alt Hacker News> Assuming there's no funny business, the Doom demo is cool.
The Doom demo seems very funny business. They're not feeding it video, they're feeding it a text description of what's going on in the game. It's not reading pixel data.
I think LLMs would play a lot better with that input too but Jev does seem to have a huge speed advantage; I don't know if the other models could do that in real-time.
Replies
nylonstrung • yesterday at 8:06 AM
In a case like this it still seems more appropriate to encode that data in tabular form and use a tabular foundation model
➕ show 1 reply
Well, the whole point of the demo was showing that transforming from game state to text to action is so quick with Jev that it can play in real time.
So no, the other models cannot play in real time. Hence why this is interesting.