How AI Works Brilliantly for Me

I find AI immensely useful. I think most people do. Yet the complainers dominate the discussion. Two years ago it was about hallucinations, now it is about offing humankind. Read my positive perspective riffing off Annie Get Your Gun’s song from 1946.

“Anything You Can Do, I Can Do Better.” I am the you and I is the AI. So it becomes “Anything I Can Do, AI Can Do Better.”

I say this because I find almost infinite usefulness in collaborating with AI. My favorites are ChatGPT and Gemini.

I am an expert user so I know when not to use AI. I would never allow it to create a storyline or draft a deck or email. But when it comes to my main task in R&D, bridging science and management, AI can do what no human I have worked with can.

I will initiate a chat based on economic or financial theory, point to a few academic papers, share my rough hypotheses, and off we go. It is a back and forth stretching over days, weeks, or even months (not full time).

In the end, I have a full substantive practical piece of work that I can then test against real world data.

Examples, the philosophy (epistemology) of model validation (not the junk of R-squared etc.). The equations behind relative market share (not the junk people usually use). The true size of companies (not the revenue junk that, e.g,, Fortune presents). The productivity of companies and countries (not junk like output per worker). How to analyze attitude & usage surveys (not the cross-tab junk).

These examples have real-world implications. Doing thing the amateur way leads to errors, being scientific reduces those errors.

Thus, bravo AI for increasing my productivity.

---

People often talk about the drawbacks of AI as a starting point. That is garbage thinking. One has to start with the prime: in this case that AI works, only after this can you talk about the secondary.