E-commerce giant Amazon, like many other companies, is actively implementing the use of generative artificial intelligence in the development of its products. However, this is causing dissatisfaction among some programmers who believe that their work is starting to resemble a high-speed assembly line, similar to the warehouse processes that Amazon automated years earlier, The New York Times reports.
Over the past year, developers say, managers have raised productivity targets and tightened deadlines, encouraging widespread use of tools like Microsoft’s Copilot and Amazon’s own AI assistants. Some teams have seen their headcount nearly halved, even as the amount of code they need to write remains the same—employees say they have to rely on AI cues to keep up or their performance will suffer.
"Building a feature for the website used to take a few weeks; now it must frequently be done within a few days," said one Amazon engineer, speaking on condition of anonymity. "We're essentially being forced to accept AI-generated code, reduce the time for discussion, and immediately launch the solution into production."
Historians of technology are already drawing parallels with the 19th century, when factory production divided artisans into small, repetitive operations on assembly lines. "Workers complained about the fast pace and the deterioration of the quality of work," says labor historian Jason Reznikoff. "Today’s implementation of AI in programming feels the same: deep thought is being replaced by quick, modular tasks."
A 2023 study by Microsoft and university partners found that AI assistants could boost programmer productivity by more than 25%. Amazon CEO Andy Jussie praised the findings, saying speed is "key" to staying ahead of the competition and praising generative AI for its "cost savings."
Not all developers are upset about the changes, though. AI advocates say it frees them from tedious tasks (updating legacy code, writing tests), allowing them to focus on high-level development. "AI has saved us thousands of ‘development years,’" Jassy wrote in a letter to shareholders. Citing that in software development, years of experience, or "development years," are often used to gauge a developer’s expertise. A junior developer is typically considered to have 1-3 years of experience, a mid-level developer typically has 3-5 years, and a senior developer is 5+ years.
But Harvard University economist Lawrence Katz warns that AI may prevent beginners from acquiring fundamental skills: "They may never master the basics of programming if AI writes the first drafts for them."
The debate among Amazon engineers resonates with the experiences of their colleagues in the company's warehouses, where robot pickers have increased productivity tenfold, but many employees have complained about the routine and physical strain.
"We see a repetition of their old problems in our new workflows," shared one Amazon developer. "Initially, AI assistance was optional, but as productivity metrics increased, it became mandatory."
Amazon says it regularly reviews its workforce and can add teams as needed. "We will continue to adapt as we integrate generative AI into our processes," said spokesman Brad Glasser. The company also notes positive employee feedback about its AI investments.
Shopify and Google are following a similar course: both have declared artificial intelligence a "core requirement" and will include AI skills in employee evaluations.
While AI can democratize app creation by enabling rapid prototyping and lowering barriers to entry, developers at Amazon and elsewhere are wondering whether speed in exchange for insight could compromise the quality of their profession. As one developer put it, "AI is a blessing for prototyping, but a challenge for gaining experience."