Live data from Hacker News

Qwen-Image-3.0: Rich Content, Authentic Details, Deep Knowledge

qwen.ai

221–230 of 241 posts

Re: Qwen-Image-3.0: Rich Content, Authentic Details, Deep Knowledge

#221

Earlier quoted context omitted.

This is the way. There's definitely more dogfooding that needs to be done. And id argue that if your purpose is truly to learn or to teach, the process of describing that image will do wonders for retention.

> if your purpose is truly to learn or to teach, the process of describing that image will do wonders for retention. If you're wanting to learn then you won't have the expertise to craft a prompt that has the correct details or to spot when the model makes a mistake in its output. If you're in a teaching position then you won't have anything to learn.

Not necessarily. A reproduction task as a spaced repetition activity can help embed learning. If you are trying to write that prompt there is no reason you can't refer back to the text book to build the prompt and check the output.

Re: Qwen-Image-3.0: Rich Content, Authentic Details, Deep Knowledge

#222
It's a bit shocking to see them showing off blatantly disinformation-al examples. See the Japanese manga one. It says "原作 監修 三浦建太郎" which says its original is written by, and it is supervised the by, the famous author (who died a few years ago so who can he possibly supervise?) and "共同制作 白泉社" saying: collaborated by a (famous Japanese) publisher.

I hope these stakeholders have good enough layers.

Re: Qwen-Image-3.0: Rich Content, Authentic Details, Deep Knowledge

#223
It's hilarious how bad image generation is. I just tried this out and I gave it two actual logos, a full design language spec, and three screenshots of the actual application and it spat out three absolutely awful images, none of which had the same logo as the one I sent.

Re: Qwen-Image-3.0: Rich Content, Authentic Details, Deep Knowledge

#225
post #33

To me it feels so weird that people are trying to push these model for online shopping like "here is how this dress/shirt/pants would look on you". But these models will always make the clothes fit your body and show you in flattering light and so on. How the actual garment fits is still as elusive as before these tools

Definitely. I think the appeal to some isn't GenAI per se but image-to-image generation like "make this item look like a talented photographer shot it".

But I can't imagine that your conversion wouldn't benefit by showing actual imperfect photos of your products, especially when you run a platform with 1k new listings per month. I'm personally turned off by something that looks like 3D, especially when all the different colour variations look the exact same.

Re: Qwen-Image-3.0: Rich Content, Authentic Details, Deep Knowledge

#226
post #4

> Supports up to 4.5k token input, effortlessly generating complex layouts such as newspapers, storyboards, and exam papers. Impressive. Btw, what is currently the best model to run locally on a 16GB Vram? Is it Z-Image Turbo?

It depends on what you need, but Krea/Klein9b/Ideogram4/Z-Image are among the best right now for text2image and Qwen Edit and Klein are probably still the best at editing.

Re: Qwen-Image-3.0: Rich Content, Authentic Details, Deep Knowledge

#227

Earlier quoted context omitted.

The short-term goal of a tool like this is to sell products. The more ambitious long-term goal is to shift cultural norms, blurring the lines between advertising and reality until the question you're asking is no longer consciously asked. At least, not by average people, and not at the point of purchase. I find it easy to envision a world, maybe 50 years from now, in which the very concept of "truth in advertising" i…

>I find it easy to envision a world, maybe 50 years from now, in which the very concept of "truth in advertising" is viewed as a lost, idyllic fantasy. Something people are nostalgic for, but feel powerless to regain. Alternatively, perhaps we'll see models fine-tuned or steer-able towards accuracy that customers can themselves use to get a more honest view on what the product would look like in person. The funny thi…

> Alternatively, perhaps we'll see models fine-tuned or steer-able towards accuracy that customers can themselves use to get a more honest view on what the product would look like in person.

we live in a world where the norm is to use filters on pictures of your own face to lie about how you look, imagine thinking than people actually care about truthness

Re: Qwen-Image-3.0: Rich Content, Authentic Details, Deep Knowledge

#228

They must have trained on GPT Image 1 outputs. The yellow tint is unmistakable. https://qianwen-res.oss-accelerate.aliyuncs.com/Qwen-Image/i... https://qianwen-res.oss-accelerate.aliyuncs.com/Qwen-Image/i... https://qianwen-res.oss-accelerate.aliyuncs.com/Qwen-Image/i... https://qianwen-res.oss-accelerate.aliyuncs.com/Qwen-Image/i...

Maybe it was trained in a Mexican data center https://en.wikipedia.org/wiki/Mexican_filter

Foreign media like the BBC were actually caught applying a yellow filter to footage of China [1] [2] so maybe they trained on that lol

[1] https://www.globaltimes.cn/page/202102/1215710.shtml

[2] https://xcancel.com/thouse_opinions/status/13616465112479539...

Re: Qwen-Image-3.0: Rich Content, Authentic Details, Deep Knowledge

#229
post #57

Earlier quoted context omitted.

WTF lol who made qwen stefani's dress spiderman into the spider verse did qwen meet Peter qwen stacey porn do blake shelton and qwen stafani have children together

I've been thinking about this too much because it's so ridiculous and funny that I had to at least try wrapping my head around it. My best guess is this: If you look at past snapshots at archive.org, you notice that the meta keywords are growing like an append-only list, which means it's probably part of some messed up seo pipeline. The other clue is that it's stuffing the meta keywords which apparently only Yandex u…

"Google's webmaster tools doesn't provide distinct queries like that"

Yes it does

Anyways your theory is wrong because they don't rank for things like "ChatGPT login". It's just a keyword list from auto complete data not their own ranking data

Re: Qwen-Image-3.0: Rich Content, Authentic Details, Deep Knowledge

#230
post #33

To me it feels so weird that people are trying to push these model for online shopping like "here is how this dress/shirt/pants would look on you". But these models will always make the clothes fit your body and show you in flattering light and so on. How the actual garment fits is still as elusive as before these tools

I see the ai product ads and it's pretty funny. The product is shown but is "broken" as far as scale is concerned.

Sort of like how they tried to make gandalf big and the hobbits small in lord of the rings (which wasn't very convincing to me)

Post reply on HN