Fastly(US:FSLY) 3 6 Ke·2025-09-02 02:28Core Insights - Fastly conducted a survey revealing that senior developers use AI-generated code 2.5 times more than junior developers, with about one-third of senior developers reporting that over 50% of their code is AI-generated [1][2][9] - The survey highlights the "hidden costs" of AI coding, particularly the time spent identifying and fixing AI-generated errors, which can negate productivity gains [1][11][14] Group 1: Developer Experience and AI Usage - Senior developers (10+ years of experience) are more optimistic about AI's role in coding, with 59% believing AI tools help them deliver code faster, compared to 49% of junior developers [7][8] - Only 13% of junior developers reported using AI for over 50% of their code, while this figure rises to 32% for senior developers, indicating that experience influences AI adoption [9][10] - The survey found that 26% of senior developers felt AI made them "much faster," compared to only 13% of junior developers [7][8] Group 2: Perceived Productivity and Challenges - Nearly 30% of senior developers reported spending more time modifying AI outputs, which offsets time savings, while only 17% of junior developers felt the same [11][12] - Despite the challenges, over half of the developers still believe that tools like Copilot, Gemini, or Claude make them overall faster [11][12] - The phenomenon of "false progress" is noted, where initial speed gains from AI usage are often lost due to frequent modifications and debugging [13][14] Group 3: Emotional and Environmental Considerations - AI tools provide emotional value, with nearly 80% of developers stating that AI makes programming more enjoyable by reducing repetitive tasks and providing instant code generation [14][15] - The survey also assessed awareness of green coding practices, showing that over 56% of junior developers consider energy consumption, while this figure approaches 80% for mid to senior developers [15] - Approximately two-thirds of developers recognize the carbon emissions associated with AI tools, indicating that green coding is becoming part of developer culture [15]