To what extent are the current work environments, workflows, and the latest AI tools positively or negatively impacting the overall developer experience? This is the question that GitHub wanted to answer. And for that, he partnered with Wakefield Research, which conducted an online survey for him among 500 developers based in the United States, who work in companies with more than 1,000 employees.
According to Inbal Shani, product manager at GitHub, the report focuses on some of the most important components of the developer experience (DevEx): their productivity, team collaboration, AI and how they think they can have the best impact in enterprise environments.
92% of developers use AI tools to code
LLM code generator like Co-pilot powered by the Open AI Codex, AlphaCode from DeepMind or CodeWhisperer from Amazon, work across a wide range of domains, tasks and programming languages and help increase developer productivity.
GitHub study published last year also showed that the use of its Copilot completion tool was associated with improved productivity among interviewed developers.
A survey by Wakefield Research reveals that AI coding tools are widely used: 92% of US-based developers say they use them both at work and outside of work.
Waiting too long for reviews, upgrades and tests
For developers, writing code and testing is the most time-consuming task, but they also cite waiting too long for code reviews (25%), running upgrades or testing (27%), which negatively impacts their productivity and experience.
Their two main daily tasks are writing code (32%), but also finding and fixing security vulnerabilities (31%), which highlights the importance they attach to security.
To the question: “What positively affects your working day?” developers replied:
- Learning new skills (43%);
- Getting feedback from end users (39%);
- Automated tests (38%);
- Designing solutions for new problems (36%).
More collaboration for better DevEx
Enterprise developers work with an average of 21 other engineers on projects, want more collaboration and, for 92%, make it one of the main performance evaluation measures alongside code quality.
They believe that regular contact with colleagues, documentation and well-managed team meetings are key to successful collaboration. They also need uninterrupted work hours, access to fully configured development environments, and formal mentor-mentee relationships.
What benefits do they expect from using AI coding tools?
Developers generally believe that AI coding tools will improve their performance:
- 70% of respondents believe it will give them an advantage at work, citing better code quality, better execution time and incident resolution as some of the biggest expected benefits. They believe that these metrics should be used to measure their performance beyond the amount of code;
- 57% of developers believe it could help them improve their skills and integrate learning and development into their daily workflow;
- 81% also believe it will make their team more collaborative.
Reference: Inbal Shani blog, GitHub Product Manager