Live data from Hacker News

Roc Camera

roc.camera

121–130 of 491 posts

Re: Roc Camera

#121

It's wild that it's already come to this: The camera itself becomes more important as the instrument to provide zero-trust proof. This is a brilliant solution to one of the most critical emergent problems. I can see a world where no digital image can be trusted if it doesn't come with a hash. There is also something called "film" which might be a retro answer to this problem.

Until people start to make AI images, print them out and then make a "real" photo of the printout to get the hash.

Re: Roc Camera

#122
lol, faq is funny

how long does the batter last

> Currently, the battery will last estimated 2~3 hours on constant use on a full charge. It can last much longer if it is off.

Re: Roc Camera

#123
post #114
post #108

Earlier quoted context omitted.

From these pics it actually looks like a whole PI4 board is used https://farcaster.xyz/faust

Interesting. I'm curious why they would do that.

All the stuff is off the shelf. Makes it way easier to develop. There is no reason to actually use RPi, compute module or not, as a base camera board (talking from experience) other than it is super easy to start with.

Re: Roc Camera

#124
post #9

This is rather expensive for what looks like a home 3D printed toy with some cute software. Other than that it's a 16MP Sony CMOS, I'd expect a pretty noisy picture... How do I get my photos off the camera? Coming soon. We're working on export functionality to get your photos off the camera. It would be more interesting if the software was open source.

> This is rather expensive for what looks like a home 3D printed toy with some cute software. This attitude really rubs me the wrong way, especially on a site called Hacker News. I think we absolutely should be supporting projects like this (if you think they're worth supporting), else all we're left with is giant corporation monoculture. Hardware startups are incredibly difficult, and by their nature new hardware pr…

You literally can't even export the photos...

Re: Roc Camera

#125
post #9

This is rather expensive for what looks like a home 3D printed toy with some cute software. Other than that it's a 16MP Sony CMOS, I'd expect a pretty noisy picture... How do I get my photos off the camera? Coming soon. We're working on export functionality to get your photos off the camera. It would be more interesting if the software was open source.

> This is rather expensive for what looks like a home 3D printed toy with some cute software. This attitude really rubs me the wrong way, especially on a site called Hacker News. I think we absolutely should be supporting projects like this (if you think they're worth supporting), else all we're left with is giant corporation monoculture. Hardware startups are incredibly difficult, and by their nature new hardware pr…

It would be cool if this was open source because looking at the pictured this is all off the shelf hardware. I am guessing only bespoke thing here is the stl for the case

Re: Roc Camera

#126
I remember reading in some Qualcomm Snapdragon document that Qualcomm integrated some image authenticity method. Not sure if this ever landed in an end-product?

Re: Roc Camera

#127
post #9

This is rather expensive for what looks like a home 3D printed toy with some cute software. Other than that it's a 16MP Sony CMOS, I'd expect a pretty noisy picture... How do I get my photos off the camera? Coming soon. We're working on export functionality to get your photos off the camera. It would be more interesting if the software was open source.

It wouldn't work at all as open source since you could just modify the source to sign your AI generated pictures.

Okay. What prevents you from printing out a AI generated picture and taking a photo of that with the camera?

Re: Roc Camera

#128
post #85

I don't understand how the "proof" part works, like, what part of the input to the "proof generation" algorithm is so inherently tied to the real world that one cannot feed it "fake" data ?

My understanding is it can't. The proof is "this photo was taken with this real camera and is unmodified". There's no way to know if the photo subject is another image generated by AI, or a painting made by a human etc.

^^This so much.

I remember when snapchat were touting "send picture that delete within timeframes set by you!" and all that would happen is you'd turn to your friend and have them take a picture of your phone.

In the above case, the outcome was messy. But with some effort, people could make reasonable quality "certified" pictures of damn near anything by taking a picture of a picture. Then there is the more technical approach of cracking a system physically in your hands so you can sign whatever you want anyway...

I think the aim should be less on the camera hardware attestation and more on the user. "It is signed with their key! They take responsibility for it!"

But then we need:

1. fully active and scaled public/private key encryption for all users for whatever they want to do

2. a world where people are held responsible for their actions...

I'm not sure which is more unrealistic.

Re: Roc Camera

#129
post #95

I’m not seeing what this is product is trying to solve? A zero knowledge proof to say it isn’t AI ? I think you could do this with a disposable camera or Polaroids and a photo scanner that makes the zero knowledge proofs .

A different thread mentions "what if you take a photo of an AI photo with the Roc camera?" - I still think that would be hard due to perspective, lighting, various other artifacts. Scanning an image would be much easier to dupe though - scanners are basically controlled perspective/lighting environments so scanning an actual polaroid vs an ai generated polaroid printed on photo paper would be pretty indistinguishable…

Maybe I am just naive, but I don't see why taking a photo of a screen, projection, or print out would be hard. Wouldn't it just need even lighting and tripod?

Adding something like a LIDAR and somehow baking that data into the meta data could be fun

Re: Roc Camera

#130
post #85

I don't understand how the "proof" part works, like, what part of the input to the "proof generation" algorithm is so inherently tied to the real world that one cannot feed it "fake" data ?

My understanding is it can't. The proof is "this photo was taken with this real camera and is unmodified". There's no way to know if the photo subject is another image generated by AI, or a painting made by a human etc.

Perhaps if it measured depth it could detect "flat surface" and flag that in the recorded data. Cameras already "know" what is near or far simply by focusing.
Post reply on HN