Prompt engineering, one of the buzziest jobs in 2023, becomes obsolete as AI models better intuit user intent and companies train staff on how to write prompts
Prompt engineering, a role aimed at crafting the perfect input to send to a large language model, was poised to become one of the hottest jobs in artificial intelligence. Bluesky: @jordisoler , @tappedlands.com , and @skynetandchill.com . Mastodon: @harrymccracken@mastodon … LinkedIn: Ariel Guersenzvaig and Isabelle Bousquette Bluesky: Jordi / @jordisoler : It's becoming increasingly difficult to predict the jobs of the future. [embedded post] @tappedlands.com : “Trendy, ultimately meaningless job proves meaningless” [embedded post] @skynetandchill.com : Prompt engineers replaced by AI. You really just need an hour to learn basic prompting, and then how to use AI to optimized prompts. Mastodon: Harry McCracken / @harrymccracken@mastodon.social : It did feel a little like hiring people to Google professionally in 1998. https://www.wsj.com/... LinkedIn: Ariel Guersenzvaig : Guess what they are talking about. If you guessed prompt engineering, you are right. “What happened?” they ask. What happened? Really? … Isabelle Bousquette : Do you actually know anyone with the job title “Prompt Engineer”? I don't. And neither did a number of tech execs I talked to …
Context & Ripple Effects
Prompt engineering emerged as a distinct specialty when AI companies were hiring people to refine prompts and expose chatbot weaknesses. By 2024, research had already suggested that models themselves could take over much of that optimization work through LLM-led prompt engineering.
This report marks the labor-market consequence: improved intent inference and broader employee prompt training reduce the rationale for maintaining a standalone role. It also sharpens a product tension visible in complaints that some AI tools remain constrained by inaccessible system prompts.
First-order effects
- Dedicated prompt-engineering roles lose differentiation as model behavior requires less hand-crafted input and prompt-writing knowledge is distributed across ordinary staff.
- Employers can shift prompt work into existing product, operations, and domain-expert roles rather than staffing a separate AI-input specialty.
Second-order effects
- AI vendors face more pressure to make intent handling and workflow defaults reliable, since customers will increasingly judge tools on outcomes rather than on users' ability to phrase instructions.
- Training demand shifts from specialist prompt craft toward organization-wide AI literacy and role-specific guidance, while teams that still need control may focus on system-level configuration rather than user prompts.
Third-order effects
- If this pattern persists, AI-related jobs may increasingly center on embedding models into accountable workflows and supplying domain judgment, not on a transient interface skill tied to a particular model generation.
- The shift could widen the gap between consumer-simple AI interfaces and enterprise systems that retain governed, editable instructions for reliability and oversight.
The trend: This is one data point in workplace AI's move from standalone interaction skills toward embedded, workflow-native systems that absorb more of the interface complexity.