Ultimately we can expect future games to use AI in real-time to generate an infinite amount of in game objects.
Sort of like procedural content generation on steroids.
Trying to code procedural generation to create an entire city of believable houses which are all explorable inside and fully furnished is a pretty complicated problem requiring a huge amount of work.
But using AI to do that will become trivial.
the problem is a game needs a purpose and a random environment made by someone that does not know purpose is pointless
It also likely means the game mechanics are banal if you can be guaranteed a random generator can make an environment suitable for the game
That is another thing I noticed, games where people are obsessed with realism rarely pay off. Its a GAME that has gameplay mechanics.
And with larger games that feel slapped together, I started to not care. Like why should this thing that got randomly made have value, be interesting, etc. Map makers usually imagine something happening, have a theme, and their work reflects it. I tried some game maps with AI generated stuff in RTS games, and the ones that are playable, are ridiculously bland, practically open
'open world' lost its sales appeal to me after I tried to play the old dumpster fire, boiling point. Big open world like requires the gameplay complexity to drop dramatically, or it basically HAS a 'tested' route that you can play, and anything else is a senseless challenge or cake walk. Good games with 'choice' give you some options, but its all like tested, because if its truely open world its impossible to make sure you get 'value' from your work. And you can't help but notice things that are along the lines of "i see why you did this" when there is a sudden gameplay shift to try to 'normalize' the open world aspect of the game. It feels MEGA orchestrated when that happens. 'scaling' works on math but in reality it feels like stupid and its a immersion killer. You feel like you are playing a mathematical function when this happens.
I would say making a good, fun, immersive game is extremely difficult and requires talent that is just not there (with genius designers). Throwing money at it seems to not do anything.
I am not sure, but suspicious, that vampire masqurade 2 which recently came out might have used some AI help for environments. The game looks good, and its big, but the reviewers are calling it a 'walking simulator'. Because all you do is walk around, environment does not feel like it has a purpose or inspiration/vision.. it just looks OK and you walk through it.
I worked on some game modding before, with a focus on interactive environment THAT IS USEFUL to the game, and I suspect, you might need 10,000 years work, to get a true 'open world' to work right in gameplay mechanics. I am talking optimizing a few hallways and a few rooms on a space ship, I thought I might have a inkling of something, after about 400 hours of work, that added something to a pretty well established game-rule / play set. You need so much tuning and imagination to even do something simple like figure out what a fair opening/closing time for a door is, how strong the door is, how far you can see and how other players would feel about it (few features on a 'space' map), and if it fits the theme of the game. A good game usually has MANY complex interaction possibilities, and making them work for a specific environment is DAMN HARD if the environment is part of the game and not just backdrop on a flight sim ( a 'basic' flight sim, not a cool one that involves flying a chopper around really low like the movie 'blue thunder', let alone having enemies, or other players, or some adrenaline.)
i would say that they should work on making john woo clones instead, that are employed as management/oversight. they better save some of his DNA so they have a fallback plan when AI fails to solve gaming industry problems
like any other engineering IMO, they say its 'solved' and 'easy', then when someone wants a few more details, turns out its not solved, not easy, not cheap, not fast and actually a path of struggles and achievements, and many of the expert services (golden bullets) for that feel like scammers/inappropriate choice s if you work hard for a week to figure out the basics of the problem lol. And you need interested+smart people and time
No easy solutions, no good 'buy it now' solutions... that is what reality for development is
Those AAA games just tell you the same thing the movie industry can tell you, that with money you can get props and extras. It often won't help make a good film, but if you make it extravagant enough, people will go just to see props and extras with the expectation of seeing props and extras (stunt). You might even say it does not 'buy' the lead actors of the film, not really. They are "there" but are they are not doing what they are expected to do (to the audience I mean). With main character A-list actors, its even funnier, because they can actually get pissed at the movie production lol.. and with games the 'character devs' get pissed too, and their character does not 'perform' (i.e. interesting tidbit about the tomb raider games, sometime after tomb raider III in 1999, the 'people' behind lara croft got so fed up with the character they demanded the developers let them kill her, to end the series... and it showed, tomb raiders IV and V are considered to be pretty trash). The extra's however are available, and props can be made by a expensive prop industry. I wonder if AI to video games is kind of like CGI to movies, which essentially made props infinite. They have been saying CGI will improve hollywood for like 50 years now lol and it keeps getting worse.. so you might say that the cost of (low quality) prop-equivalents in videogames may go down. And the greenscreen actors 'ability' to 'make believe' the enviroment they are in is often AWFUL Am I supposed to not notice they are 'visualizing' something instead of doing something?
(you might as well put some kind of icon over the actors head that indicates he is in the process of 'solving' or 'loading' his 'environment'.
have a 'real set' or you fool no one. In some of the more elaborately decorated modern games, I see it, and it might as well be a grey wall, because I instantly know its not worth looking through and it does not do anything, and if someone hides shit in there, it just feels annoying. I think they thought about what they are doing more deeply when it was harder to make. When you have elaborate pre-made libraries and stuff, it is just random shit, why should I be fascinated by random shit that was quickly put together without thought?
You need some AI (with the increased code complexity) to automate some of the labor to keep it from becoming 'big corporate management' and to keep the team-roles relevant/important (camel problem), but it becomes soulless very quickly
And to actually answer the thread, I think that
1) people are not happy with the problems that games are having
2) video game companies claim AI will solve these problems
3) people are pretty sure AI won't solve these problems and are scared it will make it worse, especially because the video game industry does not usually fall in line with the latest 'revolutionary' tech trends, which historically have seemed to not change much and annoy people (is keysight getting involved in blockchain NFT yet?).. so they see it as feeding the latest tech trend into a industry that normally had little to do with unhelpful tech trends that excite investors. AI is a 'golden bullet' solution, and usually the golden bullets some how did not reach the game industry too much, and we know what golden bullets do. Have big data IOT subscription bath tub drains magically changed your life?
I am still hopeful maybe some smaller devs can make good games because of it though, maybe 1 game in the next ten years, i feel like I would take anything at this point.
Home
Help
Search
Login
Register
