[Games] Lecture 9 - AI Making Games (Paolo Burelli) Flashcards
Technical problems makes design/creativity flourish example
War game where you have to sneak instead of killing everyone because having too many bullets on the screen wasnt technically possible
AI now is more like LLMs and transformer models, back when games like No Mans sky where being made, it was seen as “generative ai”, which we might not categorise it as today
One part of making games with AI is the fact that we cannot just store evrything, but also the part of we cannot have humans and sitting designing so much game
Explain the relationship between the generator and the discriminator, we use it when we want to make good generative AI
What is the problem with only using the generator and discriminator when generating images for your game?
The generator only learns to make things that looks like something form the data set, it will not make something “new” and interesting
What is interactive narrative?
Games like facade, using individual agents with specific objectives. Now we could use LLMs to roleplay these characters.
Who is games by angelina
An I made by Cook? who made an AI to make uo games themselfs
Online vs Offline
Stochastic vs
Deterministic
Is the product completely reproducale? If you have the same kind of seed, will it be the same random process?
Constructive vs
Generate and Test
Constructive: no feedback loop, the main difference is that there is no feedback loop
Generic vs Adaptive
Level of Control
How much is handcrafted? How much is by the algorthim?