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

341–350 of 500 posts

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

#341

Post by zarzavat »

atty wrote: post #319
dannyw wrote: post #286

Earlier quoted context omitted.

This is not new. Imagine not releasing tools like curl or nmap because it can be used for hacking. The issue is, as an industry and society, we somehow bought the "safety" and "harm" charade a little bit too much, and somehow think it's a reasonable argument instead of being completely insane.

I believe you misunderstood my post. (Or I have misunderstood yours) I was not arguing that the models should not be released. I was pointing out that the statement that no harm has been done with them was false, and that even though that is the case it is probably better for society, on the whole, for them to be open. We can both admit that the tools can and will be used for bad purposes, and come to the conclusion…

Given OpenAI’s extreme restrictions on Dalle 2, can anybody point to any harm that has been done with Stable Diffusion in particular since it launched? Even a single instance. Because I have only seen strictly positive coverage of people having fun with it.

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

#343

Post by throwaway23597 »

burkaman wrote: post #196
yreg wrote: post #188

Earlier quoted context omitted.

>You can't type any prompt that's "unsafe", you can't generate images of people, there are so many stupid limitations that the product is practically useless other than niche scenarios Imagen and Imagen Video is not released to the public at all. You might be confusing it with OpenAI's models.

They are probably confusing OpenAI with DeepMind, which is owned by Google.

No, I'm very much talking about the Google models. From the original link:

"We have taken multiple steps to minimize these concerns, for example in internal trials, we apply input text prompt filtering, and output video content filtering. 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."

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

#345

Post by matheusmoreira »

atty wrote: post #266
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…

There is a clear risk from these sorts of models as they get better - I mean recreating specific individuals’ likenesses in compromising images (or even worse, video). We’re not at that point yet, but these things are getting better fast, so it’s only a matter of time. The problem is that there’s no way to mitigate those risks except to keep the model behind an inference-only API, or not release it at all - as soon a…

These "risks" only exist because people somehow came to believe they actually have control over their "likeness". They don't. They never had. It's an illusion. The only risk here is exposing this for the lie it is.

Maybe once it's trivially easy to copy someone else's "likeness", society will finally be able to accept it and evolve past it.

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

#346

Post by matheusmoreira »

zakki wrote: post #316
Geee wrote: post #314

Earlier quoted context omitted.

Deepfakes are just an extension of imagination. We can already imagine people in any situation, making an image of it doesn't cause any more harm.

I don’t know why you put image and imagination in the same place. It is different. Imagination is the result of thinking. Image is the result of action. We can imagine killing people. Nobody harmed. But killing people in real life?

I can imagine a person killing other people. I can draw a person killing other people. I can make a computer draw a person killing other people. In all three cases, zero people were harmed.

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

#347

Post by throwaway23597 »

alphabetting wrote: post #157
throwaway23597 wrote: post #113

Google continues to blow my mind with these models, but I think their ethics strategy is totally misguided and will result in them failing to capture this market. The original Google Search gave similarly never-before-seen capabilities to people, and you could use it for good or bad - Google did not seem to have any ethical concerns around, for example, letting children use their product and come across NSFW content…

Providing search results of the internet is not comparable to publishing a tool that can create any explicit scene your fingers can type out.

This is clearly a matter of opinion. When you frame it as "providing search results of the internet" yeah sure it doesn't sound so bad. But there are things on the internet far more fucked up than anything I could imagine, let alone describe in such a specific way that a model could generate a picture of it.

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

#348

Post by infoseek12 »

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…

Google has done incredible work on basic machine learning research that has enabled other individuals and organizations to do amazing things. In terms of actually implementing new machine learning technology, they've pretty much hobbled themselves to the point of irrelevance. In some ways, it may be for the best that they've ensured that the future of machine learning will be written primarily by those who hold opposing views.

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

#349

Post by echelon »

flycaliguy wrote: post #335
atty wrote: post #322

Earlier quoted context omitted.

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…

I totally agree. People often argue that photoshop has been around forever and so on. Creating sophisticated pornographic video of any individual is brand new. Creating an app that realistically removes clothing from any photo is new technology. I actually just now came up with an idea for a browser plug-in that removes clothing from every image loaded.

And we're going to get very used to it in short order.

We may make laws that prohibit online harassment, and that should be the mechanism we use to deal with this. Not through technology bans.

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

#350

Post by echelon »

matheusmoreira wrote: post #345
atty wrote: post #266

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 (or even worse, video). We’re not at that point yet, but these things are getting better fast, so it’s only a matter of time. The problem is that there’s no way to mitigate those risks except to keep the model behind an inference-only API, or not release it at all - as soon a…

These "risks" only exist because people somehow came to believe they actually have control over their "likeness". They don't. They never had. It's an illusion. The only risk here is exposing this for the lie it is. Maybe once it's trivially easy to copy someone else's "likeness", society will finally be able to accept it and evolve past it.

Well said.

Biological twins have never had control over this. If one twin wants to be a porn star, there's nothing the other can do.

Edit: one could imagine a future dystopia where clones are created to bypass "identity IP".

Post reply on HN