The AI vs. Human Hype is Overblown: Show Us the Proof, or It Didn't Happen

Humans beware, you're about to be asked to demonstrate advanced reasoning that drives your work. At least, that's what I'd be asking if I were hiring you.

I spent a long time developing prompts and fine-tuning them for creating advanced Jobs-to-be-Done qualitative modeling. Initially, I was trying to reduce the time it takes to produce a viable model from weeks to hours. But recently, I've begun to realize that these outputs may not be trusted.

Why?

Because incumbents gonna incumbent! That's why.

I spent about a year as a financial auditor way back when, and the one thing I learned to do was trace my work back to the source, so I had proof. It was mind-numbing work. But I understood the importance of it.

The state of GenAI today allows us to do something similar. Since I'm not asking these models for their opinion, what I need is a more logical breakdown of how it came to output a job map, or success metrics. Since I'm instructing the number of items outputted, and that they need to…