Why Prompt Engineering Was Always Temporary

In 2023, Anthropic posted a Prompt Engineer job with a $375,000 salary and no requirement for a computer science degree. McKinsey had reported that 7% of organizations were hiring Prompt Engineers. The Wall Street Journal called it the job of the future. In 2025, the role has effectively disappeared (Salesforceben, 2025). 

Job postings for "Prompt Engineer" peaked at 144 searches per million on Indeed in April 2023. By mid-2025, that number had collapsed to 20-30 per million. Microsoft's research ranked Prompt Engineer second from last among new roles companies were prioritizing (Salesforceben, 2025).

The reason Prompt Engineer died is mechanically simple. AI systems got good enough to optimize their own prompts. They ask follow-up questions, iterate on their own outputs, and, in a sense, run their own prompt engineering internally. The knowledge that made Prompt Engineers valuable became something the AI systems themselves performed better than humans.

Prompt Engineer was never supposed to be a permanent role, however. It was a transitional artifact. The gap between AI capability and human understanding of how to use that capability created a niche job. Once the gap closed, the job closed with it.

The same pattern is happening to dozens of emerging roles right now. Specialized AI Trainers are being consolidated into broader "AI Manager" positions. AI Data Specialists are being absorbed into data engineering teams. The roles that existed to translate between humans and AI are disappearing as AI becomes native to existing professions (Salesforceben, 2025).

Meanwhile, the roles that are growing are completely different from what was predicted. Demand for AI fluency jumped sevenfold in two years through 2023 to 25 (McKinsey, 2025). But AI fluency is not a job title. It is a requirement. It is the baseline skill people in all jobs need to have.

On the other hand, skills that were supposed to be in high demand are facing decline. Programming in specific languages is declining because AI can code in any language. Data entry is being automated. Accounting processes are becoming routine. Specialized knowledge that took years to acquire became obsolete in months.

The occupations at highest risk of displacement are computer programmers, accountants, legal assistants, customer service, and data entry (Goldman Sachs, 2025). These are supposed to be white-collar jobs with stability. Instead, they are the first to be automated because their work is routine enough for AI to learn.

The occupations at lowest risk are, strangely enough, air traffic controllers, chief executives, clergy, and photographers (Goldman Sachs, 2025). So, the jobs that require judgment, taste, accountability, or human presence are not being automated, while jobs that require efficiency and consistency are being automated immediately.

A study reported that 84% of companies say they will replace or outsource jobs if staff do not have the right AI skills (Pluralsight, 2025). But what are the right AI skills? Not prompt engineering. Not data specialization. Is it, then, the generic ability to work alongside AI systems?

Organizations are already acting on this. IBM plans to replace 30% of back-office roles with AI within five years. BT is cutting 55,000 jobs and replacing 10,000 with AI (Tech.co, 2025). These companies are starting to run their business model on the assumption that AI can do what humans do, cheaper and faster.

Estimates say that 170 million new jobs will be created by 2030 while 92 million will be displaced (The Interview Guys, 2025). Net positive. But that net positive assumes people displaced from accounting will become AI ethicists. It assumes data entry workers will retrain as AI security specialists. It assumes people will move geographically, learn new skills, and compete for roles that did not exist even two years ago.

Udemy's CEO says the solution is "skill fitness." Learning never stops and everyone becomes a perpetual student (GovTech, 2025). 

Otherwise, what is happening to Prompt Engineering will happen to every role that sits between humans and AI. Once the gaps close, the roles will disappear. Demand shifts to roles that do not exist yet. Workers who learned Prompt Engineering in 2023 are looking for new jobs in 2025 because their entire specialization was a temporary artifact of an immature, unpredictable technology.