Introduction
A high school student recently asked a machine learning engineer whether studying technology still makes sense in an era of artificial intelligence. That conversation led to a broader examination of where AI employment is heading. The evidence points beyond Silicon Valley: the most significant opportunities may arise in factories, mines, hospitals, and other sectors that have only just begun to adopt intelligent systems. This article presents five predictions for the next five years, with a review point set for 2031.
What Happened
The software employment landscape has already shifted. Job listings for software roles on Indeed fell by roughly three-quarters from the 2022 peak to the 2025 low point, while civil and electrical engineering postings remained well above 2020 levels. Interest rates and post-pandemic over-hiring explain part of the decline, but AI appears to be why software hiring has not bounced back like other engineering fields.
Software development is uniquely exposed to automation in part because it is the most documented profession online. Nearly 400 million public repositories serve as training data, and programs can be verified instantly through compilers and test suites. That gives AI models a clear feedback signal. Because code is essentially text, transformer architectures process it effectively. Performance on real GitHub issues climbed from roughly 4% in late 2023 to 79% by the end of 2025.
At leading AI labs, the transition is already visible. Anthropic reported that more than 80% of code merged in May 2026 was produced by Claude. Google says about three-quarters of its new code is AI-generated and reviewed by engineers. Across GitHub, more than 40% of pull requests now begin as agent-created drafts. Still, many developers remain skeptical; fewer than one in three developers say they trust these tools, and controlled experiments have yielded mixed results on whether AI helps experienced developers work faster.
The developer role is changing rather than disappearing. Government labor projections still forecast growth for software developers through 2035, though the outlook has been revised downward since earlier estimates. Positions limited to turning specifications into code are expected to shrink. Meanwhile, machine learning engineering and forward-deployed engineering are expanding. Junior roles are scarce; early 2026 figures placed entry-level software postings at about 4.5% of all U.S. listings.
Beyond software, investment is flowing into robotics. Robotics startups raised a record $15 billion in 2025, and funding had already reached $18.8 billion by mid-2026. Funding for humanoid robots alone rose from $4.8 billion in 2025 to $11 billion in 2026. Prices are falling quickly, with some small humanoid models now costing less than a new smartphone. One driverless ride service logged half a million trips per week in April 2026, and synthetic data pipelines can produce in hours what would take thousands of hours of human demonstration.
Non-digital industries remain the biggest untapped market. About one in five companies in the United States and European Union say they use AI in some capacity. In construction, transportation, hospitality, and mining, usage sits between 7% and 15%. Commonly cited barriers include poor data quality, legacy systems that do not connect, unclear regulations, and a shortage of internal expertise. Early adopters in mining, agriculture, energy, and medicine are already using autonomous haul trucks, computer vision for targeted spraying, predictive maintenance, and AI documentation tools.
Why This Matters
These trends matter for anyone planning a career or managing a business outside the technology sector. If coding becomes largely automated, the value shifts toward people who can define problems, collaborate with AI agents, and verify results. Industries that have not yet digitized their operations face a steep learning curve, but they also stand to gain the most as tools become cheaper and easier to deploy.
For workers, the implication is clear: pure coding from a specification is a shrinking career path. The professionals who thrive will combine technical fundamentals with domain knowledge in fields such as manufacturing, energy, or medicine. For employers in non-tech industries, the message is urgent. Waiting for AI to mature may mean losing talent to competitors who already understand how to apply models to real-world operations.
Key Takeaways
- Software job postings have fallen sharply, but developer roles are evolving rather than disappearing.
- AI is already writing most code at some leading labs, shifting developers toward review and product direction.
- Robotics and physical AI are attracting record investment as hardware costs decline.
- Non-tech industries lag in AI adoption due to data, regulatory, and expertise gaps.
- The majority of AI-related job postings are expected to sit outside information technology by 2031.
Conclusion
The student who wondered if it is still worth studying received a straightforward answer: yes, but with a different focus. Foundational skills such as Python, mathematics, and classical machine learning still matter. Equally important is the ability to work alongside AI agents rather than against them. The best opportunities may not be at the firms building foundation models. They may be at a mine, hospital, or shipping company that needs someone who grasps both the models and the business context. Time will test these predictions, and by 2031 it should be clearer whether AI's professional future lies mainly outside the technology sector.

Discussion
Join the conversation
Thoughtful reactions, questions, and follow-up ideas help shape the next story.