The Treadmill Problem: Why Behavior Outweighs Tools
There's a belief that we should treat generative AI like a typical digital transformation.
You know the drill: Bring in new software. Offer awkward training videos. Burn down old software. Boom. Everyone’s using the new tool.
That approach falls flat with LLMs like ChatGPT. Why?
Because the hurdle isn’t learning how to use it. It’s not like teaching people how to use Excel. The hurdle in the case of LLMs is the behavior of the user.
Weird, right?
This is what I call the Treadmill Problem.
Imagine this:
You wanna get fit. You take the plunge and buy a treadmill. The moment you click "purchase," dopamine floods your brain. You’re already picturing your new beach bod emerging from the ocean in slow motion.
Transformation is guaranteed, right? Wrong.
The reason we’re not in shape has nothing to do with whether or not we own a treadmill, friends.
We want to think it is, because we're wired to believe that buying a product will magically change our behavior. Heck, QVC has built an entire empire on this notion.
But then the treadmill arrives, and reality hits. We hop on, and five minutes later, we're ready to tap out.
It's not because we're out of shape or short on time. It's because our brains are hardwired to 1) seek quick rewards and 2) conserve energy.
The treadmill does neither. Meanwhile, that couch and bowl of Doritos are singing a siren song to our limbic system.
The point is, we don't hop off and try to fix the treadmill. We don’t watch videos on ‘how to run on a treadmill.’ The problem is neither lack of knowledge nor is it the tool itself.
THE REAL PROBLEM: OUR BRAINS
Just like the treadmill, the problem with underutilizing ChatGPT is…yep, you guessed it. Us. It's our brains leading us astray.
Here’s what happens:
Your brain is conditioned for Google, and ChatGPT looks a LOT like Google. Our brain slips right into the whole command-and-response thing. Ask a question, get an answer, move on with your life.
But generative AI doesn't work that way.
Engaging with ChatGPT is more like picking the brain of Jeff Bezos about marketing strategies, or getting customized travel tips from the head of Costa Rica's tourism board, or learning from an exceptionally patient tutor.
It all comes down to your behavior. And rewiring that is tough. Because while the best results come from conversing with ChatGPT like it's human, our brains resist. It's like trying to chat with a baby as if it's a tenured professor. Our neural pathways keep yanking us back to old habits.
THE HTG PRINCIPLE: YOUR BEHAVIORAL CHEAT CODE
So, just like the treadmill, we have to recognize that ChatGPT will do its job just fine. We're the ones who need to change our approach.
If you're new to this, or you're trying to get your team up to speed (which, by the way, is deceptively challenging because there's not much to "teach" - it's like trying to teach weight loss by saying "eat less, move more"), I've got an LLM-hack to get you started.
You may have heard me talk about this from day one. But it still works. Let me tell you about the HTG Principle - Hi, Thanks, Great.
H is for Hi:
Start by greeting ChatGPT like you would a human. This isn't just about being polite; it does something neurologically. It establishes respect and rapport, which Social Identity Theory and the Cooperative Principle tell us lead to more productive dialogue. ChatGPT doesn't care if you say hi, but your brain does. Most importantly, you find it reacts very human like. Which gets you in the right mindset to start crushing this stuff.
T is for Thanks:
Thanking ChatGPT leads to linguistic matching, where you align your style, tone, and respect with your dialogue partner. This builds understanding and rapport in YOU. It helps rewire your brain to engage with ChatGPT using your "human interaction" circuitry, not your "Google command" autopilot.
G is for Great:
Provide instant feedback to guide ChatGPT's responses. If you ask for ideas and they're not quite right, don't just ask for more. Volume doesn’t help! Give specific feedback, like you would to a colleague. "I like this one because X, but I don't like that one because Y." These little iterations can have a huge impact and soon you’ll be weaving your way to amazing answers, leveraging your own expertise.
And there you have it, folks - the Hi Thanks Great Principle. Before you dive into your next prompt, remember: it's not just about the words. It's about putting yourself in the mindset that you're conversing with a human. Do that, and I promise you'll unlock so much more. (BTW, if anyone out there needed the disclaimer that ChatGPT is not a human, consider this sentence that disclaimer. Cool? Okay, it’s Dorito’clock around here.)
AI NEWS OF THE WEEK
Inflection’s Inflection Point
The AI world was rocked with the news that Inflection CEO and founder Mustafa Suleyman was leaving Inflection AI to help Microsoft start MS AI. What’s this mean for Inflection, which was getting some buzz?
Apple and Google, sitting in a tree?
In more news that caught the world off-guard, Apple is in talks to use Google Gemini as their AI engine. Maybe Apple wasn’t as advanced in this space as we thought?
The AI Pin, now with more stuff
Humane, makers of the AI Pin, released a new demo showing upcoming capabilities of the wearable device. The goal was to show how it can integrate into your life and change how you interact with tech. Super compelling stuff.
Generative AI Tips
LLMs like ChatGPT are phenomenal at mimicking human behavior. I would argue that they do this better than they do almost anything else.
For marketers, this is a HUGE boon.
I often teach marketing teams to try what I call The Infinite Focus Group. (I might write more about this later, I love it so much.) That’s when you ask it to create a focus group of a diverse group of folks to review whatever you want them to review. The responses are completely human, and you can modify the honesty, critical response level, anything. It’s an incredible time and money saver!
That's all for now, friends! See you next time.
