I’ve said before that prompt engineering is a dead end, and here’s further support, for similar reasons.
- Future generations of AI systems will get more intuitive and adept at understanding natural language, reducing the need for meticulously engineered prompts.
- New AI language models already show great promise in crafting prompts.
- The efficacy of prompts is contingent upon the specific algorithm, limiting their utility across diverse AI models and versions.
A more “enduring and adaptable skill is problem formulation — the ability to identify, analyze, and delineate problems.”
I agree, although even this may only be true in the relative a short-term. I’ve mentioned before the work being done at the Allen Institute, where language models are being trained to generate new scientific hypotheses. From my perspective, this is a form of problem formulation that can be tested.
The author goes into the ‘problem formulation’ skillset in some detail, and I think there’s value there, especially for a teacher looking for ways to get students thinking about AI in their disciplines.