How do you guys deal with those self declared experts?

The trick with good AI is to make them seem to make informed decisions. Our AI is working in fireteams, each fireteam have a Director AI that coordinates them. Also we never cheat, if the AI does not know where the player(s) are they will start a search etc and use their last known position (or last estimated sound position if a gun shot). Our game is built upon CQRS and the bots too, so we can easy have a seperate dialog system that listens to the commands sent between AI and play a sound, so if a GetIntoCoverCommand is sent the bot can play a "Get down!" sound etc. Also its important that the Director AI does not micro manage the fireteam because a real world fire team leader does not do that, so we have some commands the Director AI can issue, but other than that each bot does its own separate logic. Lastly we have the knowledge about the world. A good AI system needs to know how to efficiently perform a flanking operation, this way the player will never feel safe. We paint additional room data on the navmash, which points out choke points (door ways etc), this way we can do classic graph path finding and find additional ways to the player

 

i think its just too early for multiplayer vr games
there are too few consumers
what's the number of steam users owning a vr headset ? something like 0.3% ?
its already hard to make a succesful game where theoreticly 100% of the steambase are able to play it
i also wonder how many of these people with vr headset are coorporations or startups, who will never play any games

maybe i am wrong and @Andersmalgren has different statistiks, but thats how it looks to me

 

At the same time, like I said earlier I would never try to make a FPS for desktop and compete with DICE and the other triple A companies, There are about 2 million headsets, so yeah, the target audience is really small.