When Jack Dorsey cut Block’s workforce from over 10,000 to just under 6,000, he said the company was “building with intelligence at the core of everything we do.” The layoffs got the attention. The structural insight got lost. What Dorsey was describing — whether intentionally or not — was the first move in a transition to an orchestrator-driven economy. Not the last move, and probably not the cleanest, but the first visible one. And it is going to be a rough few years as everyone adjusts. The visible and the invisible The layoffs are the visible part. The invisible part is the infrastructure debt accumulating underneath every company that has been generating code with AI for the last two years. Google disclosed that 75% of its new code is now AI-generated. Microsoft’s CEO said 20-30% of code at Microsoft is AI-written. That is an enormous amount of code that was produced fast, reviewed lightly, and shipped into production systems that now have to be maintained by someone. The data on what that code looks like is not reassuring. NYU Tandon researchers found that approximately 40% of programs generated by GitHub Copilot contained known security vulnerabilities. Aikido Security surveyed 450 organizations and found that one in five had experienced a serious cybersecurity incident caused by AI-generated code. The Sonar State of Code 2026 survey found that 56% of developers are concerned about over-reliance on AI leading to a decline in understanding of their own codebase, and 49% worry about the erosion of their own coding abilities. As one LeadDev contributor put it: “I don’t think I have ever seen so much technical debt being created in such a short period of time.” The regret data Forrester’s Predictions 2026 report found that 55% of employers who made AI-attributed layoffs already regret it. When you ask CEOs who announced “we’re laying off 20% and replacing them with AI” whether they actually have a mature AI system in place, nine out of ten times the answer is no. Harvard Business Review surveyed 1,006 global executives and found that 60% had reduced headcount in anticipation of AI’s future impact, but only 2% said large layoffs were tied to actual AI implementation. Where we are Companies are generating more code than ever, faster than ever, with less understanding of what is in it than ever. They are laying people off using AI as the excuse, and more than half of them already wish they had not. The code is accumulating, the debt is compounding, and the people who would normally be responsible for maintaining it are being shown the door. There is now such an excessive amount of code being generated that a pendulum is going to swing. The direction is not yet clear. One possibility is that the pendulum swings toward orchestration — companies realize they need experienced people who can direct systems, debug them, and own the outcome, and the orchestrator role emerges as the new center of the technology organization. Another possibility is that the pendulum swings back toward human-written code — companies discover that the infrastructure debt from AI-generated code is so severe that core systems need to be rebuilt by hand. A third possibility is that it swings toward tooling — companies build systems to manage, audit, and maintain AI-generated code at scale, and a new layer of infrastructure emerges to deal with the mess. The honest answer is that it is probably all three at once, and the companies that survive will be the ones that figure out the balance. But the transition is going to be rough. The amount of code in production that nobody fully understands is going to keep growing, the people who could understand it are getting harder to find, and the companies that created the problem are still telling you it is a feature, not a bug. The orchestrator economy is not a future state. It is the transition happening right now, and the infrastructure debt is the bill coming due.