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

NeRF: An eventual successor for deepfakes?

metaphysic.ai

Post reply on HN
Print view View original

141–150 of 169 posts

Re: NeRF: An eventual successor for deepfakes?

#141

Post by corrral »

cwkoss wrote: post #109
narrator wrote: post #64

Earlier quoted context omitted.

Surprise Plot Twist: Maybe we're already living in a post-truth society and you are still sure you know what the truth is. How would you even know that what you were ferociously defending as the truth wasn't a lie? What makes you think you're not smart enough to not fall for lies? Largely, I think most people's means of finding the truth is just to take a vote of the information sources they find credible and go with…

Disinformation is very common in traditional news media. This technology just democratizes this tool and allows anyone to engage in it. There will probably be a net increase in disinformation, but citizens will likely also get better at being skeptical of currently unquestioned modes of disinformation.

> There will probably be a net increase in disinformation, but citizens will likely also get better at being skeptical of currently unquestioned modes of disinformation.

Russia seems to be farther along this path than we are and every account I've read of their experience of disinfo isn't that they got better at seeking the truth, but instead just assume everything's a lie & nothing's trustworthy, and disappear into apathy.

Re: NeRF: An eventual successor for deepfakes?

#142

Post by flor1s »

echelon wrote: post #66

What's the difference in NeRF from a classical photogrammetry pointcloud workflow? It seems like the representation and outputs are identical. Why would you prefer NeRF to photogrammetry? Or vice versa?

Neural Radiance Fields are a technique from the neural rendering research field, while photogrammetry is a research field on its own. However these are just turf wars and in practice there is a lot of overlap between both fields.

For example, most NeRF implementations recommend the use of COLMAP (traditionally a photogrammetry tool) to obtain camera positions/rotations that are used alongside their images. So this multi-view stereo step is shared between both NeRF (except a few research works that also optimize for camera positions/rotations through a neural network) and photogrammetry.

After the multi-view stereo step in NeRF you train a neural renderer, while in photogrammetry you would run a multi-view geometry step/package that uses more traditional optimization algorithms.

The expected output of both techniques is slightly different. NeRF produces renderings and can optionally export a mesh (using the marching cubes algorithm). Photogrammetry produces meshes and in the process might render the scene for editting purposes.

Re: NeRF: An eventual successor for deepfakes?

#143

Post by random-human »

tomgraham wrote: post #33
SantalBlush wrote: post #30

Earlier quoted context omitted.

Creating yet another product to monetize is not a solution, it's just more of the problem. It incentivizes a perpetual arms race between fabrication and verification at the cost of everyday users. No thanks.

It is free. Protecting individuals' right is more important that making money!

Free but collecting and storing peoples biometric data on your servers (per the FAQ). How do I know it's not a clearview ai clone but with easier data gathering? What is that saying about what the real product is if something is free?

Re: NeRF: An eventual successor for deepfakes?

#144

Post by joeld42 »

Not sure about NeRF's impact on deepfakes specifically, but they are looking like a seismic shift in the graphics world. Even without the "neural" part, as a scene representation they are already having a huge impact, and will have implications everywhere. It feels almost like a new primitive is being discovered.

Maybe I'm just getting caught up in all the hype, but I can't remember the last time I saw a topic with this much momentum.

Re: NeRF: An eventual successor for deepfakes?

#145

Post by mazlix »

RubyRidgeRandy wrote: post #4

Something I've wondered lately is what will life be like in a post-truth society? we already see examples of this now where a large number of people get their news from fake memes on facebook. There are huge swathes of people who live in their own make-believe world, like those believing wholeheartedly that the 2020 election was stolen. What will life be like when you can't trust any video or interview you see becaus…

I don't know exactly what you mean by "post-truth", but I think for the vast majority of human history we've lived in a time where it was rare to have easy/cheap to create and hard to fake records of what happened (photography, videos, audio recordings).

In terms of a single photograph though that's been very easy to fake in compelling ways for at least the past 30 years look at lochness and bigfoot photos.

Video's harder, but still pretty much all of this is not horribly difficult to fake. We've long been able to edit things out and mix audio to make a compelling video of a cartel leader meeting with them, no need for deep fakes.

Re: NeRF: An eventual successor for deepfakes?

#146

Post by Davonbon »

Each day technology becomes scarier, even if technology exists to detect deepfakes and all, and there are infinite methods to tell reality appart from digital images, it doesn't matter, because most people don't know about that, and it's not their fault, not everyone has to get into tech that deeply. However, technology being used to lie to people has always been a thing, think back to photoshop, you can also create fake pictures in there, doesn't matter if they're a hundred percent realistic, many people are still gonna believe it's real, videos of ARMA 3 being used as war footage has been a thing for a long time, even in the ukraine war it happened again, so it's not news, the scary thing is how easily accessible this technology is becoming, and I'm thinking a lot of new laws on usage of technology that will have to be created because of this.

Re: NeRF: An eventual successor for deepfakes?

#147

Post by mazlix »

jl6 wrote: post #124
GuB-42 wrote: post #67

Earlier quoted context omitted.

There has never been a truth society. This tech will certainly be used to frame someone for a crime, like I am sure Photoshop was used in such a way, and thousands of other techniques. And modern technology offers counters. It is an arms race but because of the sheer amount of data that is collected, I think that truth is more accessible than ever. The more data you have, the harder it is to fake and keep consistent.

I don’t know, it seems like the existence of widespread, easy photo/video/audio faking technology could be a really strong argument for dismissing any purported photo/video/audio evidence. Wouldn’t it be funny if deepfakes destroyed the blackmail industry?

I don't much about the blackmail industry, but I would imagine the reverse would be more likely.

I would think blackmail works best when being used on things which are actually true. If someone wanted to blackmail me by sharing a fake photo of my cheating on my spouse I wouldn't cave for anything because I know I haven't so I feel confident I have enough other evidence to fight that claim. My location data, who is the "other woman", etc.?

On the flip side if I had commit some crime and someone made a deepfake of me doing that crime but they didn't actually have a legitimate photo, I might cave at that point since I would think it's a genuine photo and presumably they know even more about the incident.

Re: NeRF: An eventual successor for deepfakes?

#148

Post by l33t2328 »

ketralnis wrote: post #97
mc32 wrote: post #93

Earlier quoted context omitted.

If the accused admit to something it's more believable than the alternative (that they were forced into false admission). So in this case if the Ukrainian government admit to making things up then I would think it's believable that they made something up for the sake of propaganda. We can also check more independent sources --read Japanese news, or Indian news sources, etc.

We don't know that the accused admitted to anything. We know that the Washington Post says that they did. The world becomes very solipsistic when you lose trust in reporting.

Well…yeah. If you don’t trust anything you can’t know very much.

Re: NeRF: An eventual successor for deepfakes?

#149

Post by mazlix »

idontwantthis wrote: post #135
fleetwoodsnack wrote: post #24

Earlier quoted context omitted.

Different instruments can be used to capture different segments of the population. You’re right there are gullible people who are more likely to believe things with limited or no evidence. But it isn’t necessarily about the most impressionable people, nor is it about installing sincerely held beliefs. Instead, what may be a cause for concern is simply the installation of doubt in an otherwise reasonable person becaus…

Think about how few people believe in Bigfoot when video, photographs, footprints, eye witness testimony all exist. Think about how many people believe in Jesus without any of that physical evidence. If anything, the physical evidence turns most people off. And I'd argue that most Bigfooters don't even believe in the physical evidence, but use it as a tool to hopelessly attempt to convince other people to believe in…

Completely agree.

For some reason many people react to learning about deep fakes' potential with huge concern as if photos and video used to be infallible and it's suddenly being overturned when this really hasn't been the case.

Re: NeRF: An eventual successor for deepfakes?

#150

Post by dools »

fxtentacle wrote: post #16
RubyRidgeRandy wrote: post #4

Something I've wondered lately is what will life be like in a post-truth society? we already see examples of this now where a large number of people get their news from fake memes on facebook. There are huge swathes of people who live in their own make-believe world, like those believing wholeheartedly that the 2020 election was stolen. What will life be like when you can't trust any video or interview you see becaus…

I believe we'll go back to trusting local experts that you can meet in person to confirm that they are not a bot. Because anything online will be known to be untrustworthy. Most blogs, chat groups and social media posts will be spam bots. And it'll be impossible for the average person to tell the difference between chatting with a bot and chatting with a human. But humans crave social connections and intimate physica…

So you’re predicting a return to Stone Age tribalism
Post reply on HN