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

DALL·E now available in beta

openai.com

Post reply on HN
Print view View original

481–490 of 579 posts

Re: DALL·E now available in beta

#481

Post by meowface »

antiterra wrote: post #475
gwern wrote: post #413

Earlier quoted context omitted.

It's true that image models are much less of a burden on GPU VRAM than a model like BLOOM where fitting it into a few A100s is ideal, but these diffusion models are a PITA for a ordinary hobbyist in terms of total compute: the CLIP pass over the text input is almost free, but then you feed it into the diffusion model, for one sample you'll be doing 10-100 forward passes (depending on how fancy the diffusion methods a…

> 6-9 means a joykilling minute+ wait Excet we’re already waiting 90+ seconds for dall-e mini.

DALL-E 2 tends to be 15 - 30 seconds, from my experience.

Re: DALL·E now available in beta

#482

Post by csteubs »

thorum wrote: post #182
xnx wrote: post #76

I fully expect stock image sites to be swamped by DALL-E generated images that match popular terms (e.g. "business person shaking hands"). Generate the image for $0.15. Sell it for $1.00.

DALLE images are still only 1024 px wide. Which has its uses, but I don’t think the stock photo industry is in real danger until someone figures out a better AI superresolution system that can produce larger and more detailed images.

Another commenter mentioned Topaz AI upscaling, and Pixelmator has the "ML Super Resolution" feature; both work remarkably well IMO. There are a number of drop-in and system default resolution enhancement processes that work in a pinch, but the quality is lacking compared to the commercial solutions. There are still some areas where DALL-E 2 is lacking in realism, but anyone handy with photo editing tools could amend those shortcomings fairly quickly.

On-demand stock photo generation probably is the next step, particularly when combined with other free media services (Unsplash immediately comes to mind). Simply choose a "look" or base image, add contextual details, and out pops a 1 of 1 stock photo at a fraction of the cost of standard licensing. It'll be very exciting seeing what new products/services will make use of the DALL-E API, how and where they integrate with other APIs, use cases, value adds like upscaling and formatting, etc.

Re: DALL·E now available in beta

#483

Post by lionkor »

softwaredoug wrote: post #398

Surprised by the lack of comments on the ethics of DALL-E being trained on artists content whereas copilot threads are chock full of devs up in arms over models trained on open source code. Isn’t it the same thing?

It is bad. But:

As long as DALL-E isn't caught painting out a 1-to-1, reverse searchable copy of an image, its not really as bad as copilot, IMO.

The issue isnt just that copilot is trained on my GPL code, its that it might decide to copy paste lines from it, including my comments, etc.

Re: DALL·E now available in beta

#487

Post by spaceman_2020 »

dave_sullivan wrote: post #248
peteforde wrote: post #228

I have been having a blast with DALL-E, spending about an hour a day trying out wild combinations and cracking my friends up. I cannot imagine getting bored of it; it's like getting bored with visual stimulus, or art in general. In fact, I've been glad to have a 50/day limit, because it helps me contain my hyperfocus instincts. The information about new pricing is, to me as someone just enjoying making crazy imagines…

I think people don't realize how huge these models really are. When they're free, it's pretty cool. But charge an amount where there's actual profit in the product? Suddenly seems very expensive and not economically viable for a lot of use cases. We are still in the "you need a supercomputer" phase of these models for now. Something like DALLE mini is much more accessible but the results aren't good enough. Early ear…

How hard would it be to spin off a variant of this with more focused data models that cater to specific styles or art-types? Like say, a data model only for drawing animals. Or one only for creating new logos?

Re: DALL·E now available in beta

#488

Post by darawk »

softwaredoug wrote: post #398

Surprised by the lack of comments on the ethics of DALL-E being trained on artists content whereas copilot threads are chock full of devs up in arms over models trained on open source code. Isn’t it the same thing?

Why is this any worse than an art student learning to paint by looking at other painter's work?

Re: DALL·E now available in beta

#489

Post by numpad0 »

Nition wrote: post #393
numpad0 wrote: post #165

Earlier quoted context omitted.

I'm kind of surprised that no one had found "verbatim copy" cases as were made with GitHub Copilot. Such exact copies in photography are likely easier to go for than with code snippets.

It might be interesting to find an image in the training set with a long, very unique description, and try that exact same description as input in DALL·E 2. Of course it's unlikely to produce the exact same image, or if it does, you've also discovered an incredible image compression algorithm.

Oh I don’t have problems with DALL-E doing its thing, I just think it’s wrong if the purpose will be to cleanse off copyrights from images.

Re: DALL·E now available in beta

#490

Post by geysersam »

jamiek88 wrote: post #448
commandlinefan wrote: post #371

Earlier quoted context omitted.

> trying out wild combinations and cracking my friends up Wait until the next edition comes out where it automatically learns the sorts of things that crack you up and starts generating them without any input from you.

Ai generated TikTok could be almost like wire jacking humans.

How many years away is this? 5? 10? I seriously doubt it'll be longer than that, considering the recent advances of autoregressive models, and the overall trajectory of ML the last decade.

It'll use hideous amounts of compute.

Post reply on HN