Continuing the reflection about my Artificial Intelligence experiment (see dot #4290), I knew enough from my minimal training to have some idea of how to prompt ChatGPT to give me usable content. And the more I prompted it, the more refined my prompts became.
For example, I started with: “Write a 150-word blog post in the style of leadershipdots.com that offers advice to new supervisors about how to transition from being an individual contributor to becoming a supervisor,” but I received something that read like it was from a dissertation. So, I triggered another version with: “Re-write the above post to be more informal,” and ended up with what you read as dot #4286.
By the time I was on my third dot (#4288), my prompt read: “Write a 150-word informal blog post in the style of leadershipdots.com that offers advice to supervisors about how to deploy free resources at their disposal to increase employee engagement. Include an example and unique analogy.” In addition to machine learning, it was simultaneous human learning.
And if I didn’t like the output, I could merely type: “Try again,” and I would instantly have another version. It was as magical as I’m sure the early programmers felt when their stack of keypunch cards calculated equations faster than humans could.
Back in those days when computers were first being programmed, the common mantra was GIGO — garbage in, garbage out. The 2024 version of GIGO remains the same, only it applies to prompts — the modern-day computer code.
In the dark ages of DOS computers, I was an expert on writing commands for the mainframe system to pull specialized reports generating enrollment numbers in increasingly refined subsets. To me, writing prompts is a very similar skill. The more nuanced your query, the more usable your output. I think it’s where I came to identify the Superman “S” with specificity — the ability to be granularly specific about identifying what you want to produce is a superpower.
Writing prompts is becoming so essential for businesses to harness the power of AI that J.P. Morgan announced all incoming employees will be required to learn AI Prompt Engineering as part of their orientation, and the company is training 80,000 current employees on the skill as well. AI is permeating the banking world — and just about every other aspect of life.
Writing is just one tiny fraction of what AI can do. Maybe it’s time for you to play around on your own or to take one of the many online tutorials that abound. The sooner you learn to think in the format of how prompts are written, the sooner you will have access to the unlimited resources that prompts unlock.
One last dot on my AI experiment tomorrow…
