“machine translation drove 4x incremental" daily active unique users in Q3 compared to Q2.

So it sounds like they failed to gain new users in their original core areas (like North America) but gained them by expanding into new countries.

In other words, the revolt was successful, and if we limit the area to those places, then arguably the site never recovered from the exodus.

But of course, new users aren’t going to understand what’s been lost - it’s all new to them - and spez’s site seems to have been heavily enforcing unspoken bans on alternatives to their site, so it’d be hard to reach out to them. (And that’s not including the language barrier we’d have with most of these new users, it seems.)

It seems to me that spez did win - the dropping of the free API got him the big money deals with places like OpenAI and Google that he needed to turn his company profitable.

But it was a pyrrhic victory at best - he would have been able to grow his website to gain these same new users via machine translation without having killed the API and apps and lost all those OG mods and users. Also, the use of AI here means the new data he’s getting is less valuable (since it’s no longer purely human) which risks making his site less valuable to OpenAI, Google, etc in the future.

Meanwhile, the fediverse continues to grow, and I think that the existing Bluesky[HTML_REMOVED]fediverse bridges mean that Bluesky users can participate on Lemmy instances, meaning that even Bluesky arguably has become a competitor of spez’s site now.

So he may have accomplished his short term goals, but at the same time he’s provided too much fuel to his competition and needs to watch out before someone else manages to eat his lunch.