At Amazon, some coders say their jobs have begun to resemble warehouse work

Since at least the industrial revolution, workers have worried that machines would replace them.

But when technology transformed automaking, meatpacking and even secretarial work, the response typically wasn’t to slash jobs and reduce the number of workers. It was to “degrade” the jobs, breaking them into simpler tasks to be performed over and over at a rapid clip. Small shops of skilled mechanics gave way to hundreds of workers spread across an assembly line. The personal secretary gave way to pools of typists and data-entry clerks.

The workers “complained of speedup, work intensification and work degradation,” as labor historian Jason Resnikoff described it.

Something similar appears to be happening with artificial intelligence in one of the fields where it has been most widely adopted: coding.

As AI spreads through the labor force, many white-collar workers have expressed concern that it will lead to mass unemployment. Joblessness has ticked up and widespread layoffs might eventually come, but the more immediate downside for software engineers appears to be a change in the quality of their work. Some say it is becoming more routine, less thoughtful and, crucially, much faster paced.

Companies seem to be convinced that, like assembly lines of old, AI can increase productivity. A recent paper by researchers at Microsoft and three universities found that programmers’ use of an AI coding assistant called Copilot, which proposes snippets of code that they can accept or reject, increased a key measure of output more than 25 per cent.

At Amazon, which is making big investments in generative AI, the culture of coding is changing rapidly. In his recent letter to shareholders, CEO Andy Jassy wrote that generative AI was yielding big returns for companies that use it for “productivity and cost avoidance.” He said working faster was essential because competitors would gain ground if Amazon doesn’t give customers what they want “as quickly as possible” and cited coding as an activity where AI would “change the norms.”

Those changing norms have not always been eagerly embraced. Three Amazon engineers said managers had increasingly pushed them to use AI in their work over the past year. The engineers said the company had raised output goals and had become less forgiving about deadlines. It has even encouraged coders to gin up new AI productivity tools at an upcoming hackathon, an internal coding competition. One Amazon engineer said his team was roughly half the size it was last year, but it was expected to produce roughly the same amount of code by using AI.

Amazon said it conducts regular reviews to make sure teams are adequately staffed and may increase their size if necessary. “We’ll continue to adapt how we incorporate Gen AI into our processes,” Brad Glasser, an Amazon spokesperson, said.

Other tech companies are moving in the same direction. In a memo to employees in April, the CEO of Shopify, a company that helps entrepreneurs build and manage e-commerce websites, announced that “AI usage is now a baseline expectation” and that the company would “add AI usage questions” to performance reviews. Google recently told employees that it would soon hold a companywide hackathon in which one category would be creating AI tools that could “enhance their overall daily productivity,” according to an internal announcement. Winning teams will receive $10,000.

The shift has not been all negative for workers. At Amazon and other companies, managers argue that AI can relieve employees of tedious tasks and enable them to perform more interesting work. Jassy wrote last year that the company had saved “the equivalent of 4,500 developer-years” by using AI to do the thankless work of upgrading old software.

Eliminating such tedious work may benefit a subset of accomplished programmers, said Lawrence Katz, a labor economist at Harvard University who has tracked research on the subject closely.

But for inexperienced programmers, the result of introducing AI can resemble the shift from artisanal work to factory work in the 19th and 20th centuries.

“Things look like a speedup for knowledge workers,” Katz said, describing preliminary evidence from ongoing research. “There is a sense that the employer can pile on more stuff.”

Bystanders in Their Own Jobs

The automation of coding has special resonance for Amazon engineers, who have watched their blue-collar counterparts undergo a similar transition.

For years, many workers at Amazon warehouses walked miles each day to track down inventory. But over the past decade, Amazon has increasingly relied on so-called robotics warehouses, where pickers stand in one spot and pull inventory off shelves delivered to them by lawn-mower-like robots, no walking necessary.

The robots generally haven’t displaced humans; Amazon said it has hired hundreds of thousands of warehouse workers since their introduction, while creating many new skilled roles. But the robots have increased the number of items each worker can pick from dozens an hour to hundreds. Some workers complain that the robots have also made the job hyper-repetitive and physically taxing. Amazon says it provides regular breaks and cites positive feedback from workers about its cutting-edge robots.

The Amazon engineers said this transition was on their minds as the company urged them to rely more on AI. They said that while doing so was technically optional, they had little choice if they wanted to keep up with their output goals, which affect their performance reviews.

The expectations have sped up rapidly. One engineer said that building a feature for the website used to take a few weeks; now it must frequently be done within a few days. He said this is possible only by using AI to help automate the coding and by cutting down on meetings to solicit feedback and explore alternative ideas. (A second engineer said her efficiency gains from using AI were more modest; different teams use the tools more or less intensively.)

The new approach to coding at many companies has, in effect, eliminated much of the time the developer spends reflecting on his or her work.

“It used to be that you had a lot of slack because you were doing a complicated project — it would maybe take a month, maybe take two months, and no one could monitor it,” Katz said. “Now, you have the whole thing monitored, and it can be done quickly.”

As at Microsoft, many Amazon engineers use an AI assistant that suggests lines of code. But the company has more recently rolled out AI tools that can generate large portions of a program on its own. One engineer called the tools “scarily good.” The engineers said that many colleagues have been reluctant to use these new tools because they require a lot of double-checking and because the engineers want more control.

“It’s more fun to write code than to read code,” said Simon Willison, an AI fan who is a longtime programmer and blogger, channeling the objections of other programmers. “If you’re told you have to do a code review, it’s never a fun part of the job. When you’re working with these tools, it’s most of the job.”

This shift from writing to reading code can make engineers feel like bystanders in their own jobs. The Amazon engineers said that managers have encouraged them to use AI to help write one-page memos proposing a solution to a software problem and that the artificial intelligence can now generate a rough draft from scattered thoughts.

They also use AI to test the software features they build, a tedious job that nonetheless has forced them to think deeply about their coding. One said that automating these functions could deprive junior engineers of the know-how they need to get promoted.

Amazon said that collaboration and experimentation remain critical and that it considers AI a tool for augmenting rather than replacing engineers’ expertise.

It said it makes the requirements for promotions clear to employees.

Harper Reed, another longtime programmer and blogger who was the chief technology officer of former President Barack Obama’s reelection campaign, agreed that career advancement for engineers could be an issue in an AI world. But he cautioned against being overly precious about the value of deeply understanding one’s code, which is no longer necessary to ensure that it works.

“It would be crazy if in an auto factory people were measuring to make sure every angle is correct,” he said, since machines now do the work. “It’s not as important as when it was group of 10 people pounding out the metal.”

And just as factories abroad have made it cheap and easy for entrepreneurs to manufacture physical products, the rise of AI is likely to democratize software-making, lowering the cost of building new apps.

“If you’re a prototyper, this is a gift from heaven,” Willison said. “You can knock something out that illustrates the idea.”

The Dreaded Speedup

Amid their frustration, many Amazon engineers have joined a group called Amazon Employees for Climate Justice, which is pressuring the company to reduce its carbon footprint and has become a clearinghouse for workers’ anxieties about other issues, like return-to-office mandates. (Amazon said it is working to reduce carbon emissions from its data centers; the climate justice group is pushing it to provide more information on how.)

The group’s organizers say they are in touch with several hundred Amazon employees on a regular basis and that the workers increasingly discuss the stress of using AI on the job, in addition to the effect that the technology has on the climate.

The complaints have centered around “what their careers are going to look like,” said Eliza Pan, a former Amazon employee who is a representative for the group. “And not just their careers but the quality of the work.”

While there is no rush to form a union for coders at Amazon, such a move would not be unheard-of. When General Motors workers went on strike in 1936 to demand recognition of their union, the United Auto Workers, it was the dreaded speedup that spurred them on.

The typical worker felt “that he was not free, as perhaps he had been on some previous job, to set the pace of his work,” historian Sidney Fine wrote, “and to determine the manner in which it was to be performed.”