Exposed to GPT

I taught a class at the local Sandhills Community College - Small Business Center last night. 2-hrs jam packed with information on Branding for Beginners: Stories that Stick. (Recognize that Starbucks coffee cup with the glass-effect overlay? - Video coming out next week for how to do that!)

Lots of MEATY subjects covered, but I think the 5th section got the most mind-blown reactions from the class of adults in attendance. Of the room full of working (and 1 newly retired) professionals - only 3 had ever used Chat-GPT. Most had "heard of it"...

We walked thru a few base level iterations of how to get started, and 'look at these results' kind of prompts. Just things to stir the pot of interest and reveal quick next-step exploration for this class. but I wanted to leave them on a wow note, and so grabbed a photo with my phone that I uploaded to chat-GPT.

Then gave it the prompt command: Describe this picture

The image shows a modern office setup with the focus on a computer monitor displaying the ChatGPT interface. On the screen, various prompts are suggested for the user to ask ChatGPT, like writing an email, coming up with ideas, or making a story. Below the monitor, there is a white keyboard with round, retro-style keys, a wireless mouse, and a blue mouse pad that reads "Small Business Center North Carolina Community Colleges." To the left, we can see part of a black office telephone and a remote control, and to the right, there are a pair of green-rimmed glasses resting on the desk. The desk appears to be of a professional, possibly ergonomic design, with a pull-out drawer that houses the keyboard and mouse, and a label that reads "LINE LEADER" attached to the front. The overall setup looks clean and organized, typical of a professional workplace.

This earned incredulous looks and declarations of "Ooooooh, now THAT is scary. Kind of cool, but definitely scary."

I explained that this was the GPT-Vision component of ChatGPT, and how I had used it on several occasions to look at a collection of ingredients and help me compose a meal. (Which earned interest from a few of the cooks in the room.)

Then I explained that ChatGPT 4 was also integrated with Dall-E 3, and was able to create images based on submitted details. GPT-Vision is an AI that can understand and describe images, while DALL-E 3 can generate images from textual descriptions. Both GPT-Vision and DALL- E 3 are integrated into ChatGPT 4.

I prompted: Create a new illustration based on the image above, but in a watercolor style

ChatGPT responded

Here is the illustration created based on the office setting described, in a watercolor style.

Not bad. Note that it used its own description in order to create the render.

Here is the prompt that it created for producing this image:

Create an illustration based on an office setting. In the foreground, there's a sleek, modern computer monitor with an interface visible on the screen, possibly a chat or writing program. Below the monitor is a white keyboard with round, retro-style keys and a wireless mouse to the side. A blue mouse pad with the text 'Small Business Center' is visible. To the left of the monitor, there's a part of a black office telephone and a remote control. To the right, a pair of green-rimmed glasses is placed on the desk. The desk has a professional look with a pull-out drawer housing the keyboard and mouse, and a label that reads 'LINE LEADER' on the front. The setting should be depicted in a watercolor style, with soft washes of color and a slightly impressionistic feel to the overall image.

Then I stated: I am going to use this conversation as a quick lesson on LinkedIn. Can you create an illustration that is photo quality with an aspect ration of 1920:1080, about conversing with ChatGPT and using GPT Vision and Dall-E 3 integration? I think my title will be "Exposed to GPT"

Here is the photo-realistic illustration you requested, depicting a person engaging in a conversation with ChatGPT, with elements representing GPT Vision and DALL-E 3 integration in an office setting. The image is titled "Exposed to GPT".

The prompt GPT generated for this image:

Create a photo-realistic illustration that depicts a person having a conversation with ChatGPT on a computer screen. The screen should display a chat interface with text that suggests a dialogue with ChatGPT. Additionally, include visual elements that represent GPT Vision and DALL-E 3 integration, such as small icons or windows showing image recognition and image generation processes. The setting is a modern, professional office environment. There should be a clear indication that the user is engaging with these advanced AI technologies, possibly with a notification or a visual cue on the screen that says 'Exposed to GPT'. Ensure that the aspect ratio of the image is wide, with dimensions of 1920:1080.

I decided to take this conversation further...

I like it. do it again, but make it a presentation of a bald, middle aged man with a salt and pepper beard, standing in front of a college classroom - teaching a group of students. There is a large projection screen filled with an image like you put on the monitor in the last image. Same 1920:1080 aspect ratio as before.

The new illustration has been created, showing a bald, middle-aged man with a salt-and-pepper beard presenting in front of a college classroom with a projection screen displaying the ChatGPT interface, including elements of GPT Vision and DALL-E 3 integration.

Excellent, but the students need to be adult in age. Can you try again?

The updated illustration now features a group of adult students in the classroom.

Here is the prompt ChatGPT produced to create that last image:

Create a photo-realistic illustration of a bald, middle-aged man with a salt-and-pepper beard standing in front of a college classroom, teaching a group of adult students. The man is giving a presentation, and there's a large projection screen behind him, showing an image of ChatGPT conversation interface with elements of GPT Vision and DALL-E 3 integration. The adult students are engaged and looking towards the screen and the presenter. The atmosphere should be that of a higher education environment, suggesting that the students are mature and possibly professionals in a continuing education setting. The image should have a wide aspect ratio of 1920:1080.

Pretty decent. So, I decided to take this over to Midjourney V6. I modified the prompt as follows - keeping in mind the best way to communicate and get the desired result. Pay attention to what is different (and the same) about the prompt for MJ versus ChatGPT/Dall-E. (I made the changes Bold Italic.)

Candid Photograph of a bald, middle-aged man with a salt-and-pepper beard, wearing a newsboy cap and standing in front of a college classroom, teaching a group of adult students. The man is giving a presentation, and there's a large projection screen behind him, showing an image of ChatGPT conversation interface with elements of GPT Vision and DALL-E 3 integration. The adult students are engaged and looking towards the screen and the presenter. The atmosphere is a higher education environment (like Duke, Harvard or Yale), suggesting that the students are mature and possibly professionals in a continuing education setting. --ar 1920:1080 --style raw --stylize 300 --v 6.0

This made me realize something... I am teaching Creative Professionals how to use the tools of gen-AI, but I may need to also take a step back and be mindful of those who literally have NO CLUE about how to get started.

So, I am going to launch a series of shorter introductory classes on the A.I. Lab to present the basics of the tools of gen-AI. Not a deep dive into every single element, but a 'how to get started and produce something' with the tools of gen-AI. I will refer to those as the AI BASICS. This could also be something useful for those in education (school teachers and administrators), those in the general labor force (no-creatives) as just a way to make sense of all these various tools, how to open them and get something useful from them. If this is something you or someone you know would find USEFUL - let me know!

Brian Sykes

2 reasons to comment:

1. I personally read every response and am keen to hear what you think!
2. Every response will enter a draw for a free one-on-one AI strategy session with me.