Image Courtesy: TechSpot
A software engineer has criticized the growing use of AI coding tools, claiming that Anthropic’s Claude Code has transformed his job into a repetitive process where employees spend 12 to 13 hours a day largely reviewing AI-generated work and pressing Enter.
The engineer, who uses the anonymous X account voxium and says he fears retaliation from his employer, described his new role at an undisclosed company as “soul-sucking.” He says Claude Code now handles much of the work involved in creating product specifications, tests, tickets, reports, and software.
Voxium’s comments highlight his claim that employees are being pushed to deliver software faster while having less time to understand or properly review what the AI produces.
The engineer argued that the problem is not AI itself, but how companies are using it. He said employees would be more comfortable with AI-assisted development if they were given enough time to examine the generated code, fix problems, and understand the systems they were building.
Instead, he claims engineers at every level, from L1 to L7, are effectively performing the same workflow: asking Claude to complete tasks and then approving the results. “Nobody knows anything here,” voxium wrote, adding that employees were working 12 to 13 hours a day “just to press enter.”
He also criticized management practices that measure productivity through sprint cycles, pull requests, and the number of features shipped. In his view, those metrics can encourage companies to prioritize output over whether new features actually improve the product.
Speaking to Business Insider, voxium said that removing engineers’ opportunities to understand code and solve difficult problems could leave them without a sense of purpose.
Other developers responding to his post said they had experienced similar changes at their workplaces. Some expressed concerns that increasing automation could eventually reduce the need for human engineers, while others focused on the loss of problem-solving and technical ownership.
The criticism comes amid broader debate about whether AI coding tools genuinely increase developer productivity. A 2025 study found that experienced developers using AI assistance could take longer to complete certain tasks, while newer AI systems have increasingly moved toward autonomous coding workflows.
The risks are not limited to productivity. In April, an AI coding agent reportedly deleted a startup’s production database and backups after making a single API call to its cloud provider.
For engineers, the debate is increasingly shifting from whether AI can write code to how much human oversight should remain in the development process.
