Quote:
Originally Posted by Farang
that is pretty cool
|
Only a couple years ago, when the super early AI came out, the images it made were so ugly and the earliest videos of Will Smith eating spaghetti were sooooooo bad... then I saw a recent video of Will Smith eating spaghetti and it was fucking almost spot-on!
Shit is changing so fast. One of the tools that I wanted to buy to make my Vigilante Simulator game, was this tool to make city blocks for like $500 USD.... procedurally generated urban terrian I could customize to make it look like a Vancouver slum... anyways, I wouldn't be surprised, if, by 2026, you could just type into a prompt:
"Generate a west coast Canadian urban slum based on google maps coordinates X by Y from 300 block of West Hastings street to 300 block of East Hastings street, Vancouver, Canada"
And it will create you a nice 3D model of the area you can plugin to Blender or Unreal Engine 5, premium features could add interior models and such. Maybe super-advanced AIs could add in cockroaches, mice, basic NPCs, car traffic on the street, working stoplights, weather system, day/night cycle... just a matter of time before generative AI works off some sort of default model of reality, so you could just import in large text prompts to describe the characters, story, time period, etc etc and the AI can generate you a semi-fleshed out world that would normally take a single developer years and years to build... probably not too far off. Just need to be able to tell the stories in these worlds that people would want to play.