HN user

FilipSivak

12 karma
Posts0
Comments8
View on HN
No posts found.

You clearly don't understand what multimodal means. Multimodal is for example new gemini where you can input green car and get the very same car, only with red paint. Multimodal LLM can do the edit in the latent space, which is the key.

Very misleading title, and you won't get away with it by using word "mulimodal" either.

This is most definitely fake and misleading. There is no way the narrated gameplay shown in the video is a result of an autonomous agent. The capabilities of proposed system (of which no evaluation was shown) do not fit the capabilities of the agent shown in the gameplay. For example, Lara comments that "wolf tracks are backwards" but there is no way the agent as described in the video would be capable to come at such conclusion.

Go and look into the comment section, which is full of naive people believing this with comment "actually, AI is now very advanced". Content like this skews public perception of what is and is not possible and is, imho, harmful.

This is a path trace. Such render can take hours (and there is no temporal denoising in Uneal yet). Now, VR, must render the scene twice, once per each eye. Also, VR should run at relatively high framerates, such as 90 or 140 FPS. I think achieving this quality in VR won't be anytime soon.

That said, the scene might not be too far off when using Lumen, which is a performant global illumination method. I'd like to see the scene rendered in Lumen and shown side-by-side.

This is absolutely horrible idea. There are cities that are safe (Lvov). Dont book airbnb that refugees can book themselves. If you want to send money, send money to humanitarian organisation.

As for cities that are obviously not safe, still, humanitarian org. might be better.