Re:Jev: New frontier model 40-400x cheaper and 20-200x faster
by nylonstrung on Wednesday September 16, 2026 @04:06AM
(#49723391)
In a case like this it still seems more appropriate to encode that data in tabular form and use a tabular foundation model
Re:Jev: New frontier model 40-400x cheaper and 20-200x faster
by adastra22 on Wednesday September 16, 2026 @04:13AM
(#49723440)
Forgive my ignorance. Tabular foundation model?
Re:Jev: New frontier model 40-400x cheaper and 20-200x faster
by salomonk_mur on Wednesday September 16, 2026 @09:45AM
(#49726973)
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.
Re:Jev: New frontier model 40-400x cheaper and 20-200x faster
by headgasket on Thursday September 17, 2026 @05:25PM
(#49746756)
I'be interested to see how they get doom frame json representation...
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.