ChatGPT on a Next Level: Generating Images from Text
Read my text in ChatGPT’s chat window:
“A woman is wearing an AI cape — bit worried about the emergence of AI — but also relieved the importance and use of it”
Just as I enter this prompt in the ChatGPT’s chat window, four images appeared in a couple of minutes:
Four wonderful images. Even if I’m an artist, don’t think I’d come with these images at least until I spend time imagining the art!
How cool is that!
If you are not blown away with the magic of this, I don’t know what else could.
Let me introduce you to a new killer feature OpenAI added to ChatGPT — the DALL-E project!
DALL-E is on ChatGPT
OpenAI released long awaited test-to-image generation AI model — the DALL-E AI model to the ChatGPT plus subscribers.
It was rolled into my subscription just today. I gave it a try and just as Midjourney surprised me earlier, so is the ChatGPT now!
And you can pick one and ask ChatGPT to make modifications.
AI Meets Musician
While I was playing with ChatGPT to generate new set of images and stuck in awe with the images that were being produced, my son (he’s into music) entered my office!
He modified the prompt asking ChatGPT to add an “electric violin”, picking the 3rd image.
“Let’s go with Image 3. Can you add a blue electric Violin to image 3. Show her standing inside the violin.”
This is the result:
That was awesome! I couldn’t imagine the depth and details in this image.
I was seriously amazed and surprised with the details emerged in that pic.
Converse with it as you’d chat with your Granny
I thought I’d take this to next level. I’ll ask it to add a backdrop of London and its iconic landmarks in the scene:
“Cool. Let’s create this scene on the banks of river Thames in central London. Make sure Big Ben, St Pauls, Parliament and few other historic places and landmarks are shown in the scene.”
I submitted my newly added requirements to the prompt; it duly produced the output:
How cool is that!!
Love the back scene of London! It captured famous landmarks as expected.
Finally I asked ChatGPT to fast forward to 2050:
You can see the future — fortunately the Big Ben, UK parliament, St Pauls and few others survived in the imagination.
Being cheeky, I’d want ChatGPT to show me the future of 3000. Let’s see how it can imagine the world in 3000:
The monumental landscapes were still around. Hmm. I asked if they’d survive :)
I could go on! I can spend my evening playing with this, it is so inviting and jaw dropping and inspiring and surprising and thrilling!
Enabling DALL-E
To try DALL-E, if you are a plus subscriber, click on the Dall-E plugin as shown here:
That’s a swift whirlwind tour of what our future of imagination would be with the AI coming closer to us every second!
