I'd say Minecraft's multiplayer experience is close to what Anon describes as "good multiplayer", probably because it hasn't changed much in 15 years - there's not even an in game server browser (at least on the Java edition), and playing Minecraft in and of itself is usually a big time commitment so you're more encouraged to find a couple of servers you like and stick to them.

However, the last time that I feel like I integrated into a server's community was 4 years ago - a blank server list doesn't really encourage you to go looking for more, and it's been harder to commit time as I get older and have more responsibilities (that I ignore anyways, but still).

I think Lethal Company also has a lobby system without matchmaking, but I haven't played it so I don't really know.

I do agree that the game's story/atmosphere should be designed for singleplayer first, since that's what drew a majority of people to the first game.

However, I think MP could still retain some sense of atmosphere as long as they have a good in game voice chat and keep it proximity based, like in Lethal Company.

It isn't scary at all for your teammate to tell you exactly where they just died, how, call out the leviathan heading your way, then continue with a joke they were making. But having a friendly conversation over the radio cut off by their violent death to horrors unknown could be scary, even if facing the unknown alone is fundamentally scarier.

My theory is he doesn't remember, keeps trying to, and then makes a new post about it when he remembers

Outside of straight cash and ads, off the top of my head a user could give a website data, content, or computing power. Which, as I kept writing this, I've found aren't perfect alternatives.

Personal data collection seems compelling, since the data can be sold to hungry data brokers looking to optimise their ads, but tech-savvy users want to keep their data safe, either by using plugins to block ads and tracking, or by not using your website. And you'd also have to have no soul to do this.

User generated content gives users a reason to engage and return, and it also means you could save money that you'd have otherwise used to pay someone to make content. If you rely on this too much though, ethical concerns become apparent - last I checked, Reddit mods are unpaid.

Volunteer computing could maybe lower costs by offloading some server calculations onto volunteer's computers when idle, but I don't know if it could even be used for that. It's probably a non-starter for websites, too; to a user it would seem like your site was asking them to install a crypto miner.

... this comment is getting too long and doesn't really have a point. But I can't let the 45 minutes I spent writing it go to waste so easily. Hm... what if I combined all 3 ideas?

Yes, a website that asks you to volunteer idle computer time to train an algorithm that can both be outsourced to other companies and used to analyse your personal data, which itself can be given to other companies and used to reccomend you posts you are more likely to comment on, adding value to the website! Surely this has none of the flaws that I described before.

If you still want to make the game despite that, I'd recommend watching some of Pirate Software's youtube shorts for motivation. He's got some great gamedev advice.

This one that he uploaded today feels relevant: https://youtube.com/shorts/TBxhiw-Hpxc

I'll hold myself back from sharing more for now, in case you don't care. And also because it's 3 am.