Hacker News Reader

Comment by colesantiago | Hacker News Reader

https://en.wikipedia.org/wiki/Analog_hole‎

There’s no analog hole, because they are defacing the image.

OTOH, that’s also one of the problems with it for art that is intended to be displayed to humans (they minimize this descriptively, but the full-size examples I’ve seen are bad.)

Did you read the website at all? It discusses that taking a screenshot or adding any other amount of blur, noise, etc isn't a way to circumvent this method.

https://en.wikipedia.org/wiki/Analog_hole‎

I also don't see how analog hole is relevant. This is not some novel DRM style, but a way to slightly modify the original image such that it's less useful for AI model training.

https://en.wikipedia.org/wiki/Analog_hole‎

That doesn’t make sense in the context of this tool.

Then you either don't understand the tool, how generative AI models work, or what the analog loophole is

It makes sense because the protective mechanism depends on the AI ingesting the picture as-is, with the added noise.

If the ingestion workflow alters the picture sufficiently, the protection could be lost, same as DRM qualities are lost if someone alters the data-stream by recording what is shown on the screen of the display device.

But this is an analog protection mechanism.

If your human eye can look at the post-Galze'd work and understand what it is, so will the (next) AI model

Maybe, but that is irrelavent as to whether or not the analog hole is a relavent attack.

[It should be noted that the comment 3 up the chain i was replying to got edited so my response 2 up the chain makes less sense now]

It’s hype, quite literally very impossible.

What's impossible about it?

There have even been other drm systems that have fairly succesful at attacking analog hole, e.g. https://en.wikipedia.org/wiki/Cinavia

Of course, no solution prevents everything,but preventing a specific system from using the data in a way that is hard to get around is highly possible.

Great link, in any case.

Reply on news.ycombinator.com