The software engineers are not doing well.
AI was meant to simplify programming. Despite being more productive due to AI tools, Siddhant Khare claimed that his work is now more difficult than ever.
Khare told, “It is like a reviewer now, but we used to call it an engineer.” “Every time it feels like you are a judge at an assembly line and that assembly line is never-ending, you just keep stamping those PRs.”
One of Khare’s long essays was headed “AI fatigue is real and nobody talks about it.” He described AI fatigue in it as “the kind of exhaustion that no amount of tooling or workflow optimization could fix.”
He wrote, “Last quarter, I shipped more code than any other quarter in my career.” “I also felt more drained than any quarter in my career.”
Khare, a software developer at ONA, is also an AI tool maker. He stated that while he has nothing against AI, he encourages everyone—including software programmers like himself—to use the technology in a more sustainable manner.
The difficulty, Khare noted, is that AI creates a productivity paradox that falls almost totally on the user to solve by lowering the cost of output while increasing “the cost of coordination, review, and decision-making.”
Khare noted that his usual day now consists of context switching between up to six different design challenges rather than serious concentration on just one.AI “only takes an hour for each one.” However, the human brain suffers greatly when it has to switch between six problems,” he noted. Between issues, the AI is not fatigued. I do.
Khare is not alone, according to early reactions. Other engineers have commented on Lobsters, Hacker News, and X. According to their experiences, vibe coding might result in burnout, as Khare wrote.
“I joke that I’m on the ‘Claude Code workout plan’ now,” one Hacker News user commented. “While my standing desk is working, I do a couple of squats or pushups or simply walk around the house to stretch my legs.” Much more pleasurable than sitting at my desk, hands on the keyboard, all day. Taking my eyes off the screen also allows me to focus on the next task.”
Others describe their day as the “C’mon do something meme” as they wait for a code agent to return results that they must then alter and refine.
“I used to lose myself in focused work for hours,” another user commented. “That has changed. Now I’m continuously lured away, and I’ve seen a pattern: I send a prompt, wait for a response, and then browse.”
On Monday, Harvard Business Review published a report on an eight-week research at a US software company with 200 employees that “discovered that AI tools didn’t reduce work, they consistently intensified it.”When the excitement of trying wears off, employees may see that their workload has quietly increased and that they are feeling stretched from balancing everything that has suddenly appeared on their plate,” the survey stated. “That workload creep can in turn lead to cognitive fatigue, burnout, and weakened decision-making.”
AI tool FOMO and skill atrophy
Khare claimed that as OpenAI, Anthropic, and other AI startups promote their well-liked models and agents, he has also found it difficult to keep up with their frequent upgrades.”Once you open your phone or laptop, you get to Slack, Twitter, LinkedIn, or GitHub, and all of these are like, ‘Look at me,'” Khare declared.
Companies must continue to innovate in order to keep up with the AI race, but for programmers like Khare, the upgrades make them anxious about missing out on what might be the next big thing.
“I was spending weekends evaluating new tools,” he wrote. going over each changelog. watching each demonstration. Fearing that I would slip behind, I tried to stay at the frontier.
Khare concluded by writing that the most frightening aspect was the way AI was altering him. When asked “to reason through a concurrency problem on the whiteboard” without a laptop or artificial intelligence, he claimed to have found it difficult. It is comparable to navigation and GPS. You made mental maps before GPS. Your city was familiar to you. Routes could be reasoned about,” he wrote. “You need GPS to navigate after years of using it. You stopped using the skill, thus it atrophied.”
The engineer who first used the term “vibe coding” last year, Andrej Karpathy, the former head of AI at Tesla, recently noted that although the industry was witnessing a “phase shift” among increasingly powerful AI coding tools, he had also started to notice talent atrophy. I’ve already observed that my ability to create code by hand is gradually deteriorating,” Karpathy remarked.
Identifying appropriate responses to AI usage
Khare claimed that he had developed a set of guidelines to control his usage of AI because he was determined to help himself.
He’s still having trouble keeping his word. He claimed that the first step was to almost ignore the AI discussion throughout his 14-day vacation. Although he stated that a 30-minute timer on AI use is beneficial, Khare clarified that the timer was not meant to be a rigid restriction.
The engineer added that AI firms themselves may also be of use. He likened how addicting games aim to keep players playing to the sensation of being only one suggestion away from the ideal response. “You must maintain some kind of safeguards for the humans to prevent them from destroying themselves,” he stated.
