As generative AI infiltrates the mainstream, prompt engineering appears to be the most voguish job. Yet, is it a sustainable career path or just a fleeting fascination? This article delves into the hype surrounding prompt engineering, scrutinizing its job market prospects, and providing a pragmatic perspective on its worth as a skill.
The hype around prompt engineering
Prompt engineering, which involves using natural language to extract useful content from AI models, has recently been spotlighted in mainstream media. Reports in prominent publications such as The Washington Post, Time, and Forbes have stirred the hype, projecting prompt engineers as 'AI whisperers' and suggesting lucrative salary prospects. Silicon Valley companies like Anthropic have further fanned the flames, seeking 'creative hacker spirits' for prompt engineering roles. Even influencers have joined the fray, depicting prompt engineering as a modern-day gold rush, ready for the taking for anyone willing to learn a few tricks.
Job market realities for prompt engineering
The reality of the job market for prompt engineering appears to be significantly less rosy than the media's portrayal. Six months after Anthropic's famous job ad, the demand for 'prompt engineers' seems more hype than substance. There is a notable lack of standalone prompt engineering positions in job listings, even in AI-friendly markets like the United States. Instead, responsibilities associated with prompt engineering are being absorbed into broader roles, such as machine learning engineers or AI specialists.
Domain experts and technical experts in prompt engineering
A closer look at prompt engineering reveals two distinct types of value creators: domain experts and technical experts. Domain experts, specialists in a specific subject, often excel at prompt engineering because they know the right questions to ask and can identify value in AI-generated responses. In contrast, technical experts, skilled in AI intricacies, can leverage their knowledge to improve overall results. While both types of prompt engineers are valuable, they serve different purposes and bring diverse skill sets to the table, making it crucial for organizations to recruit both if they're using generative AI at scale.
Determining the effectiveness of prompt engineering is a complicated process. The value of generative AI outputs, such as advertising copy or images, hinges on their real-world success, which is often subjective and hard to measure accurately. For instance, judging the success of an AI-designed ad campaign requires domain experts who can evaluate the campaign's impact. Furthermore, it is suggested that evaluation of generative AI should move closer to semiotics, connecting natural language to the real world, to help narrow the evaluation gap over time.
The value of learning prompt engineering is debatable. Because AI models are continually updated and replaced, specific prompting techniques that work today may quickly become obsolete. Instead of investing time and resources in mastering prompt engineering, individuals might find more value in developing their ability to pair AI with problem formulation in their domain of expertise. This approach not only enhances their domain expertise but also bolsters their aptitude for generative AI, possibly leading to a breakthrough application. Over time, the ability to prompt AI may become a basic skill, akin to using a search engine, making it less of a unique selling point on a resume.