Live data from Hacker News

Drag Your GAN: Interactive Point-Based Manipulation of Images

vcai.mpi-inf.mpg.de

31–40 of 90 posts

Re: Drag Your GAN: Interactive Point-Based Manipulation of Images

#31
post #22

Has anyone built the "online photoshop that incorporates all of the latest AI image editing tools asap and sells access as a premium subscription with lots of GPU access for smooth editing" business yet? I'd be curious to know.

Adobe Firefly already did it https://www.adobe.com/sensei/generative-ai/firefly.html

Re: Drag Your GAN: Interactive Point-Based Manipulation of Images

#32
post #27

Earlier quoted context omitted.

Please try to make an original movie that meets the standard of, say, The Godfather, with AI.

I understand your point but I think it would be easier with AI than without. Many movies are not made to the standard of The Godfather because they don't sell like MCU Movie #53 and if you include more humans in the creation you're more likely to run into the current system's restrictions. Making a movie as beloved as the Godfather would still be challenging of course.

Of course it would be easier. I agree. I just take issue with the "why humans" thing because if anything, the recent advancements highlight just how big the human element really is.

Can you imitate a Bach prelude? Sure. And only people who aren't actually familiar with his music would be impressed.

Much of AI approaching "human performance", is it approaching the lowest bar. There's a Wittgenstein thing going on here. That an LLM can ace the LSAT or GMAT is mostly an indictment of those tests.

A little off topic.

Re: Drag Your GAN: Interactive Point-Based Manipulation of Images

#34
post #22

Has anyone built the "online photoshop that incorporates all of the latest AI image editing tools asap and sells access as a premium subscription with lots of GPU access for smooth editing" business yet? I'd be curious to know.

Adobe Firefly already did it https://www.adobe.com/sensei/generative-ai/firefly.html

It's trained on their stock art and under-performs Stable Diffusion and Midjourney.

It's really poor, comparatively.

Re: Drag Your GAN: Interactive Point-Based Manipulation of Images

#35
post #22

Has anyone built the "online photoshop that incorporates all of the latest AI image editing tools asap and sells access as a premium subscription with lots of GPU access for smooth editing" business yet? I'd be curious to know.

Playground AI ( https://playgroundai.com/ ) does a lot of this.

There are a million of these. It's a super crowded space.

https://civitai.com/

https://lexica.art/

https://openart.ai/

(Many more)

Re: Drag Your GAN: Interactive Point-Based Manipulation of Images

#36
post #22

Has anyone built the "online photoshop that incorporates all of the latest AI image editing tools asap and sells access as a premium subscription with lots of GPU access for smooth editing" business yet? I'd be curious to know.

Adobe Firefly already did it https://www.adobe.com/sensei/generative-ai/firefly.html

Adobe AI is a crap compared to Midjourney

Re: Drag Your GAN: Interactive Point-Based Manipulation of Images

#38
post #32

Earlier quoted context omitted.

I understand your point but I think it would be easier with AI than without. Many movies are not made to the standard of The Godfather because they don't sell like MCU Movie #53 and if you include more humans in the creation you're more likely to run into the current system's restrictions. Making a movie as beloved as the Godfather would still be challenging of course.

Of course it would be easier. I agree. I just take issue with the "why humans" thing because if anything, the recent advancements highlight just how big the human element really is. Can you imitate a Bach prelude? Sure. And only people who aren't actually familiar with his music would be impressed. Much of AI approaching "human performance", is it approaching the lowest bar. There's a Wittgenstein thing going on here…

>That an LLM can ace the LSAT or GMAT is mostly an indictment of those tests.

These kind of comments are always the funniest. You can just tell the person who makes them has never looked at those tests nevermind attempted them.

Re: Drag Your GAN: Interactive Point-Based Manipulation of Images

#39
post #34

Earlier quoted context omitted.

Adobe Firefly already did it https://www.adobe.com/sensei/generative-ai/firefly.html

It's trained on their stock art and under-performs Stable Diffusion and Midjourney. It's really poor, comparatively.

It makes way fewer visual mistakes (like wrong number of limbs) than Stable Diffusion, or even Bing Dall-E ~3. The latter is still the best at understanding your prompt though.

Re: Drag Your GAN: Interactive Point-Based Manipulation of Images

#40
post #22

Has anyone built the "online photoshop that incorporates all of the latest AI image editing tools asap and sells access as a premium subscription with lots of GPU access for smooth editing" business yet? I'd be curious to know.

I think less effort has gone into image editing compared to image generation so far. That said, we're building some photo realistic image editing tools at https://www.faceshape.com, focused on face editing for now. Current models don't perform as well, but next generation currently under training will.

I'm always curious to know what kind of AI image editing people are interested in, can you share what kind of edits you'd like to do? There's the usual edits like background removal or object removal, but those are more general tools that are getting incorporated into lots of apps natively (say Google Photos).

Post reply on HN