AI is turning Amazon coders into warehouse workers

A decade ago, Amazon introduced knee-high robots to its fulfillment centers. Warehouse workers no longer walked miles along rows of shelving, but instead stood in place as machines brought the goods to them, rapidly picking hundreds of items every single hour.

Now, a very similar assembly-line shift is hitting the company's highest-paid knowledge workers: software engineers.

According to an investigation by The New York Times, Amazon developers report that heavy corporate pressure to adopt generative AI has turned creative engineering into a relentless, high-speed grind that increasingly feels like warehouse labor.

Fewer workers, tighter deadlines

Tech executives have embraced AI coding assistants with open arms. A study by researchers at Microsoft and three universities found that developers using tools like GitHub Copilot boosted output by more than 25%.

Amazon CEO Andy Jassy made the business case clear in his recent shareholder letter, writing that generative AI delivers massive returns through productivity gains and cost avoidance. Jassy noted that AI would fundamentally change software norms, having previously claimed that AI automation saved Amazon roughly 4,500 developer-years during a major internal migration.

On the ground, however, engineers say those efficiency gains have simply become higher quotas. Three Amazon engineers reported that managers have relentlessly pushed AI adoption over the past year. One engineer said their team is roughly half the size it was a year ago, yet leadership expects the same volume of code via AI assistance.

Features that previously took weeks to build and refine are now expected within days. To meet these accelerated timelines, developers are cutting architectural discussions, skipping peer feedback meetings, and leaning on auto-generated code to hit their marks.

The push is spreading across Big Tech. In April, Shopify's CEO told staff that AI usage is now a baseline expectation and tied it directly to performance reviews. Google recently announced an internal hackathon offering a $10,000 top prize for AI tools that boost daily employee productivity.

From writing code to checking it

Beyond sheer speedup, engineers say the nature of their day-to-day work is degrading.

Instead of designing complex systems, coders are increasingly tasked with reading and proofreading machine-generated output. Longtime developer and blogger Simon Willison pointed out the downside: writing code is fun, but doing code reviews is universally considered a chore. With modern generative tools, review has become the primary job.

Amazon engineers noted that while AI can generate large chunks of logic scarily well, reviewing unfamiliar code requires intense focus and double-checking. When you skip that deep thinking, you lose agency over your work.

Harvard labor economist Lawrence Katz noted that knowledge workers are experiencing the modern equivalent of 19th-century factory work intensification, where employers use technological gains to pile on more volume. There is also widespread concern about junior engineers, who may miss out on the trial-and-error fundamentals required to master the craft and earn promotions.

Former Obama campaign CTO Harper Reed offered a counterpoint, arguing that developers should avoid romanticizing manual coding. He compared it to modern auto manufacturing, where workers no longer need to hammer sheet metal by hand.

Why it matters

Generative AI is rapidly turning software engineering from craftsmanship into an industrial throughput contest.

For solo developers and rapid prototypers, these tools are genuinely incredible. As Willison noted, turning a raw concept into a working demo in hours is a massive advantage. But for working engineers inside enterprise hierarchies, the reality is far less liberating.

For the tech industry, the dynamic has fundamentally shifted. Companies are not using AI to give developers breathing room to design better, safer architecture. Instead, management is translating AI gains directly into headcount reductions and faster shipping sprints. When platforms like Shopify and Amazon measure engineers on prompt volume and rapid turnaround, rivals will inevitably follow suit to stay cost-competitive.

We think this aggressive speedup looks smart on quarterly financial sheets, but carries serious structural risk. Forcing engineers to scan and merge thousands of lines of auto-generated code under compressed deadlines is a surefire way to pile up invisible technical debt. More importantly, when you turn creative problem solvers into factory inspectors, you risk burning out senior talent while destroying the training pipeline needed to build the next generation of software architects.