AI Futures: An Anthology Excerpt. Part 3 (of 4)

“The Great RenAIssance” Approaches: Why 2023 Was the New 1995

Surreal Self-Portraits: My Gateway to AI

I’ve been curious about AI for the past decade or so, tantalized by promises of cures to cancer and big AI wins like Big Blue finally taking down Gary Kasparov in 1996, hinting that something big was afoot. And yet, AI remained the purview of the Stanford mathematicians and Silicon Valley engineers. While I’m geeky, I’m not a geek. Intrigued by the possibility of AI, I would try to dip my toe in the waters every year or so when Google would tease a new “much easier” version of Tensor Flow which promised to make AI accessible to the masses. And like Charlie Brown, I would run toward Lucy and that tempting football, ready to finally kick it down field and POOF, a kick and miss. Machine learning and AI was simply not at an inflection point where non-engineers could explore it.

In early 2022, after a deep dive into blockchain and NFT projects the year before, I gained access to an art project called Pixelmind which put a simple interface on top of Stable Diffusion (an AI image generator from Stability AI) and used a “mad libs” style prompt to create magical images with zero technical knowledge or artistic talent. It was pretty magical and eye-opening. You could even watch videos of the AI dreaming your images into existence, step-by-step. For months I would cycle in and out of making 100’s of images in a few days, sometimes marveling at how empowering it was, and other times wondering what the point was. As it often does in my life, my curiosity got the better of me and I wanted to understand what was actually happening here.

It was time for me to see if I could figure out how to get Stable Diffusion up and running on my own. The Stable Diffusion subReddit was my tutor (a pretty crappy one if you don’t already know your way around Github and similar code repositories.) My instinct told me it was worth the effort, and I learned enough about Google Co-Pilot and it’s “simple notebooks,” to understand that I was woefully unprepared to get things working. I failed NINE TIMES to get Stable Diffusion up and running... Not "nine tries" one night. Nine separate attempts over a three week period before I finally got it up and running!

Eventually my stubbornness won out and I figured out how to make images, and even to upload pictures of myself to create surreal self-portraits that seemed to be from some parallel universe where I made some very different life choices. I got enthralled by what I was creating and ended up creating an art project called Kyle Shannon Dreams (instragram.com/kyleshannondreams & tiktok.com/@kyleshannondreams) where I could bring these characters to life. The experience was eye opening and I found myself producing creative content significantly more accomplished than the amount of effort I put into it. Something was different here. Something was shifting within me.

This era wasn’t just about incremental advancements in existing technology; it was about a paradigm shift, opening doors to capabilities that were once the stuff of science fiction. This experience was more than just technological experimentation; it was a glimpse into a future where AI could augment and amplify human creativity. The more I explored AI in image generation, the more I realized its potential to revolutionize the way we perceive and interact with digital content. It wasn't just about the technical prowess of these tools but their ability to redefine the boundaries of artistic expression.

The Launch of ChatGPT: “Well, this Changes Things!”

ChatGPT is doing for AI and Machine Learning what the World Wide Web did for the Internet. It made the power of decades of research, development and infrastructure instantly accessible to “the rest of us.” The Internet had been around for decades and while it was powerful and transformative, the user base was mostly researchers and engineers and academics. Adding a “point and click” interface via the HyperText Transfer Protocol (HTTP) meant that non-scientists could use the Internet as a tool for self- expression. The hyperlink, as simple as it seems from a technical perspective, is the key breakthrough that democratized the power of networked computers.

ChatGPT does the same thing for AI. Even though GPT3 and the OpenAI Playground (a tool for developers and researchers) have been around for a few years, ChatGPT reveals the power of AI via an interface as simple as the Google search engine. When people look at it, they instinctively know what to do. They don’t have to read the manual. (Which is good because there isn’t one.)

The OpenAI team that created ChatGPT didn’t think it was a big deal. They didn’t think anyone would even notice it. Why? GPT3 had been out for some time, and the instructions to build ChatGPT had been in the documentation, and no one had bothered to build it. When Microsoft was about to announce their first $1 billion investment in OpenAI, Sam Altman suggested the engineers at OpenAI go ahead and build ChatGPT so they had something to show that wasn’t so technical. They had a handful of weeks to build it. Not a big deal. They had drinks the night before the launch and thought maybe there’d be a blog post or two and a few thousand comments. About that...

Within 5 days of launch, ChatGPT had 1 million users. Within 6 weeks, it had 100 million users. It was the fastest adoption of technology in history and was entirely driven by word- of-mouth. It’s easy to underestimate the impact of a simple interface. There were no “switches and dials.” You didn’t have to “choose an AI model.” You simply typed a prompt into a simple text input box and what seemed like magic emerged. In Arthur C. Clark’s 1962 book "Profiles of the Future," he posits, “Any sufficiently advanced technology is indistinguishable from magic.” That’s what ChatGPT feels like. Even the scientists and researchers don’t fully understand how Large Language Models do what they do. They call unexpectedly successful results—such as when it can accomplish something they didn’t explicitly train it on—“Emergent Behavior.” Yeah, call it what you want, it’s magic.

When I first started playing with ChatGPT, I did what I now refer to as “Parlor Tricks.” I would toss ridiculous challenges to it to see if I could break it. ‘Create a blog post about the impact of authentic video on building customer trust.’ BOOM. Done. ‘Create a table comparing gummy bears to sushi.’ BOOM. Done. ‘Explain quantum mechanics to me in an Eminem Rap.’

(Verse 1)

Yo, check it, we're diving deep, quantum mechanics in play, Tiny particles in the game, acting in a bizarre way. Not like the physics that's classic, they got a different story, Talking 'bout atoms and photons, in their quantum glory.

Wave-particle duality, it's a crazy notion, Particles acting like waves, in quantum motion. They could be here or there, sometimes they're in between, In the world of the very small, they're rarely ever seen.

(Chorus)

Quantum mechanics, it's the rule of the small, Unpredictable and wild, it's got us all enthralled. With probabilities and mysteries that we can't ignore, It's a rap about the quantum, gonna explore some more.

(Verse 2)

Heisenberg's uncertainty, it's a fundamental limit, Position and momentum, can't know both, ain't no gimmick. The more you know one, the less you know the other, It's a quantum world, unlike any other.

Entanglement, it's a spooky kind of action, Particles connected, no matter the fraction. Change one, the other reacts, no matter the distance, Quantum entanglement, defying our existence.

(Chorus)

Quantum mechanics, it's complex, it's unique, In the microscopic realm, it's the physics we seek. From superpositions to wave functions that evolve, In the quantum world, there's so much to solve.

(Bridge)

Yo, it's not just theory, it's technology in action, Quantum computing, it's gaining traction. Bits to qubits, processing with speed, Solving problems classical computers can't heed.

(Outro)

So that's the story, the quantum rap game, Complex and wild, never tame. It's the physics of the small, the foundation of tech, Quantum mechanics, deserves respect.

BOOM.

The first thing that struck me as I interacted with ChatGPT was that I should learn more about the Turing Test (you know, the one where it passes the test if you interact with a computer and can’t tell if it’s human or not,) because it felt a lot to me like we were suddenly beyond that. In trying to get ChatGPT to act like the computer systems from SciFi movies of my youth, I realized it was already acting like them... or better. I wasn’t having to think logically about how to “program the system,” I was simply talking to it, and it was responding... like a really smart person.

My true breakthrough moment with ChatGPT happened a week later, when I was absent-mindedly playing with it while my Wife and I were watching TV. I asked it to write me a Python script. (It can also code if you didn’t know that.) Sure enough, it very quickly generated code that looked impressive. To be clear, I’ve never written a line of Python in my life. After it generated the code, I realized, “I don’t know how to run Python code.” So I confessed to ChatGPT my ignorance and rather than judging me, it patiently walked me through how to get it up and running, and then it helped me troubleshoot bugs, and even suggested other things I could make the code do. Within 90 minutes, I had written a FULLY FUNCTIONAL python application that generated original social media content (5 different types) based on user input! This wasn’t just ‘hello world.’ The app was interacting programmatically with OpenAI’s GPT3 API! (Application Programming Interface, the thing that let’s my app talk to other computer systems.) I was now a Python programmer, and I’ve STILL never written a line of Python in my life.

“The world just changed, but no one knows it yet.”

OMG, here we go again! This is the identical epiphany I had about the World Wide Web in 1994. My heart started racing. My mind was flooded with endorphins and fear and euphoria and recurring variations of “Holy CRAP! What does this even mean?!” As I delved deeper into the capabilities of ChatGPT, I realized that we were standing at the precipice of a new era in human-AI interaction. This technology was not just a tool; it was a harbinger of a future where AI would play an integral role in our daily lives, from simplifying tasks to enhancing our creative expressions.

To be continued. CLICK for Part 4

AI Futures: An Anthology is a exploration at the intersection of artificial intelligence and human creativity. This collection brings together a diverse group of humans ranging from teachers to storytellers, civil service workers to writers, technologists to thinkers, creatives, engineers and futurists - all united by their early adoption and experimentation with AI tools. This anthology showcases the stories of these pioneers who delved into the world of generative AI not as technologists but as novices venturing into unknown territories.

The book features essays from: Lee Chazen, Cyndi Coon, Dr Wendy Elford, Camilla Esser, Peter Kaminski, Michael Lennon, and Kyle Shannon.

Next: Part 4 (of 4)

"The 'Great RenAIssance' has begun." and "AI Makes Humanity More Important, not Less"