Xbox developer tools are on the way for AI NPCs and more.
I’m most looking forward to a jump in general NPC intelligence and ability. I just want to exist in a world where I can ask an NPC to do “quests” for me, so I can run my shop in peace.
The sad fact is that most of the generative AI advances don’t really help with this. They might make npc dialouge more flexible and less static, but nothing in the toolbelt Microsoft is offering would improve general NPC AI.
Maybe down the road we’ll see something, but I suspect MS is really just offering text and modeling automation.
I’m not sure about the Microsoft toolkit specifically, but I’m talking about something along these lines:
https://arxiv.org/abs/2304.03442I’m actually experimenting with something similar at the moment, though it’s very early stages right now.
Or just ask them to move out of the way, or tell them to go a specific direction when their path finding fails.
You can do that without ai
And still it is not really a thing, hopefully ai can fix it.
ai isn’t a magic wand that just magics up whatever you want. a lot of engineering has to go into making it do things. that engineering effort can go into doing other things. it’s just where developers choose to spend engineering time.
You can do that without ai
And yet, it would be far more dynamic - not to mention easier to implement - if it were powered by an LLM.
no, it wouldn’t. you would spend 10x the time to make a generative model that fucks up constantly from biases you never imagined.
Honestly, fair enough, it would take a lot of work to have an LLM direct your game in the intended way.
That said though, to create an AI system equivalent to an LLM would be even more work.
I think a lot of this comes down to AAA vs Indie, as well. For a AAA game, there is a lot more pressure to keep the LLM in line - which is of course very difficult if not impossible. For an indie game, though, the goofiness can be part of the charm, I think.
I guess my point is that I’m just excited to see what people can come up with. The point of games is to play, and I personally think LLMs are fun to play with.
I think the idea that it takes more work to do things by just engineering it in is flawed, to say the least.
Building llms is incredibly difficult. Building good training data is even more difficult. Then ironing out all the problems and biases in your model is even more difficult than that. You often need to build new models just to correct the old models.
There are places ai does well, it is however not a magic wand that just makes everything easier. It is far far far from that.
Sure, but once the model is trained and all that, the developer doesn’t need to worry about any of that - it becomes a black box as far as actual implementation goes.
I don’t think anyone is proposing that game devs create an LLM from scratch.
I don’t really think you grasp the problem at hand here. You think as long as their is a conversational llm everything else is solved. No.
The dialogue stuff is pretty interesting at least. As long as the models needs to run in the cloud and not on-device, I don’t see it catching on - but who knows in a few years?
milo?
