I think we move slowly, like a glacier, working in the background on stuff like an AI for mass-converting PDF to Asciidoc, and performance improvements, and containerizing the server landscape and etc., so it seems like nothing is going on, even tho it's a small horde of active developers, admins, and testers scrambling around.

But it has to handle millions of publications and one publication could be the Encyclopedia Britannica. We have to revamp and reengineer a lot, under the hood, but it seems to be working. Better than we originally expected.