The AI system GIVE HER - currently in version 2 - can currently only be tested with an invitation. Public use isn't possible yet, so I haven't been able to take a close look at DALL-E 2 yet.
By chance, however, I came across a way of using DALL-E without an invitation. I would like to show you this way briefly here. To set the mood, here is a small work of art that I created with the AI.
Steve Jobs as Superman - that's what Jasper Art and DALL-E 2 made of it.
I use AI here and there to write texts for my job "Jasper“. This understands many languages and can even generate German texts with few specifications. But that is a topic for another article.
What I noticed in Jasper's menu is a new menu item called "Jasper AI". Behind this is the offer to have pictures created for USD 20 per month. These can then be used, for example, for your texts or other areas of application.
My research into which neural network is behind Jasper AI has shown that DALL-E 2 is doing its job under the hood. So the art AI, which you can not actually use as a mere mortal.
Jasper Art's brief was to create a bear typing on an iMac in an office. I would say: mission accomplished!
DALL-E 2 at a fixed price – without paying per image
Of course, the thing is not available for free, but on the one hand you have to have a current subscription jasper.ai and then book the Jasper Art option for 20 USD per month. The Jasper subscription starts at $29 per month, which ends up at $59.
The advantage, however, is that you can use the Jasper Art option indefinitely, while with DALL-E you pay 2 (if you have an invite link) per image generation - currently around 15 USD for 460 images, including the 4 previews, of course. Count the pictures.
In an Apple store in New York – I think the AI did a pretty good job of that.
Jasper Art: DALL-E 2 with restrictions
Those who use Jasper Art will also notice that there are a few feature limitations. For example, the inpainting function and the image upload (to edit an existing image with DALL-E 2) cannot be found.
The resolution that Jasper Art delivers is 512×512 pixels, so for use on the Internet you can actually still use a tool like AI Gigapixels must increase. On the DALL-E 2 website I was able to find example images with 1024×1024 pixels. So it may be that Jasper Art has built in a limitation here as well.
A bear typing on the iMac in an office - that's what it looks like. I doubled the resolution with the Gigapixel AI, which worked great.
My conclusion: Interesting for researchers
Of course I booked the subscription to see what DALL-E 2 can do. In principle, however, it's more for little explorers, because who needs a photo of a bear typing on an iMac? However, the creation of images with AI is so exciting that I just had to take a look.
A publicly available alternative to DALL-E 2 is midjourney, which I also booked. While DALL-E is better at "realistic" images, Midjourney can produce great artwork, get higher resolution, and specify different proportions as well.
If you have a spirit of discovery, check out both Als. 😊
Jens has been running the blog since 2012. He appears as Sir Apfelot for his readers and helps them with problems of a technical nature. In his free time he drives electric unicycles, takes photos (preferably with his iPhone, of course), climbs around in the Hessian mountains or hikes with the family. His articles deal with Apple products, news from the world of drones or solutions for current bugs.
Did you like the article and did the instructions on the blog help you? Then I would be happy if you the blog via a Steady Membership would support.
9 Responses to “Use Kunst-AI DALL-E 2 without invitation”
Beatrice Willius
I tried DiffusionBee a few days ago. This is a desktop app for Stable Diffusion. The app is still very beta. For example, I had an empty interface on the first start. Weird stuff like a cancel button is considered exotic. Unfortunately the program is much too slow on my MacBook Air. I only tried two simple prompts. The result was bad.
Ah, is there an app that can be used for Stable Diffusion yet?! I'll test it, but you can certainly wait a few more weeks before trying the app. 😂 I've had good experiences with Midjourney. I had to get used to Discord first because they took it as a GUI, but basically it was a good move. And you can try Midjourney for free. But I booked the 30-dollar subscription relatively quickly in order to be able to test properly.
Short feedback: CHARL-E is meant nicely, but incredibly lame even on my M1 Max. I specified 25 steps and wanted 4 thumbnails. Now 5 minutes have passed, in which I would have already created 5×4 pictures with Midjourney. I don't see anything at CHARL-E. According to the log, he's trying to download something and there are error messages... but they don't tell me anything. So I'll stick with Midjourney. 😊
Hi Jonas! It took me 3 minutes now to understand what's wrong there. 😊 It's just too late, I think... But thanks for the tip. The spelling mistake has been corrected!
Use Apple AirTag for NFC actions on the iPhone: Here's how!
I tried DiffusionBee a few days ago. This is a desktop app for Stable Diffusion. The app is still very beta. For example, I had an empty interface on the first start. Weird stuff like a cancel button is considered exotic. Unfortunately the program is much too slow on my MacBook Air. I only tried two simple prompts. The result was bad.
Ah, is there an app that can be used for Stable Diffusion yet?! I'll test it, but you can certainly wait a few more weeks before trying the app. 😂 I've had good experiences with Midjourney. I had to get used to Discord first because they took it as a GUI, but basically it was a good move. And you can try Midjourney for free. But I booked the 30-dollar subscription relatively quickly in order to be able to test properly.
CHARL-E, another new Stable Diffusion Mac app with a little more options (multiple imaging):
https://www.charl-e.com
Joachim, thanks for the tip! This will be loaded and tried out immediately. 😊
Short feedback: CHARL-E is meant nicely, but incredibly lame even on my M1 Max. I specified 25 steps and wanted 4 thumbnails. Now 5 minutes have passed, in which I would have already created 5×4 pictures with Midjourney. I don't see anything at CHARL-E. According to the log, he's trying to download something and there are error messages... but they don't tell me anything. So I'll stick with Midjourney. 😊
Hi Jens,
The name of the link is wrong in the text above:
but on the one hand you have to have an ongoing subscription with Japser.ai
Hi Jonas! It took me 3 minutes now to understand what's wrong there. 😊 It's just too late, I think... But thanks for the tip. The spelling mistake has been corrected!
Sorry, I could have written better.
All good no problem! I could have understood it right away, but I was probably too tired. 😊