Pobody’s Nerfect – That Goes for LLMs, Too.

When discussing AI engagement, I encourage people not to marry into an LLM. Experimentation and testing are key–even if you are experienced. These models advance and leapfrog each other every day.

Case in point:
Today, I was working with my dear Gertrude (ChatGPT) on some tasks we perform regularly (data reorganization). My prompt is solid. We’ve completed these tasks successfully before. Nothing changed on my end.

But today, the results just weren’t there. I had to correct ChatGPT several times, and the errors persisted. Eventually, the time spent outweighed the time I should have saved.

My frustration was mitigated because I knew I could pivot to a different model. However, I will retest this with ChatGPT at a later date.

Subscription models (by the way, please pay for the subscription model) currently have a price point that’s fairly easy for most users to swallow. There’s no reason right now to enter into a monogamous relationship with an LLM.

Have you ever had an experience in which you thought your AI took a sick day? Did it break your heart?

Keep playing – keep testing – keep experimenting!