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

491–500 of 579 posts

Re: DALL·E now available in beta

#491

Post by anhner »

anhncommenter25 wrote: post #430
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?

I recently talked with a concept artist about DALL-E and first thing they mentioned was "you know that's all stolen art, right?" Immediately made me think of GitHub Copilot. However the artists being featured in DALL-E's newsletters can't stop gushing about 'the new instrument they are learning how to play' and other such metaphors that are meant to launder what's going on. My theory is that the professions most at-r…

I first read the artist's reply as "you know all art is stolen, right" which made more sense to me. If you look at the history of art, you'd also know that it's true.

Re: DALL·E now available in beta

#492

Post by andybak »

monkmartinez wrote: post #426
blueboo wrote: post #330

Earlier quoted context omitted.

If you’re technically inclined, I urge you to explore some newer Colabs being shared in this space. They offer vastly more configurable tools, work great for free on Google Colab, are straightforward to run on a local machine. Meanwhile we should prepare ourselves for a future where the best generative models cost a lot more as these companies slice and dice the (huge) burgeoning market here.

Can you share a few, please? I am already out of credits on Midjourney and I don't even feel like I got the hang of it until I was almost out.

Do you have a decent GPU and a Windows box? If so install VOC: https://softology.pro/tutorials/tensorflow/tensorflow.htm

Re: DALL·E now available in beta

#493

Post by anhner »

ZeWaka wrote: post #319
eigenvalue wrote: post #302

Earlier quoted context omitted.

I've been using this app to upscale the images to 4000x4000, and it works amazingly well (there is also a version for Android): https://apps.apple.com/us/app/waifu2x/id1286485858 I paid extra to get the higher quality model using the in-app purchase option. It crushes the phone's battery life, but runs in only ~10 seconds on an iPhone 13 Pro for a single 1000x1000 input image.

I mean, waifu2x and similar waifuxx libraries are free and open-source, there's really no reason to pay for it if you're working on a desktop.

what's a desktop? /s

Re: DALL·E now available in beta

#494

Post by msoad »

nuclearsugar wrote: post #405
thorum wrote: post #182

Earlier quoted context omitted.

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.

DallE2 + Topaz Gigapixel AI works amazingly well.

Photoshop also has super resolution

Re: DALL·E now available in beta

#495

Post by imdsm »

anhncommenter25 wrote: post #430
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?

I recently talked with a concept artist about DALL-E and first thing they mentioned was "you know that's all stolen art, right?" Immediately made me think of GitHub Copilot. However the artists being featured in DALL-E's newsletters can't stop gushing about 'the new instrument they are learning how to play' and other such metaphors that are meant to launder what's going on. My theory is that the professions most at-r…

Having used copilot for over a year now, it isn't there to replace programmers. It isn't called GitHub Pilot, and it doesn't do well with generating original ideas. Sure, if your job is to create sign up forms in HTML then sure, it'll do your job in a second, but if you're creating more complex systems, copilot is just there to help save you time when writing code (which is just implementing ideas).

Think of it like a set of powertools saving you time over manual tools.

Re: DALL·E now available in beta

#496

Post by imdsm »

marcosdumay wrote: post #462
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?

Is it trained on unlicensed work gathered at random at the web? (I really don't know, and I didn't find anything about it on their site.)

In a way, it's no different than an artist walking through an art gallery then going home, inspired, to paint a dark portrait a la Rembrandt

Re: DALL·E now available in beta

#497

Post by andybak »

ZeWaka wrote: post #319
eigenvalue wrote: post #302

Earlier quoted context omitted.

I've been using this app to upscale the images to 4000x4000, and it works amazingly well (there is also a version for Android): https://apps.apple.com/us/app/waifu2x/id1286485858 I paid extra to get the higher quality model using the in-app purchase option. It crushes the phone's battery life, but runs in only ~10 seconds on an iPhone 13 Pro for a single 1000x1000 input image.

I mean, waifu2x and similar waifuxx libraries are free and open-source, there's really no reason to pay for it if you're working on a desktop.

Things have moved on a considerable amount since waifu2x

Try https://github.com/n00mkrad/cupscale

Re: DALL·E now available in beta

#498

Post by andybak »

smusamashah wrote: post #286
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.

They won't. DALL-E images are mostly not as high quality. The high quality stuff which everyone has been sharing is result of lots of cherry picking.

It's a lot better than you are claiming. Mind if I ask if you have access personally?

Re: DALL·E now available in beta

#499

Post by andybak »

JulianaRestrepo wrote: post #461
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.

remember DALL-E is not allowed to generate faces

No longer true.

Re: DALL·E now available in beta

#500

Post by lostgame »

bemmu wrote: post #58

I was supposed to be making a video game, but got a bit sidetracked when DALL·E came out and made this website on the side: http://dailywrong.com/ (yes I should get SSL). It's like The Onion, but all the articles are made with GPT-3 and DALL·E. I start with an interesting DALL·E image, then describe it to GPT-3 and ask it for an Onion-like article on the topic. The results are surprisingly good.

This a fucking fantastic site, it’s absolutely hilarious, and I’ve bookmarked it - I kinda unironically want to set it as my home page - but just a heads up that the CSS is broken for me on my iPhone SE2.

The images don’t scale properly with the rest of the site, they’re massive compared to the content.

Post reply on HN