Right now we’re in an amazing time. AI and ML have been around for a long time; but ChatGPT made the non-technical world sit up and take notice. And now LLMs are taking the world by storm. Within the next 10 years, everyone will be using them. IMO, right now, prompt engineering is an essential skill to learn for keeping up with the information tidal wave surrounding LLMs as it continues to evolve.
Non-Technical Users
The average person should learn prompt engineering to use LLMs more effectively. Even simple Persona patterns of prompting (e.g., You're an expert marketer. Write me some marketing copy for X. ) can make a LLM far more effective than a raw, basic prompt ( Write me some marketing copy for X ).
Software Engineers
Conventional web software engineers (those building web apps, mobile apps, APIs, and so on) can and should be using LLMs to generate code and reduce repeated work. Quite often, they’ll be asked to build features using LLMs, so prompt engineering will be at the center of making these features effective. In addition they’ll need to learn all the logistical implications of doing this - how to test and measure the effectiveness of LLM-driven features, how to host and scale them, and so on.
ML Engineers
Even machine learning engineers should learn prompt engineering. Why, when they already specialize in the science that makes LLMs run? Because prompts are the user-facing interface by which their customers use the things they’ve built. In the same way that it’s beneficial for web developers to have at least a sense of what constitutes good UX or bad UX, it’s valuable for ML devs to know a little prompt engineering so they can anticipate how their users will interact with their work so that they can make life easier for them. ML engineers also will gain the benefits that non-technical users get from learning a little prompt engineering; and more effective use of LLMs for their work and personal lives.
Will prompt engineering be obsolete?
As of right now I don't think so. I think it will definitely change a lot, and some techniques will become unnecessary as models become more intuitive. Some prompting patterns will become defunct, and others will emerge as LLMs continue to evolve and change. But in the same way the English language (and all others) don't change frequently, I think the way we interact with LLMs will
At its core, prompt engineering is becoming a crucial skill across various domains, from everyday users to specialized professionals. As Large Language Models (LLMs) continue to reshape our digital landscape, mastering the art of crafting effective prompts will not only enhance personal productivity but also drive innovation in numerous fields, making it an indispensable tool for navigating the AI-driven future.