[1/2] Atlassian "State of Developer Experience 2025" (Filed under DevEx)
[1/2] Atlassian "State of Developer Experience 2025" (Rubric #DevEx)
I studied interesting. report From Atlassian on DevEx, which came out last year. This is the second report in a series that they've launched into 2024 In partnership with the then independent company DX (autumn 2025 year bought this company for $1 billion). The study was based on an online survey that 3500 Developers and managers from six countries: USA, UK, Germany, France, Australia and India. The survey was conducted in conjunction with independent firm Wakefield Research. 13–23 March 2025 years. The sample is divided equally by roles: 1 750 Respondents – Leaders (not below the level of development manager)and 1 750 - Engineers. Participants were invited by e-mail.
The main results are presented below.
🤖 Widespread adoption of AI tools Almost all the developers interviewed (99%) Now they use artificial intelligence tools and celebrate time savings. 68Percent save more money 10 hours a week due to their use. For comparison, in the previous report 2024 Most developers have yet to see tangible benefits from AI – the year’s progress has been dramatic.
🚀 Use of released time The hours saved by AI go to work: developers devote them to improving code quality, creating new features and writing documentation. This redistribution of resources suggests that engineers are eager to channel the benefits of AI to add value to the product. (refactoring, development of functionality) team support (documentation, knowledge sharing).
🤡 Despite the benefits, there are still significant losses of time. AI accelerates individual tasks, but organizational inefficiencies negate these gains. 50Percentage of developers lose 10 more than a week due to various friction factors 90% at least 6 weekly. In other words, on average, each developer loses as much time due to interference as they gain from AI.
👾 Main sources of friction in the development process Finding the necessary information (Service data, documentation, API)The need to master new technologies and constantly switch context between disparate tools.3 It takes time for engineers. Engagement with other teams was also among the top five hurdles, while technical debt, which had led the way, fell out of the top five this year.5. Learning new technologies and switching between them are mentioned as notable new types of friction in the workflow.
👀 Coding is not a bottleneck According to the survey, developers spend directly on writing code only ~16% of working time, and the coding process itself is not among the main "pain". This explains why popular AI assistants for programmers can make life easier. (speeding up)But they don't solve the problem of productivity: the main losses occur at other stages. Without the parallel elimination of non-productive tasks, the effect of accelerating coding is limited.
💔 The gap between engineers and management The report revealed a growing disparity of views: 63Percentage of developers think leaders don’t understand their daily challenges, compared to the previous year 44%. The authors attribute this growth of the “empathic gap” to the fact that management hastily reads the hours saved by AI to the teams, without eliminating existing obstacles in the processes. In other words, management may overestimate the effect of AI, while engineers are still drowning in routine and organizational issues.
In the sequel, I will talk about the main conclusions and recommendations of the authors of the report.
#AI #ML #PlatformEngineering #Software #Architecture #Processes #DevEx #Devops