Imagine this in a future GTA game where the news loop is closed and self generating. Endless radio content and commentary based on havoc in the city, winning online gambles etc.
```
That's all for the weather report. Now we have some breaking news. A maniac has stolen a tank from a military base and is leading the police on a wild chase through the city streets. We have our reporter on the scene with more details. Stay tuned for this developing story.
```
Pre-programmed common stuff sure which many games do as they can tell where you are, vehicle in and any weapons used etc, steal a (police car|military jet|ambulance) etc and they can craft scenarios using audio for those but for more natural random somewhat unpredictable stuff you would have endless combinations that need to be accounted for. It would have to act in response to a bunch of actions so not feasible.
A good example is Mortal Kombat where each character has several lines for every OTHER character (https://www.youtube.com/watch?v=L85QApISlvA) they are about to fight, that's a LOT of voice work as opposed to using the character lore and their history and or relationship with the opponent to come up with something fresh that's (witty|snarky|sad) to say etc.
Yes but this stuff is taking 100 gigabytes per game
Once we get flexgen running Llama (or some other combination of optimizations) there will be no audio files and these LLM’s can be run client side on consumer hardware
Exactly, imagine every NPC just has some ai generated backstory, personality traits, dislikes, habits etc and then these interact with some other NPC or group with same and they use these to have natural flowing convos. Now imagine debates with groups that learn and have the possibility to "evolve", I'm getting a mix of Red Dead Redemption and Dwarf Fortress here for content, the possibilities are endless.
