So Long “Prompt Engineering,” We Hardly Knew Ya

🌈 Abstract

The article discusses the declining importance of "prompt engineering" in the field of AI, and argues that the more valuable skill for the future is "problem formulation". It explains that as AI models become more advanced, they can generate their own prompts and work around the limitations of human-crafted prompts. Instead, the author suggests that the key skill for humans is to focus on identifying, analyzing, and defining problems that AI can then solve.

🙋 Q&A

[01] The Decline of Prompt Engineering

1. What are the three main reasons why prompt engineering is facing a "cruel sunset"?

• AI is getting smarter and can grasp language like humans, reducing the need for finely tuned prompts
• AI is already capable of crafting its own prompts, with GPT-4 and future models having this capability baked in
• Prompts are custom-made for specific AI models and have limited versatility, which AI can work around faster than humans

2. What is the author's view on the role of AI compared to "techbro billionaires"? The author is of two minds about AI - while they like the educational and helpful capabilities of GPT and LLMs, they are wary of how quickly AI is being pushed into our lives without input or discussion. The author trusts AI more than the "techbro billionaires" who are manipulating it and us.

[02] The Importance of Problem Formulation

1. What are the four ways the author suggests to examine and define problems?

• Diagnosis: Discovering the problem that AI can solve by asking the right questions and looking at different perspectives
• Decomposition: Breaking down big problems into smaller, bite-sized ones that AI can help solve
• Reframing: Shifting perspectives and seeking new interpretations to identify hidden solutions
• Constraint design: Setting boundaries and parameters for the solution that AI can understand

2. Why does the author believe problem formulation is more important than prompt engineering? The author argues that without a well-defined problem, the best-crafted prompt is just "a beautiful set of words". Problem formulation is the "bigger picture" and "broader strokes" that provide the foundation for AI to provide efficient solutions.