Hold your horses, AI enthusiasts, because the oh-so-hyped “job of the future”- prompt engineering – might just be a fleeting fancy. But fear not, for there’s a more enduring and adaptable skill waiting in the wings to help us harness the potential of generative AI: problem formulation.
Now, before we dive into the nitty-gritty of problem formulation, let’s take a quick look at why prompt engineering’s claim to fame might be short-lived. Firstly, future AI systems will likely get better at understanding natural language, rendering meticulously engineered prompts less necessary. Secondly, AI itself is on the verge of crafting its own prompts, making human intervention in this area somewhat pass. And lastly, the efficacy of prompts is tied to specific algorithms, which limits their utility across diverse AI models and versions.
While prompt engineering is all about crafting the perfect textual input with the right words, phrases, sentence structures, and punctuation, problem formulation focuses on defining the problem itself. And let’s face it, even the most sophisticated prompts won’t do you much good if you’re barking up the wrong tree.
You might be wondering why problem formulation hasn’t gotten the attention it deserves. Well, one reason could be the disproportionate emphasis on problem-solving, as evidenced by the misguided adage, “don’t bring me problems, bring me solutions.”No wonder 85% of C-suite executives think their organizations stink at diagnosing problems.
So, how can you become a problem formulation whiz? According to research and experience, there are four main components: problem diagnosis, decomposition, reframing, and constraint design.
Problem diagnosis is about identifying the core problem that needs solving. For instance, InnoCentive (now Wazoku Crowd) used the “Five Whys” technique to get to the root cause of the subarctic oil cleanup issue, leading to a breakthrough solution.
Next up is problem decomposition, which involves breaking down complex problems into smaller, manageable sub-problems. This approach proved successful in the InnoCentive Amyotrophic Lateral Sclerosis (ALS) challenge, where the focus was shifted to detecting and monitoring the progress of the disease.
Problem reframing is all about changing the perspective from which a problem is viewed, leading to alternative interpretations and potential solutions. Remember the GE Adventure Series MRI machines for kids? That’s problem reframing in action.
Lastly, problem constraint design focuses on setting the boundaries of a problem by defining input, process, and output restrictions. This can be particularly useful when directing AI to generate valuable solutions for specific tasks.
In conclusion, although prompt engineering might be hogging the limelight for now, its limitations become apparent when compared to the sustainability, versatility, and transferability of problem formulation. So, instead of obsessing over the perfect combination of words, let’s focus on mastering problem formulation to truly unlock the power of AI. It might just be as game-changing as learning programming languages was in the early days of computing.
