Loop engineering is not just a new job title. It is a signal that the first generation of AI work — writing clever prompts — is already being treated as obsolete inside the companies building the tools. That shift has consequences far beyond Silicon Valley job boards.
When Anthropic co-founder Boris Cherny says the era of manual AI prompting is ending, he is speaking from a position of authority. His company makes Claude, one of the most widely used AI assistants.
He also created Claude Code, a coding tool. Cherny previously declared that “software engineering is dead.” Now he is pointing at what comes next. That next thing is loop engineering.
The difference matters for anyone who uses AI at work. A prompt engineer writes a single instruction.
A loop engineer builds a system that runs multiple steps on its own. The report gives a concrete example: a prompt engineer asks an AI to summarise a legal document. A loop engineer builds a system that reads the document, identifies ambiguities, searches relevant regulations, reviews its own conclusions, and flags uncertain cases for human review.
That is a fundamentally different kind of work. It is less about talking to the machine and more about designing a process the machine can run. This changes who gets hired.
The report states that the best loop engineers are often not the people who know the most about AI. They are the people who understand a domain — media, law, medicine, finance, software development.
That is a direct challenge to the idea that AI expertise is the most valuable skill. A lawyer who understands how legal documents actually get used may be a better loop engineer than a computer scientist who has never seen a contract. The practical effect is that companies will start hiring domain experts into roles that look more like process designers than AI whisperers.
Job listings already show terms like “forward deployed engineer” and “loop engineer.” These are not buzzwords. They reflect a genuine shift in how AI systems are being built and deployed, according to people inside the industry.
For workers, this means the window for “prompt engineering” as a standalone career may be closing. The skill is being automated into the system itself. Cherny’s point is that manual prompting is a transitional phase.
Once AI becomes agentic — capable of taking multiple steps on its own — the challenge shifts from writing better prompts to designing better processes. That shift will hit different industries at different speeds. The report mentions media, law, medicine, finance, and software development as domains where loop engineers will be needed.
Each of those fields has deep, complex workflows that AI cannot yet run end-to-end. The loop engineer’s job is to encode that expertise into a repeatable system.
That is harder than it sounds. It requires understanding not just what the AI can do, but what the human expert actually does. What to watch next: watch which companies start hiring loop engineers and how they describe the work.
Watch whether universities start offering courses in loop engineering or whether domain experts just learn it on the job. Watch whether the term sticks or gets replaced by something else.
The vocabulary is moving fast. Terms like “prompt engineering” entered the mainstream a couple of years back. Now they are being pushed aside.
The core fact is simple: the people building the AI say the era of manual prompting is ending. That means the work of using AI is changing. The question is whether the rest of the economy catches up before the next shift arrives.
{“@context”: “https://schema.org”, “@type”: “NewsArticle”, “headline”: “Anthropic co-founder declares end of manual AI prompting era”, “description”: “Loop engineering is not just a new job title. It is a signal that the first generation of AI work — writing clever prompts —…”, “datePublished”: “2026-06-26T04:07:03+00:00T09:30:00+00:00”, “author”: {“@type”: “Person”, “name”: “Merry Gel Sigui”}, “publisher”: {“@type”: “Organization”, “name”: “InfoPulse Today”, “url”: “https://newsfeed.infopulsetoday.com/”}}
