Saturday, June 01, 2024

I started my conversation with GPT 4o with the most complex question I could think of. Here’s what happened.


The question, by the way, was derived from some graduate thesis research many years ago and was something I incorporated into a recent story called “When Everything Became Possible: The Story of Three Characters, Set Free to Help us Understand our Future with AI.” I was hoping it might be a good “thought partner” and help me piece together a 2.0 version of the story.

Could AI demonstrate all the ways that it is not being used? Could it find solutions for some of the biggest problems in the world, like what is happening in Ukraine or Israel, for example? Would it understand what I meant when I said that music might help us to communicate things in the world where language fails?


Before I decided to make the video public, I shared this with a colleague in education. Her response made me realize there is some deeper potential in generative AI than most people realize. In fact, I’m going to include her whole response because when you’re out here experimenting, it’s great to get this kind of validation and I am truly thankful for this.

Note: she agreed to let me share this anonymously.

G’morning Lee ☀️ Thank you so much for sharing your rough cut! 🤓
It is absolutely stunning and you are just amazing 🤩 I am in awe of your intellect and ability to create the conversation I witnessed (as odd as it is to say “conversation”).
I was unaware that we were at this level of AI, and the example of its application to “discuss” world issues is nuts. I fully understand the vital importance of prompt engineering. You are a master.
The other lightbulb moment for me was regarding the expansion of knowledge you offered to be considered went beyond humanity to nature. I just believe so strongly that the inclusion of knowledge beyond our 5 senses is critical to world peace. Opening to the natural world brings us to the root and the once strong connection to the beginning of human-hood and connection to each other and that which connects us all. It is who we really are-spiritual beings living a human existence for a short while. Nature is the path to the spiritual. This is what indigenous cultures exemplified, and that which our greed and ego bastardized.
You were also right on in terms of trauma responses. That is the work of Dr. Gabor Mate (Where Olive Trees Weep) and included discussions amongst his other work. The understanding that we are all acting out our trauma and what we can do. SO much rests on our ability to examine ourselves; our thoughts, biases, justifications, and behaviors.
To include all of these aspects into this AI tool and knowing this is just the start -wow!
Here’s what I’ll offer for consideration: I found that I was initially so wowed by what the AI was offering and its ability that when I actually really cued in and listened for solutions offered, there wasn’t much more there than what you had originally prompted. It summarized and/or restated differently without saying much more. I’m sure that will improve over time but it was my one-time observation.
Thanks again for sharing. You really are a brilliant person. 🤩

I say thanks, and there’s a reason the AI acted this way. The reason it did this is that it had been instructed to perform this in the very first sentence, i.e. “can you put this together into a coherent narrative?”

But, she is right in that I was never refuted. The premise was accepted easily. No push back. had the prompt been, “discuss this issue with me as though you are professor or professional journalist or television news host” that would have been a lot different.

Another friend said this:

I watched the whole thing. The music is cool. While I understand your friend’s observation that ChatGPT isn’t necessarily adding new insights, it’s still impressive that it can handle this level of complexity (and with a voice interface, no less). I also like the fact you can interrupt/redirect it midstream to augment your prompt.

The real test, and perhaps my next video segment, will be to see what it does when asked to provide counter arguments or to go into more of a dialogue or discussion mode. Will we have to wait for AGI for that? I’ll find out soon.

If you are curious to know how this will become part of my Summer workshop, you can visit the slideshow here. 


If you are interested in setting up a similar workshop, please sign up on my Calendly page

No comments: