Is DALL-E Groundbreaking Technology or Not? Costs ≈0.13$ Per Image and More DALL-E 2 Access Details
9/1/2022 by David Smooke
Photo Credit, inputting "pixelated green clock on a laptop screen in san francisco" via DALL-E.
This Slogging thread by richard-kubina, Guy, David, Limarc and Sheharyar Khan occurred in hackernoon's official #random channel. It is about the first couple hundred images we generated with DALL-E.
I've got access to Dall-e :excited:
Bruh where’s my beta key?!?!
Those featured images - gonna be dope!
:party_parrot:
So for image creation, is it groundbreaking technology or not?
From what I’ve messed around with it’s pretty good! $15 for 115 image generations so not tooo pricey either! Seems to be a bit more consistent than GPT3 too
Revolutionary!
0.13$ per image range
DALL E has impressive floor. Almost every image feels a little like a human made it
This looks like a lot of fun
Sheharyar Khan
Since Dall.E couldn't generate an accurate picture of you David, here's Turnstile vocalist Brendan Yates, who I think comes https://www.google.com/search?q=Brendan+Yates&client=firefox-b-d&sxsrf=ALiCzsZ3aQ77IocJY8gao_CR-srnKGXNFQ:1661960185620&source=lnms&tbm=isch&sa=X&ved=2ahUKEwj0gfqTtPH5AhXJ8bsIHWi1DHQQ_AUoAXoECAIQAw&biw=1920&bih=927&dpr=1#imgrc=yf1ABBCoDMyUnM.
Sheharyar Khan Ha! That's a new one. Last 'celebrity' lookalike someone called me was https://www.google.com/search?q=Milo%C5%A1+Teodosi%C4%87&newwindow=1&client=firefox-b-d&hl=en&sxsrf=ALiCzsa-WuCP_p8ISXM36EguvClgWWhoyw:1661967043844&source=lnms&tbm=isch&sa=X&ved=2ahUKEwii-5razfH5AhWOAjQIHU11ApkQ_AUoAnoECAIQBA&biw=1364&bih=790&dpr=1.8
Limarc it is fun. can try it out here https://labs.openai.com/waitlist
Oh that's interesting they don't want you to use it to create deep fakes of famous people
yeah, I've seen folks using stable diffusion to do that though https://www.reddit.com/r/MediaSynthesis/comments/x2j01s/modern_celebrities_alternate_timelines/
This one has credible Adult Swim vibes for a show called something like "Suzette & Steed"
Sidenote: Who owns this image? Me or http://OpenAi.com. My words. Their tech. On the free trial plan...
Ok got to my first viable turn company logo into featured image art about company moment
I can't seem to find the right way to tell the DALL-E it's noon time!
Comments
undefined
9/1/2022
Stable Diffusion is 1000x better. You can run it for free I colab, you get img2img, inpainting, masking, and so much more coming out. There's even vid2vid where you can feed a video clip in, and have the AI change it. So you could already make Captain America into a cartoon or make Seth Rogan Captain America like a deep fake. Add in audio(soon) and you can generate your own movie on demand, given the processing power and provided Klaus “You will own nothing, and you will be happy" Schwab allows you to have electricity.
Malhar Thanki
9/2/2022
It’s Exciting😦! and it’s just 2022 yet👀
Samantha Cook
9/2/2022
Get outta town. Buckaroo Banzai!