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

Adobe Firefly: AI Art Generator

adobe.com

Post reply on HN
Print view View original

271–280 of 496 posts

Re: Adobe Firefly: AI Art Generator

#271

Post by brookst »

usrusr wrote: post #106
hn_throwaway_99 wrote: post #57

Earlier quoted context omitted.

I think you're missing a fundamental reason: adding AI functionality into products is simply easier. That is, these companies are largely not doing the hard part, which is creating and training these models in the first place. The examples you gave of of cloud and iPhone have both huge capital barriers to entry, and in iPhones case other phone companies just didn't have the unique design talent combination of Jobs an…

> That is, these companies are largely not doing the hard part, Are they the hard parts though? The short time it took from the first waves of public excitment around DALL-E to stable diffusion being the well established baseline looks more like the class quantity of problems that can be reliably solved by shoving enough resources at it. What I consider hard problems are those full of non-obvious dead ends are where…

The hard part is building a product customers want, delivering it at scale, and iterating on product value and revenue.

The rest is just technology.

Re: Adobe Firefly: AI Art Generator

#272

Post by sebzim4500 »

ClumsyPilot wrote: post #261
grondo4 wrote: post #75

Earlier quoted context omitted.

Yes, AI should be trained on every piece of information possible. Am I allowed to become a better programmer by looking at private, (illegally leaked) closed-source, proprietary code?

> Am I allowed to become a better programmer by looking at private code? Your argument is based on the idea that you and AI should have the same rights? I do not see how this works unless AI going to be entitled to minimum wage and paid leave? Otherwise it is just a money grab

He's not saying that he and the AI have the same rights, rather that he and the person running the AI have the same rights.

Re: Adobe Firefly: AI Art Generator

#273

Post by golergka »

int_19h wrote: post #270
lmarcos wrote: post #225

Earlier quoted context omitted.

That artists will be able to do more not necessarily mean that fewer of them will ne needed. I bet the opposite actually. Companies will want to produce more (with more resources) not the same (with less resources).

But do the companies actually need more art? I guess a better way to rephrase that: would it make them more profitable to produce more art? Or to keep producing the same amount but paying fewer people to do so?

When something becomes much cheaper than it was before, people tend to find much more uses for it. In game development, for example, amount of money that could be spent on art almost always amount and quality of content; if art becomes 10 times cheaper, a typical indie game will have 2 times more different pieces of content with 5 times variants of each.

Re: Adobe Firefly: AI Art Generator

#274

Post by ClumsyPilot »

O__________O wrote: post #214
superbatfish wrote: post #109

Earlier quoted context omitted.

There are at least two potential issues pertaining to copyright law here, and it's not clear which one you're asking about. That's why the responses you're getting here seem to be answering different questions. 1. Are the AI systems violating the copyright protections of the images they were trained on? If so, are users of such AI systems also in violation of those copyrights when they create works derived from those…

As I am sure you’re aware, already posted response US Copyright’s ruling related to authorship here, so will not be repeating myself: https://news.ycombinator.com/item?id=35247377 Will say that post you linked to also states, “17 U.S.C. 101 (definition of “compilation”). In the case of a compilation including AI-generated material, the computer-generated material will not be protected outside of the compilation.” — t…

> a prompt is somehow a recipe and not a novel expression that merits copyright

Is it not? Does typing in 'cat' in SD, as millions of people will, count as novel expression?

Re: Adobe Firefly: AI Art Generator

#275

Post by sebzim4500 »

andybak wrote: post #67
l33tman wrote: post #15

Earlier quoted context omitted.

I would consider both Stable Diffusion and MidJourney to be startups and both are better by far than the established companies like OpenAI, and there are dozens of LLMs soon catching up with GPT3/4. As you say, it will be a very interesting next 6-12 months..

Dall-E is still ahead of the competition for some tasks. Midjourney has a very polished look but it lacks the depth of understanding that Dall-E can manage. I regularly hit prompts that I need to jump into Dall-E for.

For playing around, sure. For serious work though you probably need control-net in some way, otherwise you end up with a bunch of images which are great on their own but make no sense together.

Re: Adobe Firefly: AI Art Generator

#276

Post by int_19h »

brucethemoose2 wrote: post #34
oidar wrote: post #26

The high compute costs for training and legacy software foothold seems that it may encourage rent seeking in the big tech companies. I hope the huge costs and compute demands won't only enrich the big tech companies that can afford to run the models on heavy duty hardware. If it does, it will potentially lead to the rich companies, individuals, and persons greatly outpacing their peers.

These models can be run locally, but I see far too many businesses and bloggers resign themselves to renting OpenAI's cloud instead of tinkering with LLaMA/Alpaca, the community Stable Diffusion finetunes and such. Adobe does have a legal edge here, which is interesting and perhaps actually worth a subscription if needed.

The problem is that GPT-3.5 is so cheap, and the results that you get out of it are still quite a bit better than LLaMA/Alpaca. There just doesn't seem to be any solid economic reason to run it locally except to keep inputs and/or outputs private.

Re: Adobe Firefly: AI Art Generator

#277

Post by lelandfe »

spoiler wrote: post #267
ryanjshaw wrote: post #237

Earlier quoted context omitted.

I'm curious, do you hold the same beliefs about text? Do you think ChatGPT should not be allowed to read books and join ideas across them without paying the original authors for their contribution to the thought?

I do! If they aren't in some way public domain, then the authors should have a say, or be if the work is purchased. I have a bit of cognitive dissonance on the subject of blog posts or articles in general, since those are kinda public domain? But I still think it should be opt in/out-able. I realise I'm also a bit of a hypocrite since I've enjoyed playing with these AI tools myself, and I realise they'd be nowhere as…

IANAL: Authorship is protected in the US by default https://www.copyright.gov/engage/writers

In order for blog posts (or other written works) to be in the public domain, authors must explicitly waive those rights. But, not that it needs saying, copyright's applicability in training data is basically the entire subject of debate right now. https://creativecommons.org/2023/02/17/fair-use-training-gen...

Re: Adobe Firefly: AI Art Generator

#278

Post by aschearer »

mesh wrote: post #258
aschearer wrote: post #191

While this is very exciting it's worth pointing out most of the page falls under "WHAT WE’RE EXPLORING." That is to say, these are from the marketing department and it's impossible to know whether they are strictly aspirational or just around the corner. From the page: > Looking forward, Firefly has the potential to do much, much more. Explore the possibilities below ... We plan to build ... We’re exploring the poten…

Yeah, we are being a little more open, a little earlier, in part because this space is so new, and we really want feedback / direction from the community. Currently in the beta, there is text to image, and text effects, and will have vector recoloring in the next couple of weeks.

Keep up the good work. I raise my point solely because people are comparing these trailers against Stable Diffusion and the like, when in reality the examples are artists renditions. There's no point in comparing.

I hope someday y'all get to the point where we can change a video's season, time of day, etc. and have things work seamlessly. That would be quite incredible!

Re: Adobe Firefly: AI Art Generator

#279

Post by usrusr »

brookst wrote: post #271
usrusr wrote: post #106

Earlier quoted context omitted.

> That is, these companies are largely not doing the hard part, Are they the hard parts though? The short time it took from the first waves of public excitment around DALL-E to stable diffusion being the well established baseline looks more like the class quantity of problems that can be reliably solved by shoving enough resources at it. What I consider hard problems are those full of non-obvious dead ends are where…

The hard part is building a product customers want, delivering it at scale, and iterating on product value and revenue. The rest is just technology.

And that had been true enough even before the "quantity matters!" of ML entered the stage.

Re: Adobe Firefly: AI Art Generator

#280

Post by theFletch »

lelandfe wrote: post #50

From that page's FAQs: > trained on a dataset of Adobe Stock, along with openly licensed work and public domain content where copyright has expired > We do not train on any Creative Cloud subscribers’ personal content. For Adobe Stock contributors, the content is part of the Firefly training dataset, in accordance with Stock Contributor license agreements. The first model did not train on Behance. Not sure what "firs…

> trained on a dataset of Adobe Stock, along with openly licensed work and public domain content where copyright has expired

As someone who has contributed stock to Adobe Stock I'm not sure how I feel about this. I'm sure they have language in their TOS that covers this, but I'm guessing all contributors will see nothing out of this. Fine if this is free forever, but this is Adobe.

Post reply on HN