Let's talk about the pink elephant in the room!

How DALL-E 2 makes AI progress visible to everyone.

I've been waiting for this for a long time. Now it's finally here: access to DALL-E 2. My anticipation was huge and I'm absolutely thrilled to see what's possible with the new system. When I received beta access last week, I immediately set out to try out the AI. Often, artificial intelligence is seen as abstract and something that can't be grasped. With DALL-E 2, AI becomes really tangible for the first time. We immediately see the results of AI's work - and they never cease to amaze me.

For those who don't know what DALL-E 2 is, here's a quick summary: It's an artificial intelligence based on a neural network from OpenAI. The AI system can create original, realistic images and artwork from a text description. It can combine concepts, attributes and styles to create something never seen before. Even your own images can be uploaded and modified. For example, how about a sea monster suddenly rising out of the sea from a vacation picture on the Côte d'Azur? Obviously, there are also more practical applications, for example to remove or add elements to an existing picture, make touch-ups or expand the image canvas.

 

A weekend of trying out DALL-E 2

When I sat together with my friends on the weekend, we had fun testing DALL-E 2 extensively. We fed the system with the weirdest combinations and were excited to see the results. How do you think a 19th century European city looks when drawn in steampunk style? How does DALL-E 2 transpose a pink elephant in a nightclub? Within seconds, the AI created photos or images in the desired style - whether realistic, as a 3D render or oil painting. DALL-E 2 also knows styles like cyberpunk or neon, or even imitations of famous artists. The impressive thing is that it doesn't use existing images and change them, but actually creates something completely new and unique. With each image, my curiosity grew as to what could be created from even more absurd concepts - and we were surprised again and again at the high quality with which our wishes were implemented. There was something enormously fascinating about seeing AI come to life for the first time.

 

With DALL-E 2, the pink elephant made of balloons comes into the small white room.

 

There are no limits to the imagination

I, like my friends, am absolutely thrilled with what DALL-E 2 does. Whether it's cats handling 80s hardware in space, children's drawings of a whale with a boat and a house, or a duck dressed as a captain holding a glass of whiskey. Anything is possible - and there are no limits to your imagination. There are suitable images for every description. These can also be used without restriction for articles, blogs, book covers or advertising campaigns, for example. For my part, I can spend hours with DALL-E and always discover something new that inspires my creativity.

 

My creativity is challenged.

 

It goes much further

DALL-E 2 is just a taste of what AI is capable of in the future. Images and text can be created almost perfectly with AI today with just a few parameters. Looking further into the future, in a few years we will also create videos and even entire series and movies using AI. If today we feel like a comedy with black humor or we want a tragic love story - we will get it in a few minutes. And as always, we can't even imagine the possibilities that are yet to come.

 

Job killer or helpful tool?

For all the euphoria I feel for DALL-E 2, there are also people who see it as a job killer. Artists, creatives and drafters sometimes feel threatened by AI and fear for their jobs. That may be a corrected fear, but we still have to differentiate here. Would we rather hang the picture of a good artist in our living room or the one created by an AI? Those who are really good at their jobs have nothing to fear, but can use DALL-E 2 as a valuable tool to draw new inspiration or expand their own artistic horizons.

 

In the future, the possibilities of AI will be all-encompassing.

 

Safety first

To ensure that such a powerful AI is not misused for negatives, such as deep fakes, OpenAI has taken some safety precautions. For example, no photos with faces can be uploaded. Public figures are also taboo. Likewise, it is not possible to create violent, lewd or political images on DALL-E.

 

Conclusion - DALL-E 2 is just the beginning

I am curious to see how DALL-E 2 and other image generators in general will develop and also what AI will be capable of in the future. We are at the beginning of an exciting journey that will change our world. Now it's up to us how we shape it.

Fancy a coffee?
© Philipp Neuberger, Berlin
|
|
|