It’s critical thinking, not prompt engineering that we’re in need of.

In his article, AI Prompt Engineering Isn’t the Future , Oguz A. Acar makes a case that prompt engineering is going to be a very transient need. Instead, he says that problem formulation is the real long-term need when working with AIs. I tend to agree. As LLMs continue to improve it gets easier for them to intuit our intentions–even from poorly designed prompts. What they can’t do, however, is tell us how to describe the problems we need help solving. This is where critical thinking becomes so important. We need to be able to articulate problems, break them down into subproblems, reframe problems to explore alternative solutions, and play with the problems constraint space. All good engineers will be familiar with these concepts. Perhaps some software engineers should considering moving up the “problem stack” ,so to speak, to work in this area.