[HN Gopher] Adobe will charge "credits" for generative AI
       ___________________________________________________________________
        
       Adobe will charge "credits" for generative AI
        
       Author : tambourine_man
       Score  : 53 points
       Date   : 2023-09-16 21:28 UTC (1 hours ago)
        
 (HTM) web link (helpx.adobe.com)
 (TXT) w3m dump (helpx.adobe.com)
        
       | tomschwiha wrote:
       | It doesn't read too bad: Creative Cloud and Adobe Stock paid
       | users can keep taking generative AI actions, but the use of
       | generative AI features may be slower.
        
       | NikolaNovak wrote:
       | Is adobe trained on their own library of licensed images, as
       | opposed to scraping whole internets?
       | 
       | If so, even as a private individual just fooling around, I'll
       | start using it from both legal and ethical perspective as long as
       | it's reasonably equivalent to other models. And this from a
       | person who's been fairly vocal against adobe's cloud subscription
       | model ;-<. I can only imagine for anybody with a commercial need
       | it would be an immediate no brain er - they'll have an
       | established relationship, account and billing, they'll perceive
       | it as integrating in their work flow, and it'll just become
       | another part of the pipeline.
        
         | greensoap wrote:
         | It is trained on Adobe Stock and supplemented with openly
         | licensed and public domain content.
         | 
         | https://blog.adobe.com/en/publish/2023/03/21/responsible-inn...
        
         | codetrotter wrote:
         | > ethical perspective
         | 
         | If a photographer licenses a photo to Adobe Stock, they get
         | paid every time someone pays to use the photo, right?
         | 
         | But if Adobe trained their AI on photos you had licensed to
         | Adobe Stock. Do you get compensated at all?
         | 
         | If not, it's not really different from what everyone else was
         | doing in terms of ethics.
        
       | kristopolous wrote:
       | Btw, there's already an open source way to do this
       | 
       | https://github.com/AbdullahAlfaraj/Auto-Photoshop-StableDiff...
        
         | xnx wrote:
         | Nice. It looks like there are extensions for Photopea
         | (https://github.com/yankooliveira/sd-webui-photopea-embed) and
         | Gimp (https://github.com/blueturtleai/gimp-stable-diffusion)
         | too.
        
       | asimpleusecase wrote:
       | I'll not pay for it, I've tried the tool in beta and it's nothing
       | to write home about. And if they decide to get cute about other
       | creative pricing of existing features - I hope people will move
       | on.
        
       | samwillis wrote:
       | I'm convinced this will be a short lived business revenue
       | structure - paying per use of generative AI in the cloud.
       | 
       | I'm sure that in the not too distant future (a few years at most)
       | we will be happily running these on customer level hardware.
       | 
       | I do wander if companies working to develop these type of revenue
       | models truly think it's a long term structure?
        
         | coder543 wrote:
         | I would have said the same thing about CAD simulations, but
         | then Autodesk decided to move backwards and remove an existing
         | feature just so they could charge for per-simulation credits:
         | https://hackaday.com/2022/08/12/local-simulation-feature-to-...
         | 
         | Whether Adobe ever decides to let their model run locally or
         | lock it forever into the cloud is a choice they will have to
         | make. A lot of people trust Adobe products, so it's entirely
         | conceivable that some people will always choose to pay for a
         | pay-per-use generative solution from Adobe rather than try to
         | run competing solutions locally. The question is probably
         | whether it generates more revenue than negativity for Adobe. If
         | _most_ Adobe users are running their own models locally and
         | avoiding the feature, then I think Adobe will be more likely to
         | follow suit and move away from the pay-per-use cloud approach.
        
           | flangola7 wrote:
           | Long term they may offer both, with the local model consuming
           | fewer credits per request. The new GPUs all have secure
           | enclave remote attestation architecture, so Adobe would be
           | able to offer this without the risk of someone jailbreaking
           | the model and running it for free.
        
         | artursapek wrote:
         | Isn't this the revenue model for most cloud computing, like EC2
         | (compute) S3 (storage) etc?
        
         | heavyset_go wrote:
         | > _I 'm sure that in the not too distant future (a few years at
         | most) we will be happily running these on customer level
         | hardware._
         | 
         | The models themselves will be hoarded as IP. Doesn't matter if
         | they're in the cloud or on devices, they'll be licensed like
         | commercial proprietary software with the same restrictions
         | commercial software has.
        
           | whiddershins wrote:
           | Yeah but that isn't t pay - per - operation
           | 
           | I mean I guess my electricity provider gets paid per compute.
        
             | heavyset_go wrote:
             | You can write software that is pay per use, and you can do
             | the same with local models.
        
           | api wrote:
           | This will be true as long as they are crazy expensive to
           | produce. Eventually the cost of training a base model could
           | drop to the range where crowdfunding could do it.
           | 
           | Or alternately someone could make a major advance in
           | distributed training and we could all contribute cycles in a
           | distributed effort like Folding@Home. As it stands training
           | requires far too much bandwidth for synchronization and
           | moving model data around. Some approach to sharding training
           | would have to be discovered. It's an open problem area.
           | 
           | Neural networks are very parallelizable and training is
           | stochastic so my intuition is that it should be possible.
           | Even if it were less efficient than synchronous training you
           | could make up for that by harnessing 100X the compute from a
           | huge crowd.
        
             | heavyset_go wrote:
             | Platforms and rightsholders know what they're sitting on
             | now, and I worry that the datasets required to train
             | sufficient models in the future will also be hoarded as IP.
             | 
             | It's one thing to train on Common Crawl in 2023, but what
             | about when you have to shell out millions of dollars just
             | for access to data sets to train on in the future? Same
             | thing with human reinforcement. The customers for both are
             | willing to pay much more than a crowdfunding campaign
             | would.
             | 
             | Training is expensive now, but data sets can be expensive
             | in the future.
        
               | krasin wrote:
               | Collecting training data is actually a perfect for
               | crowdsourcing. Images/videos are easier than text, and
               | text is easier than high-quality text, but all are
               | doable.
        
         | yieldcrv wrote:
         | I agree, these models are going to get smaller and more
         | performant, the software to leverage your hardware is going to
         | have major improvements, the hardware is going to change to
         | prioritize processing these with more memory to specific
         | coprocessors, and the OS is going to start having these models
         | baked into them, with improved default models being a core
         | feature of upgrading each OS version
         | 
         | people are looking at an extremely limited view of "bigger
         | models on better hardware will always be in the cloud" when
         | that reality simply won't matter for most use cases
        
         | YurgenJurgensen wrote:
         | Wouldn't this require a complete reversal of course of, like,
         | the last fifteen years of computing? Basically everyone wants
         | to get out of selling products and into selling services.
        
         | api wrote:
         | I run stable diffusion XL with control nets on a laptop, albeit
         | a high end one. It's pretty decent.
         | 
         | I also run LLMs such as trains of llama2, though LLMs on
         | commodity hardware are not as "there" yet as image generators.
         | It's a decent question and answer bot and summarizer but isn't
         | GPT-4 level. I could see another iteration approaching that but
         | I'd probably need more RAM.
        
         | thorum wrote:
         | When we reach the point when average people can easily run
         | today's models on their own devices, today's models will no
         | longer be SOTA and there will still be demand to run better
         | models in the cloud.
        
           | brucethemoose2 wrote:
           | Image generation parameter count is hitting diminishing
           | returns, going by what we've seen with SD/SDXL. The tooling
           | around them is far more important.
           | 
           | However, Stable diffusion already _can_ run on mobile
           | devices. There is already a good iOS app for it (and the dev
           | is here on HN) but the problem seems to be that no one cares.
           | There are 700,000 cloud imagegen apps crowding it out,
           | because thats what 's easier and more profitable to spam
           | across the store and web.
        
             | thorum wrote:
             | > Image generation parameter count is hitting diminishing
             | returns, going by what we've seen with SD/SDXL.
             | 
             | For image quality, sure - language understanding is still
             | an issue. SDXL can generate a beautiful image, but if it
             | doesn't show _exactly_ what you asked for in the prompt, on
             | the first try, there is still room for improvement. The gap
             | between LLMs and image generators in this regard is huge.
        
             | Closi wrote:
             | Depends if the limit on parameter count is a 'real' limit,
             | or just a limit based on what current-technology models can
             | effectively use.
             | 
             | Back in the 'Google Daydream' days, Google might have found
             | that they didn't get any more image-generation performance
             | by raising the parameter count - but that's just because
             | the technology at the time couldn't effectively utilise
             | more parameters. It's impossible to know what next-gen
             | models might be able to use, but I suspect we will find
             | ways to allow the models to take advantage of even higher
             | parameter counts.
             | 
             | Stable diffusion can run on mobile devices, but it's
             | painful and image generation takes a fraction of the time
             | via cloud services.
        
         | GaggiX wrote:
         | We already run these on customer level hardware, you need a
         | good PC but still, in the future the pool of people that can do
         | is just going to be bigger.
        
         | Closi wrote:
         | I'm unconvinced that local deployment will be the preferred way
         | to run generative models anytime soon.
         | 
         | They seem like pretty much the perfect fit for cloud - burst
         | compute which would result in very low hardware utilisation if
         | ran locally.
         | 
         | Why would it be better to have a $1,500 GPU that is weak and
         | used infrequently, when you could share a big cluster of better
         | GPUs shared between a big group of people, and have it more
         | heavily utilised?
         | 
         | There is a philosophical argument about owning your own
         | hardware etc, however I think the economics and performance
         | will eventually push this to the cloud for most use-cases (most
         | people will just get better bang-for-buck in the cloud).
        
           | baz00 wrote:
           | You can run this on an M2 Mac Studio. That'll be consumer
           | level hardware in a few years.
        
             | wayfinder wrote:
             | I have an M1 but I first remember thinking it was very
             | impressive that my Mac could run these AI models...
             | 
             | ...until I tried the same on my RTX 4070 and it made my Mac
             | look like a joke.
             | 
             | For the 30 seconds my Mac would have taken for 1 result,
             | which will probably need revising, the RTX would give me 30
             | results.
             | 
             | However the RTX was half the cost of my Mac, so it's not a
             | good investment if I just want to generate some images. I'd
             | rather pay for the cloud if I didn't have the RTX already.
        
             | Closi wrote:
             | Run what? Adobe's offering is cloud-only, and I don't think
             | the hardware requirements are disclosed.
        
         | bhouston wrote:
         | > I'm sure that in the not too distant future (a few years at
         | most) we will be happily running these on customer level
         | hardware.
         | 
         | I doubt that is what Adobe will do. This is a new revenue
         | stream for them, why would they remove it?
         | 
         | Gimp will use local generation but Adobe is using a proprietary
         | dataset that they can keep secure in the cloud.
         | 
         | So yeah this is going to be sticking around.
        
         | financypants wrote:
         | Sure we will be happily running these on customer level
         | hardware, but there will always be a stronger version in the
         | cloud "worth paying for," won't there?
        
         | giancarlostoro wrote:
         | > we will be happily running these on customer level hardware.
         | 
         | This is the stage of AI that will impress me most. If I can use
         | your AI completely offline on my device on a spaceship orbiting
         | Pluto, then I will say we have achieved an AI capacity that is
         | impressive, even if its got the quirks of chatgpt today.
        
           | brianwawok wrote:
           | We can already do that. But why give away the keys to a money
           | machine?
        
         | gochi wrote:
         | They can upgrade hardware to bleeding edge faster than
         | consumers can get their hands on remotely comparable products.
         | Most consumers also don't upgrade every year to stay on
         | bleeding edge, they just tolerate what they can currently do.
         | While this will be fine for most, those actually using
         | generative AI for work aren't likely going to tolerate
         | stagnation as easily.
        
         | janehdoch wrote:
         | I think that we still need a few years 3-5 alone for better
         | models, different architectures and more finetuning.
         | 
         | Those models will affect us more than today already and change
         | how we perceive AI.
         | 
         | Than we will start to see AI optimized hardware (much more
         | optimized).
         | 
         | And than perhaps in 10 years we all run a lot more models
         | locally.
         | 
         | Nonetheless or despite this, the normal consumer doesn't run
         | open models and will probably not do that for a very long time.
         | Searching, keeping up-to-date and running models is still
         | effort and the usage model makes a ton of sense. Escpecially in
         | time of SaaS.
         | 
         | Im not running wikipedia locally. And none of my social circle
         | operates infrastructure / server.
         | 
         | People just want to use it.
         | 
         | Besides that, whatever local models or open models will be able
         | to do, AIaaS will have faster models, better models and more
         | convinient models.
         | 
         | I'm just waiting to pay for google assistent if it becomes
         | smart and can manage my emails my calendar and everything else.
         | After all my gmail account already has access (through email
         | and password reset) to most services i use.
         | 
         | I'm more curiuos when we will see AI service integration
         | through much more system to system communication. Machine
         | friendly apis (which partially already exist anyway)
         | 
         | PS: Look at how fast hardware development currently is. Not
         | much change in Memory etc. Models will not just become 100x
         | smaller in just a few years. We are right now at optimizing
         | those models to be cost efficient. Alone this phase will take a
         | few years.
        
         | amelius wrote:
         | > on customer level hardware
         | 
         | Unless nVidia changes their monetization model, and for example
         | introduces an App Store for AI, with subscriptions, of course
         | on locked down hardware.
        
         | code51 wrote:
         | Adobe is planning for a post-regulation, post-biggest-lawsuit
         | world for sure. All their steps show that they'll base their
         | offering on their own commercial data - paying training license
         | fees to stock images and artists.
         | 
         | Whether the amounts they pay would make licensing your work
         | sensible or not, Adobe is surely assuming this will ultimately
         | end up as Napster-to-Spotify transition.
         | 
         | If we end up "happily" (means legally as well) running these on
         | customer level hardware, then the question won't be about
         | credits of computation. It'll be about credits to use licensed
         | work.
        
           | sebmellen wrote:
           | This is the most plausible reason from what I see.
        
           | whiddershins wrote:
           | That's a big bet. My instinct is it doesn't go that way. We
           | will see.
        
           | GaggiX wrote:
           | > It'll be about credits to use licensed work.
           | 
           | If this is true (which I kinda doubt), is it going to matter
           | to most people? Like you can't really tell the images used to
           | train a model from the images it generates (if it's trained
           | correctly), so I doubt the majority of people would care,
           | like those who already use MJ for example. Training models on
           | copyrighted data for academic research will be allowed, the
           | models will be published, and good luck enforcing the
           | licence; and here I'm talking about the worst case scenario
           | where a court would find an image generated by AI to be
           | derivative of another image in a pool of billions in a
           | dataset (this goes way beyond any definition of derivative
           | work for now).
        
       | tikkun wrote:
       | I suspect many companies will do something like this - prepaid
       | credits or tokens for AI features that have high inference costs.
       | Inference costs per user are high, at least much higher than most
       | traditional software costs. This way, the costs are aligned with
       | the usage.
       | 
       | We'll also see occasional subscription products, but only when it
       | can be done in a way that is comfortably gross margin profitable
       | for most users of a company. (Eg ChatGPT Plus, Claude Pro,
       | Midjourney, 365 Copilot)
       | 
       | This will only change when the cost of inference goes down by a
       | lot.
        
       | xwdv wrote:
       | Hmm, what if they expand this to charge for small credits for
       | various other tools? Filters, brushes, etc.
        
         | Etheryte wrote:
         | Please stop giving Adobe ideas.
        
           | pphysch wrote:
           | ERROR: 0.25oz of liquid remains in your Adobe(tm) x Mountain
           | Dew(tm) Verification Can(tm). Please finish drinking to
           | proceed using the software product. Thank you for your
           | compliance.
        
         | Traubenfuchs wrote:
         | At this point you could just call it microtransaction. Imagine
         | free, but extremely bare bones implementations of photoshop
         | where anything but one brush and eraser is hidden behind
         | individual or package deals.
        
       | xu_ituairo wrote:
       | "Generative credits provide priority processing of generative AI
       | content across features powered by Firefly in the applications
       | that you are entitled to. Generative credit counts reset each
       | month."
        
       | slowhadoken wrote:
       | Create a problem, charge for the solution.
        
       ___________________________________________________________________
       (page generated 2023-09-16 23:00 UTC)