So far as I know, there aren't a lot of 8-player local multiplayer games. The only obvious answer is the Jackbox games, using your phones as controllers.
Beyond that, I did find this Steam curator, who seems to specialize in 8-player games. From thier list, I recognize Gang Beasts, and Pico Park: Classic Edition. Party Golf, Screen Cheat, and Cobalt also all looked interesting, but I've never seen anyone play them.
Maybe I'm wrong, but I don't think Boomerang Fu or Overcooked support 8 players?
I'd expect that, although I've noticed this trend continuing (and seemingly getting worse) for months now.
Looking at the more detailed breakdowns, it looks like there are a couple of servers (Lemmit.online, alien.top among others) with huge numbers of posts/comments that appear to be entirely bots. Are those counted in the stats? Could those be messing with the overall graphs? If Lemmit's quarter of a million posts a month are counted, its going to make the monthly posts stat useless when even .world only has about 15k posts a month.
Edit: Comparing the graphs to the server list, it looks like Lemmit is counted, so the main graph is likely misleading. I did look through some of the bigger servers, and their rate of posting seemed fairly linear, but there isn't a good way to check overall.
I've definately noticed it too. I've tried to look for stats, and most seem to indicate that there is plenty of activity, but I dont really see it. At this point, I can scroll through the day's all feed in like 20 minutes, nonetheless my subscribed feed. I kind-of wonder if theres one or two instances with a lot of bot activity effectively inflating the numbers.
Edit: Is there a way to see monthly posts by instance, or compare percentage of posts? That would be an easy way to prove or disprove my bots theory.
Edit 2: fediverse.observer shows monthly (Or rather, total by month) local posts by instance but not federated, and their overall stats are warped by a few bot instances that you can't filter out. That said, for local posts on a few of the big instances, the rate seems stable. That said, smaller instances are shutting down so I don't know if that has an impact on the overall posting rate.
The economy is terrible with both hardware and software becoming more expensive, theres a good selection of free and long-lifetime games (be it live-service or just very long and replayable), and a lot of the newer paid games have become worse.
I'd be significantly more suprised if this wasn't the case.
The main problem is it turning itself on with no input from or feedback to the user, and not giving the user access to the key without using a Microsoft account. I've heard of people getting screwed by this because they set up with a local account and thus never got their secureboot key (or did, but it was hidden somewhere and they were never told to save it).
Mastercard? Yes, but then they'd have to admit that they were in the wrong, so it seems unlikely.
Tl;dr: Mastercard says they didn't "force" Valve to remove nsfw games. They just told them that if they didn't remove the games that were complained about by Collective Shout, they'll block them.
In my opinion Luanti is a living proof that top-down extensibility aka "we make monolithic engine in C++ and then provide some APIs for scripting via bindings for some scripting language on the side" doesn't work well. You can't change main menu, you can't fix player controller (and the default one sucks), you can't write your own renderer, etc. Because developers didn't imagine someone would want that (actually they probably did, but they simply don't have capacity to provide this). Good extensibility/modability should be automatic, on binary level. Like what you get by developing in bytecode/JIT-compiled languages like Java/C# or in old Unreal Engines where everything was done in bytecode-(de)compilable special language called Unreal Script.
The last graph is total posts across Lemmy, so its only about 300,000 posts a month, although notably, about 250,000 of those are on a bot server no one is federated with.