Hacker News Reader

Comment by visioninmyblood | Hacker News Reader

Wow! I was able to combine Nano Banana Pro and Veo 3.1 video generation in a single chat and it produced great results. https://chat.vlm.run/c/38b99710-560c-4967-839b-4578a4146956. Really cool model

View on HN

Neat use-case, though the sword literally telescopically inverts itself at the beginning of the scene like a light saber where you would have expected it to be drawn from its scabbard.

I'd be interested to see how Wan 2.2 First/Last frame handles those images though...

yeah sadly veo 3.1 has not caught up to the image generation capabilities. May be we need to work on how to make video generation more physically consistent. but the image generation results from banana pro are great.

another interesting use case with synth https://chat.vlm.run/c/1c726fab-04ef-47cc-923d-cb3b005d6262. made a puppet from a image of a model and made the puppet dance.

The feet are doing unusual movements. Reminds me of leaf node cumulative error in overcompressed hierarchical animation.

yeah the video models still do not understand physics the way humans do. We are getting there one step at a time. By the way, I am seeing a lot of people complain about google billing not working well. I was able to generate these for free without signing. Look at the results and try to come up with your own failure and working use cases.

That is an interesting error actually. It happened because both orientations of the sword are visually plausible, but not abrupt transitions from one to the other; there needs to be physical continuity.

Here is a reproduction of the Matrix bullet time shot with and without pose guidance to illustrate the problem: https://youtu.be/iq5JaG53dho?t=1125

I see many recent accounts posting vlm.run links and if this is what I suspect it is, that's normally not allowed here.

If you have concerns about spam, the right thing to do is to email the mods at hn@ycombinator.com with examples.

Reply on news.ycombinator.com