In a recent LinkedIn post, 🌀 Luca Rossi discusses a concerning paradox emerging with the integration of Artificial Intelligence in engineering workflows. Rossi highlights research suggesting that while AI tools can increase perceived productivity, they may simultaneously degrade the developer experience, leading to decreased job satisfaction and a hit to the crucial ‘flow state’.
The Disconnect Between Output and Experience
🌀 Luca Rossi points to research by Annie Vella, which tracked engineers over a six-month period. The findings presented by Rossi are striking: the percentage of engineers reporting a negative developer experience nearly doubled, increasing from 14% to 27%. Perhaps more alarmingly, the share of engineers experiencing a decline in their ability to achieve a ‘flow state’ rose from 7% to 20%.
“The group reporting a negative developer experience nearly doubled, from 14% to 27%. Flow state took the largest hit, with the share saying it got worse rising from 7% to 20%.”
This trend challenges a long-held assumption in the industry, as 🌀 Luca Rossi explains. Historically, developer experience and productivity have been closely linked. Improvements in factors like focus, efficient feedback loops, and the ease of shipping code typically led to both better experiences and higher output, and vice versa.
AI’s Challenge to the Status Quo
The core of 🌀 Luca Rossi’s analysis is that AI appears to be disrupting this established relationship. While AI-powered tools can accelerate code generation, Rossi raises a critical question about the nature of the work being produced and its impact on the individuals performing it.
“AI seems to be breaking this relationship. Code gets produced faster, but the work becomes more fragmented and less satisfying.”
According to 🌀 Luca Rossi, the danger lies in an over-reliance on quantifiable output metrics. If development teams or management solely focus on the speed at which code is produced, they might overlook significant declines in employee well-being and engagement.
The Question of ‘Success’
🌀 Luca Rossi concludes by posing a vital question that cuts to the heart of performance evaluation in the age of AI:
“The problem is that, if we only measure output, it can still look like success. But is it a win to go faster by making people more miserable? 🤷♂️”
This perspective underscores the need for a more holistic approach to measuring success in software development. It suggests that organizations must consider not only how much code is being written but also how the process impacts the people doing the work. Ignoring the qualitative aspects of the developer experience, Rossi implies, could lead to long-term issues with retention, innovation, and overall team health, even if short-term output figures appear to improve.
📝 About This Content
This article is based on insights shared by 🌀 Luca Rossi on LinkedIn.
📅 Originally posted on August 3, 2026 | View original post on LinkedIn →