News for hackers. Stuff that matters.
Re:Veo 2: Our video generation model by zb3 on Monday December 16, 2024 @12:28PM (#42433151)
We should collectively ignore these announcements of unavailable models. There are models you can use today, even in the EU.
Re:Veo 2: Our video generation model by ilaksh on Monday December 16, 2024 @01:20PM (#42433695)
Actually there is a pretty significant new model announced today and available now: "MiniMax (Hailuo)Video-01-Live" https://blog.fal.ai/introducing-minimax-hailuo-video-01-live...

Although I tried that and it has the same issue all of them seem to have for me: if you are familiar with the face but they are not really famous then the features in the video are never close enough to be able to recognize the same person.

Re:Veo 2: Our video generation model by creativenolo on Monday December 16, 2024 @05:42PM (#42436310)
It was announced weeks ago.

50 cents per video. Far more when accounting for a cherrypick rate.

Re:Veo 2: Our video generation model by the8thbit on Monday December 16, 2024 @02:50PM (#42434650)
I don't see why, unless you think they're lying and they filmed their demos, or used some other preexisting model. I didn't ignore the JWST launch just because I haven't been granted to ability to use the telescope.
Re:Veo 2: Our video generation model by zb3 on Monday December 16, 2024 @04:36PM (#42435720)
Back when Imagen was not public, they didn't properly validate whether you were a "trusted tester" on the backend, so I managed to generate a few images..

..and that's when I realized how much cherry picking we have in these "demos". These demos are about deceiving you into thinking the model is much better than it actually is.

This promotes not making the models available, because people then compare their extrapolation of demo images with the actual outputs. This can trick people into thinking Google is winning the game.