Amazon programmers feel as if they were working in a factory. They blame artificial intelligence
In interviews cited by the New York Times, engineers say that teams can decrease by half, but the production goals remain unchanged, because “AI is to compensate for the lack of hands to work.” The code, which was written for weeks, is now to be delivered within a few days. The one who does not reach for AI tools risk falling out of the competition for promotion or bonus.
Accompancation is accompanying increasing automation of side tasks. Algorithms generate entire programs, write tests and company notes – activities that previously gave younger employees a chance to learn specialization and present talent. Programmers are afraid that after slimming these areas it will be more difficult for them to build competences and justify increases.
Comparisons with Amazon magazines come by themselves. There people stand still in place today, and robots will give them shelves – and here the algorithms will “discover” fragments of the code, and the human task is to quickly verify and pat the finished solution. “I feel like I'm changing from the workshop to the factory,” says one engineer.
Check also: In these works, AI is the greatest threat [LISTA ZAWODÓW]
Intellectual work? Not necessarily
The nature of the intellectual work itself is also changing. “Writing the code is more pleasant than reading it. Meanwhile, AI is absorbed in Code Review most of the time, i.e. assessing what the algorithms wrote,” notes the programmer and blogger Simon Willison, noting that Vending machines generate more code lines than a man is able to scrupulously view.
Amazon defends a new approach. Brad Glasser spokesman emphasizes that the goal is to strengthen the skills of engineers, not replace them, and the career paths remain clear and based on the results. In the annual list to shareholders Andy Jassa, head of the giant, called the generative AI “tool of productivity and avoiding costs”especially in coding. He warned that who would not accelerate, he would be overtaken by the competition.
Scientific data partially confirms the optics of the management board. The experiment with Github Copilot showed that AI developers carry out tasks by an average of 56 percent. faster, and the study at Google estimates acceleration at about 21 percent. with more complex projects. Profits from efficiency, however, go hand in hand with a new risk: concentration of routine activities, lowering the quality of junior training and burning threat.
Economists of the labor market point out that for beginner engineers, automation may resemble a nineteenth-century transition to the production tape. “In mental work, everything accelerates – the employer may impose more tasks, because the algorithm reduces their performance time,” warns prof. Lawrence Katz from Harvard.
Read also: His company wants to automate the entire economy. We are to live in prosperity and give away AI work
What happens in development teams announces a broader transformation. In the same list, Jaszy predicts that AI will “prescribe” almost every customer experience – from Call Center service to medical diagnostics. If the pace remains so high, pressure on the result, instead of creativity, can affect lawyers, financial analysts or content authors.
The history of Amazon shows that Automation does not always eliminate jobs, but changes their character. For programmers, this means shifting the accent from creating to supervision and optimization of tools, which can lead to monotony and weakening of the sense of agency. For enterprises – the potential for productivity increases is tempting, but it is easy to miss the border behind which effectiveness kills innovation.
