The best programmers in the world have quietly redefined their job, and it doesn't involve writing code anymore. According to former Google CEO Eric Schmidt, elite engineers now spend their evenings writing specifications and test functions, then let artificial intelligence systems handle the heavy lifting while they sleep. By morning, work that once took six months and ten programmers is complete .

What's Actually Changing in How Programmers Work?

Schmidt described a programmer at one of his startups who exemplifies this shift. Every evening at 7 PM, the engineer defines what he wants the AI to build, writes a test function to evaluate the output, and launches the system. Then he has dinner with his wife and goes to bed. The AI churns through the night, finishing around 4 AM. "This stuff would've taken me six months and 10 programmers at Google," Schmidt said. "This poor guy's sleeping" .

Schmidt

This isn't a one-off example. AI systems at leading research labs are already handling a significant and growing share of programming work, making the engineers who direct them more critical to business operations. The real skill now isn't writing code; it's framing the problem correctly and knowing what "good" looks like.

"If you can define the evaluation function and you can let it run, and if you have enough hardware, you're inventing worlds," said Eric Schmidt.

Eric Schmidt, Former CEO of Google

How to Prepare for AI-Driven Programming Changes

• Master Problem Definition: The ability to articulate exactly what you want an AI system to build is becoming the core skill. Engineers need to think deeply about constraints, edge cases, and success criteria before handing work to AI.
• Develop Evaluation Expertise: Writing effective test functions and judging whether AI output meets quality standards is now the bottleneck. This requires understanding both the technical requirements and business context.
• Learn System Orchestration: Understanding how to parallelize work, manage multiple AI agents, and scale solutions across hardware infrastructure is increasingly valuable. This is the "directing" skill that separates elite engineers from those doing routine coding.

Will AI Replace Programmers?

The natural assumption is that artificial intelligence will make programmers obsolete. Schmidt thinks the opposite is true, at least for the elite ones. Top-tier engineers were always disproportionately valuable, worth roughly 10 times more than those just below them. That gap is about to widen significantly .

The grunt work is disappearing, but the ability to orchestrate AI systems at scale is becoming the entire job. The engineers who can grasp parallelization, define the right constraints, and judge when output is good enough will become even more indispensable. Meanwhile, mid-level programmers doing routine coding tasks face genuine disruption.

What's the Bigger Economic Story Beyond Programming?

Yet the larger impact of AI won't come from automating programmer work at all, Schmidt argued. The biggest economic payoff will come from automating the dull, expensive backbone of corporate operations. Billing systems, accounting processes, inventory management, and delivery logistics silently drain billions from company budgets every year .

"If anything, it's under-hyped because you are fundamentally automating businesses," Schmidt stated at Harvard's John F. Kennedy Jr. Forum. He predicted that artificial general intelligence (AGI), systems capable of matching or exceeding human capabilities across all cognitive domains, could arrive as early as 2029, powered by recursive self-improvement where AI systems learn and plan on their own without waiting for human instruction .

Beyond business operations, Schmidt pointed to medicine, climate solutions, and engineering as fields where AI-driven automation could unlock breakthroughs that remain largely untapped today. The transition from "talkers" (conversational AI systems) to "doers" (autonomous agents capable of executing complex workflows) represents a terrifying leap from theoretical capability to practical, scalable utility .

The real story isn't that AI will replace programmers. It's that the definition of what makes a programmer valuable is fundamentally shifting. The elite will thrive by becoming orchestrators of AI systems. Everyone else will need to adapt or risk becoming obsolete.