Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

No elephants: Breakthroughs in image generation

oneusefulthing.org

Post reply on HN
Print view View original

221–230 of 373 posts

Re: No elephants: Breakthroughs in image generation

#221

Post by HelloMcFly »

munksbeer wrote: post #192
wincy wrote: post #158

Earlier quoted context omitted.

I was generating pictures to use for a little game I made with my six and ten year old kids. They were so excited to see us go from idea to execution so quickly, they were laughing and we had a ton of fun. The only thing that disappointed me was I got throttled. We’d need to pay for API image gen to get it even faster. I made a logo for an internal product that wouldn’t have had a logo otherwise at our company. I als…

Thanks. Which service do you use please? I'm wanting to try a paid service, just want to know which ones people recommend.

I've used Midjourney and chatGPT. Midjourney is better for rapid iteration, cycling through options faster, and to a large extent getting "weirder". It's easier to tweak using parameters.

ChatGPT is far, far superior (especially now) when you want something more specific that you've already imagined. But it's slower, and unlike Midjourney you don't get four versions to choose to build and iterate on, you get a single image that takes longer to load.

Re: No elephants: Breakthroughs in image generation

#222

Post by GaggiX »

satvikpendem wrote: post #218
__loam wrote: post #171

Earlier quoted context omitted.

Most of the professionals in this industry actively despise this technology.

That's true, but it will likely be newer studios with younger professionals who are going to be using it, much as Miyazaki doesn't like CGI either yet it's widely used now in anime. The young drive the advances while the older eschew them, that's generally how human progress has been.

I don't even think it will be newer studios and that's it, probably well known studios will adopt it after a production ready solution is presented and battle tested, I doubt many will complain about not making in between frames. Probably newer studios with a smaller budget will just test it first.

Re: No elephants: Breakthroughs in image generation

#223

Post by yencabulator »

DeathArrow wrote: post #41
probably_wrong wrote: post #12

> Is it okay to reproduce the hard-won style of other artists using AI? Who owns the resulting art? Who profits from it? Which artists are in the training data for AI, and what is the legal and ethical status of using copyrighted work for training? These were important questions before multimodal AI, but now developing answers to them is increasingly urgent. I have to disagree with the conclusion. This was an importa…

>> it's unfair for artists to have their works sucked up What framework can we use to decide if something is fair or not? Style is not something that should be copyrighted. I can pain in the style of X painter, I can write in the style of Y writer, I can compose music in the style of Z composer. Everything has a style. Dressing yourself has a style. Speaking has a style. Even writing mathematical proofs can have a st…

It's already gone quite a bit further than "style". https://www.404media.co/listen-to-the-ai-generated-ripoff-so...

Re: No elephants: Breakthroughs in image generation

#224

Post by m4thfr34k »

I am very impressed with the current image generators out there, 4o / Leonardo / etc., but I cannot wait until they include some step to actually "check their work". Ask it to produce a watch with the time of 6:37. It fails every time, because almost all watch photos out there are set to a specific time, and seems like something an initial "did I do this right" check could confirm. The time example is trivial but a general "does this output actually make sense considering what the user asked" checked would be tremendously valuable.

Re: No elephants: Breakthroughs in image generation

#225

Post by __loam »

ldoughty wrote: post #196
__loam wrote: post #170

Earlier quoted context omitted.

If you use this technology, you're actively harming creative labor.

Can you elaborate how there's no possible way to use this technology without actively harming artists? If a classroom of 14 year olds are making a game in their computer science class, and they use AI to make placeholder images... Was a real artist harmed? The teacher certainly cant afford to pay artists to provide content for all the students games, and most students can't afford to hire an artist either.. they perh…

[flagged]

Re: No elephants: Breakthroughs in image generation

#226

Post by genewitch »

danielbln wrote: post #182
genewitch wrote: post #174

Earlier quoted context omitted.

All the stable diffusion software I've used names the files after some form of the prompt, and probably because SD weights the first tokens higher than the last tokens, probably as a side effect of the way the CLIP/BLIP works. I doubt any of these companies have rolled their own interface to stable diffusion / transformers. It's copy and paste from huggingface all the way down. I'm still waiting for a confirmed Diffu…

Auto1111 and co are using the prompt in the filename because it's convenient, not due to some inherent CLIP mechanism. If you think that companies like OpenAI (for all the criticisms they deserve) don't use their own inference harness and image models I have a bridge to sell to you.

i give less weight to your opinion than my own. I'm not sure how you misunderstood what i said about clip/blip, as well. I was replying to a comment about "populating the front end with the filename" - the first tokens are weighted higher in the resulting image than the later tokens. And therefore, if you prompt correctly, the filenames will be a very accurate description of the image. Especially danbooru style, you can just split on space and use them as tags, for all practical purposes.

I guess the "convenience" just happened to get ported over from "Auto1111", or it's a coincidence, or

Re: No elephants: Breakthroughs in image generation

#227

Post by wongarsu »

YurgenJurgensen wrote: post #16
Retr0id wrote: post #5

Earlier quoted context omitted.

I've never used a stock photo site before, so I suppose it's no surprise I have no real use for "generate any image on demand".

Their main application appears to be taking blog posts and internal memos and making them three times longer and use ten times the bandwidth to convey no more information. So exactly the application AI is ‘good’ at.

If anything stock image websites are even worse at this then AI. With AI you come up with an image idea, then try to make the AI produce something close to it. With stock images you come up with an image idea, then hope some photographer had a similar idea and uploaded it to a stock website.

Re: No elephants: Breakthroughs in image generation

#229

Post by fc417fc802 »

__loam wrote: post #225
ldoughty wrote: post #196

Earlier quoted context omitted.

Can you elaborate how there's no possible way to use this technology without actively harming artists? If a classroom of 14 year olds are making a game in their computer science class, and they use AI to make placeholder images... Was a real artist harmed? The teacher certainly cant afford to pay artists to provide content for all the students games, and most students can't afford to hire an artist either.. they perh…

[flagged]

The straw man is yours. No claim of entitlement was made. A scenario was provided that appears to refute your unconditional assertion that using this technology actively harms creative labor.

You've presented all sorts of wild assumptions and generalizations about the people who don't share your vehement opposition to the use of this technology. I don't think it's the person you're responding to with the implicit bias.

You've conflated theft with piracy (all too common) and assumed a priori that training a model on publicly available data constitutes such. Do you really expect people to blindly adopt your ideological views if you just state them forcefully enough?

> If using AI is okay for the creative labor, why shouldn't the students also use it for the programming too?

They absolutely should! At least provided it does the job well enough.

Unless they are taking a class whose point is to learn to program yourself (ie the game is just a means to an end). Similar to how you might be forbidden to use certain advanced calculator features in a math class. If you enroll in an art class and then just prompt GPT that likely defeats the purpose.

Re: No elephants: Breakthroughs in image generation

#230

Post by fc417fc802 »

satvikpendem wrote: post #217
becquerel wrote: post #211

Earlier quoted context omitted.

All labor is bad.

Interesting philosophy, what is this predicated on? Do you mean that people should not have to work for a living, ie labor versus play?

Is that not self evident? When people engage in labor for the task itself (as opposed to a heavily abstracted version of not wanting to starve) we generally refer to that as a hobby.

So stating that people shouldn't need to worry about starving (metaphorically or otherwise) would be roughly equivalent.

Post reply on HN