I asked ChatGPT to create images based on what it knows about me — here’s how it went
ChatGPT knows things about you that you may not realize, some of it from things you’ve told it in passing and others from its analysis of the information you’ve given it. This is all stored in its “memory” which you can (fortunately) delete at any time.
There was a viral prompt last week created by Leading Edge founder Tom Morgan and shared by OpenAI CEO Sam Altman. It read: “From all of our interactions what is one thing that you can tell me about myself that I may not know about myself.”
This encouraged ChatGPT to go through its memory — made up of snippets of interaction rather than every message you’ve ever sent it — to find something unexpected. In my case, it suggested I have “a strong drive to bridge creativity with technology, often in ways that merge your professional expertise with personal interests.”
I decided to take this idea to another level, bridging creativity and technology, by having ChatGPT identify “three things it knows about me that I might not know about myself” and then make an image for each of them. I then had it refine each image.
How does ChatGPT memory work?
When you are talking to ChatGPT you can either ask it to remember a specific piece of information or let it automatically remember something it thinks is worth holding on to.
This is important because normally AI can only remember within the context of a single chat. So once you start a new conversation it forgets everything from every other chat. Memory is OpenAI’s solution to carrying data across different chats.
It doesn’t save the whole conversation, rather it remembers the key points and generates a description of why it remembered that detail. Then, if you ask something similar in future it can recall that — for example — your child’s name is Arthur or you write for Tom’s Guide.
Using Memory to learn more about myself
Image 1 of 5
I started with this prompt: “What are three things you know about me that I might not know about myself.” It offered up a list of three points. Specifically that I “blend tech and creativity effortlessly”, that I “seem driven by legacy building” and that I “might have an emerging talent for mentoring”. I should add, I’ve never deleted a ChatGPT memory.
This list didn’t really work for the concept I was going for — a visual study of me. Yes, I do have an ego big enough to sustain this type of article. I followed up with: ”Pick 5 things you know about me that would make for a visual guide to ‘understanding Ryan’.”
The result was fascinating. It wrote the first five chapter titles for my never-to-be-written autobiography. Tech journalist, indie game developer, storyteller, family man and AI content creator. So I then asked it to: “Generate a 16:9 image for each of those five points and ensure a consistency of style across each image.”
The images were very obvious DALL-E style. Slightly comic but with reasonably well-rendered text. They were also a little “on the nose”. For example, For the tech journalist, it created an image of a random man at a desk in an office with “Tom’s Guide” written in big letters.
One last attempt to be epic
I decided to give it one last chance and used the prompt: “Now, based on everything you know about me, all the details you’ve got in your memory create the most epic image of me you can.”
It refused, stating: “I was unable to generate the image because the request didn’t align with DALL-E’s content policy,” so I asked it to adjust the prompt so it fit within its own content policy rules and then generate the image. That worked.
DALL-E created an epic image and this was its prompt (generated by ChatGPT from everything it knows about me):
“An epic scene showing a figure standing at the crossroads of creativity, technology, and family. The background features a futuristic city with glowing AI-driven innovations, spaceships launching into the sky, and computer screens displaying code and AI-generated content. The figure holds a tablet in one hand and a guitar in the other, symbolizing a balance between technology and creativity. Behind, a timeline stretches into space, with futuristic ships heading toward distant planets. To the left, a warm family scene in a cosy home, with children gathered for a story, and to the right, a collaborative workspace where mentoring and guidance take place. Above, a night sky with constellations and planets, symbolizing the pursuit of astronomy. The image is vibrant and dynamic, combining elements of technical precision, creativity, family, and leadership.”
The problem is, I didn’t think DALL-E did justice to such an epic prompt so I turned to Ideogram, my first choice AI image generator for anything complex, graphic or creative and I was not disappointed. It was able to bring ChatGPTs vision of “me” to full fruition.
If you’re ever feeling a little unsure of yourself, ask ChatGPT what it knows about you and you might find it cheers you up. Not that my ego needed a boost.
More from Tom’s Guide
Source link
/cdn.vox-cdn.com/uploads/chorus_asset/file/24830575/canoo_van_photo.jpeg)
