Right? Like, figure out how to sort and fold laundry.

Work is hard, modern games are easy.

Shoot this dot, go to that dot, bring "x" to this dot...

Anything requiring critical thinking like a puzzle, the bot just brute forces trying every possible outcome until something works.

It works because games have a finite set of options and at least one progresses. The real world isn't that clean, and you don't get infinite chances to get 1 correct solution.

Shut up and get back down the mines or your robot manager-security-cop will have to terminate you. Meet your billionaire-enforced quota and you'll live to enjoy another day of mining.

Next: game companies can subscribe to NVIDIA anti-AI-cheating AI. This is innovation under capitalism.

So you can spend more time working shit jobs for shit pay. Aren't you happy? 😊

I want LLMs to be used to create thousands of characters that have agency in like, Skyrim or something, and to watch the absolutely unhinged fucking chaos of what it creates.

I don't want it to play games for me.

I can't believe I'm not really not really flying a spaceship!

Leonard > Amy

how cheating (particularly single player) can be fun

Have you never played a game with unreasonable grind that saps all the fun out of it ? Often just to drive player hours and 'engagement'. Multiplayer cheats however can die in a fire.

Cheats in single player games is a blast.

It gets old quick but it's fun for a while.

I don't watch others playing games, either, but someone who likes those streams told me he didn't see a difference between watching good gamers play games and good football players play football.

You never used cheat codes in GTA San Andreas and the like? It's a lot of fun

Depends, are you rich?

Ive got incredible news for you

Except you don't need a neural net to do this, nor would you really want one for it because it would be incredibly slow unless you use the GPU to run it which might be a bit busy running the actual game. It's like a "make game run worse" option.

In game AI can be pretty sophisticated even without the addition of neural nets which could be OK with determining strategy based on player action, but really isn't going to be good for the entire control.

This kind of thing is only useful as tech demo for playing platformers and the like with the same inputs a human player would, basically more like a bot.

Soylent green I guess. No money, no fun, no job. You're free to die.

I suspect it's not a lack of playtesters that's the problem, but harsh deadlines and crunch. That type of environment leads to tech debt to get things working fast, which leads to hard-to-manage code, which leads to bugs...