101–110 of 718 posts
Maybe I'm an obscure case, but I'm just not sure what I'd use an image generation model for. For people that use them (regularly or not), what do you use them for?
but concept art, try-it-on for clothes or paint, stock art, etc
Earlier quoted context omitted.
It solves some problems! For example, if you want to run a camgirl website based on AI models and want to also prove that you're not exploiting real people
Your use case doesn't even make sense. What customers are clamoring for that feature? I doubt any paying customer in the market for (that product) cares. If the law cares, the law has tools to inquire. All of this is trivially easy to circumvent ceremony. Google is doing this to deflect litigation and to preserve their brand in the face of negative press. They'll do this (1) as long as they're the market leader, (2)…
How can they distinguish from real people exploited to AI models autogenerating everything?
I mean right now this is possible, largely because a lot of the AI videos have shortcomings. But imagine in 5 years from now on ...
Maybe I'm an obscure case, but I'm just not sure what I'd use an image generation model for. For people that use them (regularly or not), what do you use them for?
At the end of the day, a tool is a tool, and the computer had the same effect on the creative industry when people started using them in place of illustrating by hand, typesetting by hand, etc. I don't want my personal bias to get in the way too much, but every nail that AI hammers into the creative industry's coffin is hard to witness.
The interesting tidbit here is SynthID. While a good first step, it doesn't solve the problem of AI generated content NOT having any kind of watermark. So we can prove that something WITH the ID is AI generated but we can't prove that something without one ISN'T AI generated. Like it would be nice if all photo and video generated by the big players would have some kind of standardized identifier on them - but now you…
Labelling open source models as "grey market" is a heck of a presumption
Google needs to pace themselves. AI studio, Antigravity, Banana, Banana Pro, Grape Ultra, Gemini 3, etc. This information overload don't do them any good whatsoever.
Google needs to pace themselves. AI studio, Antigravity, Banana, Banana Pro, Grape Ultra, Gemini 3, etc. This information overload don't do them any good whatsoever.
Earlier quoted context omitted.
In 25 years we'll reminisce on the times when we could find a human artist who wouldn't impose Google's or OpenAI's rules on their output.
the open-source models will catch up, 100%
ChatGPT's imagegen has been released for half a year but there isn't anything remotely similar to it in the open weight realm.
I've tried to repaint the exterior of my house. More than 20 times with very detailed prompts. I even tried to optimize it with Claude. No matter what, every time it added one, two or three extra windows to the same wall.
I tried this in AI studio just now with nano banana. Results: https://imgur.com/a/9II0Aip The white house was the original (random photo from Google). The prompt was "What paint color would look nice? Paint the house."
Careful with that kind of thing.
Here, it mostly poisons your test, because that exact photo probably exists in the underlying training data and the trained network will be more or less optimized on working with it. It's really the same consideration you'd want to make when testing classifiers or other ML techs 10 years ago.
Most people taking to a task like this will be using an original photo -- missing entirely from any training date, poorly framed, unevenly lit, etc -- and you need to be careful to capture as much of that as possible when trying to evaluate how a model will work in that kind of use case.
The failure and stress points for AI tools are generally kind of alien and unfamiliar because the way they operate is totally different than the way a human operates, and if you're not especially attentive to their weird failure shapes and biases when you want to test them, or you'll easily get false positives (and false negatives) that lead you to misleading conclusions.
Earlier quoted context omitted.
Can you give us an example?
'athlete wearing a health tracker under a fitted training top' Failed to generate content: permission denied. Please try again.
If you triggered the safeguard it'll give you the typical "sorry, I can't..." LLM response.