THE USE CASE TRAP: WE’RE TEACHING GENAI WRONG
There is a universal misconception that the best way to learn genAI is through use cases.
“Tell us the use cases for our industry/organizational function!” I get this all the time when running genAI workshops for organizations.
I would scurry to prepare the Ten Best Use Cases for the client. And guess what? It worked like a charm!
“This will transform our business!” senior management would cry, tossing me in the air and chanting my name.
But.
I would learn later that it wasn’t actually transforming their organizations.
Here was the issue - when I checked back in with them, I found they would still be using iterations of the use cases I gave them.
Which is bad, because I’m not the expert in the client’s business.
It took me a long time to discover the problem. This is what I found:
Learning from use cases - copying prompts, etc. - is instant gratification. The brain loves that.
The real power of ChatGPT, though, is innovating and iterating with it.
That means we have to change the way we learn.
PHRASEBOOK VS FLUENCY
Learning ChatGPT through use cases is like trying to learn French with a phrasebook. But way worse.
A phrasebook is, if you think about it, use cases of a language. It’s incredibly useful! It can help you buy eggs and inquire (frantically) where the nearest toilet is.
It’s how to survive with a new language.
But you can’t innovate off a phrasebook.
Compare somebody with a phrasebook to somebody speaking French fluently. It's a wildly different experience.
That person speaking French - even at a rudimentary level - can begin to communicate what they actually need. So instead of being stuck with “Where is the bathroom” you can build on your basic French to say “Where is the fire extinguisher - please tell me very quickly.”
The best part? Unlike learning a new language, where your brain resists the energy required, ChatGPT is more like learning how to use your iPhone. Insanely intuitive.
Better still, the dopamine hits come fast and furious. You’ll be off and running in no time.
2. CHATGPT REACTS TO YOUR IMMEDIATE NEED, NOT USE CASES
Here’s a thought experiment: Imagine clicking on a YouTube video for How to Install a Sink. Let’s say it has 9000 views.
How many of those 9000 viewers do you think were trying to install a sink in that moment?
Hint: 100% of them.
When was the last time somebody said "Copy and paste this prompt!" and you actually did it?
The odds that you, reading that prompt, actually NEED that prompt in that moment, are basically zero.
Not that it’s not helpful as an example, it is. But by the time you need that use case, that LinkedIn post is buried deeper than wherever the US government is keeping UFO wreckage.
So why do we love it?
Because of FOMO. For the same reason we save lists of AI tools. We feel like we might need it sometime. The brain clings to that due to its survival instincts.
But ChatGPT doesn't work like that.
It's like me ringing your doorbell and saying "Hey I have a way of fixing that whirring noise in your air fryer!" You're probably not going to need my help at that moment, and when you do, I’m long gone.
Let’s state the obvious:
You need help with whatever it is you need help with.
That’s why you must create your OWN use cases. That requires fluency, not a phrasebook.
You want to be a highly trained handyman as opposed to a highly specialized guy who can only fix chair legs. When something breaks in your house, you're going to be able to fix it, no matter what.
Forget use cases. We have to learn how to use ChatGPT by applying it systematically to every task you have and seeing where and how it works. That's how you get that fluency. That's how you begin to use it for whatever pops up.
3. STOP HOARDING USE CASES - TEST LLMS WITH YOUR DAILY TASKS
Forget about hoarding use cases like a squirrel hoarding acorns (that’s a thing, right?). That's not how you master ChatGPT.
Here's what you do instead:
Every task that comes across your plate, ask yourself: "Can ChatGPT help with this?"
Then dive in. Try it with that task. See what works and what doesn't.
It's like one of those choose-your-own-adventure books. (Remember those? No? Cool.) Every prompt leads you down a different path. Sometimes you end up in a dead end. Sometimes you discover a hidden treasure.
That's how you unlock truly innovative applications. Not by copy-pasting somebody else's prompt, but by exploring uncharted territory.
YOUR NEXT STEP: PUT DOWN THE PHRASEBOOK
Forget the phrasebook. We’re gonna Matrix ChatGPT into your brain, without the pain and suffering of French 101. Get started, task by task. Subject everything you do to ChatGPT, one thing after another.
Your use cases will start flowing fast and furious.
Then post them all on LinkedIn. People love that stuff.
AI NEWS OF THE WEEK
1. OpenAI + Robots
If you haven’t yet seen the video of the new Figure 01, you gotta check it out. Figure is the robotics company that’s gotten investments from Nvidia, Bezos, and Microsoft. This is happening fast.
2. Elon goes open source
Elon is suddenly making his AI chatbot, Grok, open-source. He almost had to, after suing OpenAI over becoming a for profit company. Where Elon is going with all this is a mystery.
3. Sam is back on board
Lots of OpenAI news this week! Sam Altman has rejoined the OpenAI board. An independent investigation says that while the board had the right to fire him, they had no cause. Sam is more powerful than ever.
Generative AI Tips
If you haven’t used ChatGPT Voice, you’re missing out.
One of my best thinking times is when I’m driving. In the past, I would open Apple Notes and dictate my ideas.
Now, I process it with ChatGPT Voice - available with a Plus subscription. I tell it everything I’m thinking, ask for feedback, ask questions, etc. ChatGPT talks back and records the entire conversation. Then I ask it to summarize everything into, for example, an draft of a post or a newsletter.
It’s not just a time-saver; it’s a new way of processing and activating the ideas that strike you at the inopportune times.
That’s all for today!
Thanks for reading, and share with anyone you think can benefit! We gotta learn together!
See you next time!
