% thinking/design and 5% typing/doing - once you've done the first version you know what works and what doesn't and where the pain points are, so it's much easier to write a second version which is vastly improved (as long as you're not overly ambitious).