Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Imagen Video: high definition video generation with diffusion models

imagen.research.google

Post reply on HN
Print view View original

391–400 of 500 posts

Re: Imagen Video: high definition video generation with diffusion models

#391

Post by azinman2 »

gundmc wrote: post #360
azinman2 wrote: post #333

> However, there are several important safety and ethical challenges remaining. Imagen Video and its frozen T5-XXL text encoder were trained on problematic data. While our internal testing suggest much of explicit and violent content can be filtered out, there still exists social biases and stereotypes which are challenging to detect and filter. We have decided not to release the Imagen Video model or its source code…

Speculation, but I think the most straightforward read of that statement is not about preventing this type technology from negatively impacting society broadly (since as you and others pointed out, there are numerous similar actors creating similar systems), but Google doesn't want to the bad publicity or legal risk of problematic outputs of their models. I think they're terrified to be honest.

I worry that everyone across academia and industry are creating these models as if they’re burning needs without any recognition of what’s happening collectively. It’s not neutral at all, yet everyone involved seems to think they can put in an “ethics” paragraph to absolve themselves. As we’ve seen in the last 20 years with technology (and certainly many technologies in the last 200 years before that), it simply isn’t true.

Re: Imagen Video: high definition video generation with diffusion models

#392

Post by Morgawr »

zarzavat wrote: post #369
rgmerk wrote: post #315

Earlier quoted context omitted.

Be serious for a moment. If I circulated a convincing-looking video of Rick fucking goats to his parents, partner and his boss at the school where he works, that could easily do considerable harm to Rick.

I don’t think that is a good example, because the harm in that case is the embarrassment and you can achieve the same results in 10 minutes with any image editing software. For AI image generation to cause harm specifically, the harm has to be consequent to the additional realism. IMO most of the harm from AI is likely to come from people not believing things that are real, and dismissing reality with “that’s just a…

> the harm in that case is the embarrassment and you can achieve the same results in 10 minutes with any image editing software

The harm is not the "embarrassment" of seeing someone in the likeness of yourself (or your son, your friend, your partner, etc) doing something shameful. The harm is the fact that people are very likely to believe it is true and it's not a fake obviously edited photo or video.

You can disagree on the seriousness of the harm or risk or danger or whatever but I think the distinction between an obviously silly/embarrassing fake (a puppet, papier mache, badly done photoshop picture) and a realistic convincing deepfake video is pretty obvious. They aren't even in the same ballpark.

> IMO most of the harm from AI is likely to come from people not believing things that are real, and dismissing reality with “that’s just a deepfake”.

This is also a really good point and I agree it's a danger.

Re: Imagen Video: high definition video generation with diffusion models

#393

Post by azinman2 »

dogcomplex wrote: post #363
azinman2 wrote: post #333

> However, there are several important safety and ethical challenges remaining. Imagen Video and its frozen T5-XXL text encoder were trained on problematic data. While our internal testing suggest much of explicit and violent content can be filtered out, there still exists social biases and stereotypes which are challenging to detect and filter. We have decided not to release the Imagen Video model or its source code…

Cryptographic trust (combinations of identity proofs, including passport, passwords, social and family networks of vouching, fingerprints, behavioral data, etc) will be the only way to trust any digital information very soon. If it's not vouched for by someone with proof that they're a real person, it's fake.

That’s not how human nature works. Show a video that fits cognitive bias, and the non-technical non-sophisticated people of the world will believe it. And that’s assuming it’s even technically feasible to solve, which it likely isn’t.

Re: Imagen Video: high definition video generation with diffusion models

#394

Post by bredren »

fzysingularity wrote: post #278

What's next? Dreamfusion Video = Imagen Video (this) + Dreamfusion ( https://dreamfusion3d.github.io/ ) Fundamentally, I think we have all the pieces based on this work and Dreamfusion to make it work. From the looks of it, there's a lot of SSR (spatial SR) and TSR (temporal SR) going on at multiple levels to upsample (spatially) and smoothen (temporally) images that won't be needed for NERFs. What's impressive is th…

This direction will provide the visuals but what also must be brought in is a language model and text to speech (TTS) so that you may talk and interact with these things.

Re: Imagen Video: high definition video generation with diffusion models

#395

Post by girvo »

bringking wrote: post #28

If anyone wants to know what looking at an Animal or some objects on LSD is like, this is very close. It's like 95% understandable, but that last 5% really odd.

Yeah! I've tried to explain to people what taking LSD can be like, to those who've never experienced it. It's very similar to the output from these tools: the same stimulus but exaggerated, wrong in subtle or not so subtle ways, uncanny and fascinating. Basically never creates something from the whole cloth, out of nothing so to speak.

Re: Imagen Video: high definition video generation with diffusion models

#396

Post by bredren »

ugh123 wrote: post #260
fassssst wrote: post #241

How long until the AI just generates the entire frame buffer on a device? Then you don’t need to design or program anything; the AI just handles all input and output dynamically.

Sounds like the human brain. Scary!

Reminds me of this oft-quoted Wozniak bit between my pal and me:

[The Future of AI Is] "Scary and Very Bad for People"

https://finance.yahoo.com/news/steve-wozniak-future-ai-scary...

Re: Imagen Video: high definition video generation with diffusion models

#397

Post by melagonster »

origin_path wrote: post #379
rgmerk wrote: post #315

Earlier quoted context omitted.

Be serious for a moment. If I circulated a convincing-looking video of Rick fucking goats to his parents, partner and his boss at the school where he works, that could easily do considerable harm to Rick.

How? Rick says I did not do this, it's an ai fake and I'm being harassed. At that point, assuming people believe Rick, then it's likely he'll receive sympathy and support rather than considerable harm. So there seems to be an implicit assumption here that the risk is faked material where people don't believe it's fake, for some reason. And the fix for that would be to ensure that the easy to use versions of generator…

but when everyone see him, there is always the photo in their brain. this is about goat, but they fake pedo photo? Rick will lose his job first.

Re: Imagen Video: high definition video generation with diffusion models

#398

Post by colordrops »

jquery wrote: post #355
i_like_apis wrote: post #261

The concern trolling and gatekeeping about social justice issues coming from the so-called "ethicists" in the AI peanut gallery has been utterly ridiculous. Google claims they don't want to release Imagen because it lacks what can only be called "latent space affirmative action". Stability or someone like it will valiantly release this technology, again and there will be absolutely no harm to anyone. Stop being so to…

You call it "concern trolling", I call it responsible research. It's not "social justice" to be concerned about the ability of any entity to make propaganda videos for virtually free. "Ok Google, produce a CDC-type public information video about how vaccines cause autism and disability". Not that I would have a complaint if social justice was the sole thing keeping it from being released. Facebook managed to cause ge…

That would make sense if Google were somehow in a deserved position of authority to decide who is allowed access and what it's used for, rather than an advertising company with a heavily skewed bias that doesn't necessarily take the public good into account.

Re: Imagen Video: high definition video generation with diffusion models

#399

Post by hobofan »

splatzone wrote: post #271
javchz wrote: post #267

Earlier quoted context omitted.

Imagine you click a youtube video in a bad network envoirment, then the server sends like an alt tag equivalent for the video as a promnt, and the Neural Engine chip inside your phone create the first seconds of the video while it loads. We're fay away from it now, but I've seen less sketchy solutions being implemented.

I wonder if this is an area actively being researched, using models like these for video compression?

Yes, with NVIDIA Maxine probably one of the most prominent examples of it. I haven't dug into the SDK to see if they actually delivered it, but they announced that with NVIDIA Maxine they can do live videoconferencing with 1/10th the bandwidth.

Re: Imagen Video: high definition video generation with diffusion models

#400

Post by uwuemu »

atty wrote: post #322
sbierwagen wrote: post #296

Earlier quoted context omitted.

>There is a clear risk from these sorts of models as they get better - I mean recreating specific individuals’ likenesses in compromising images And the risk behind that is...? If you drill down with such claims the core is always "someone might use this to lie online" and the proposed solution every single time is: more surveillance. End anonymity. Have a Facebook account required to use the internet. Real name and…

I strongly suspect you’ve never been on the end of an internet doxxing/hate brigade if you can’t imagine how this could be used to make someone’s life a living hell. I’ll explain again that I think they can be used for bad actions, and also that they should still be released, because the benefits will outweigh the negatives. It does not hurt to admit that some things can be dangerous when used in nefarious ways. No o…

No post body was provided.
Post reply on HN