A few years back, if someone told us that we could describe an application in plain English and some AI tool would generate working code for us, most of us would have laughed and called it science fiction. But today, that has become almost normal. We ask AI tools to create functions, explain errors, generate tests, refactor code, build UI components and sometimes even create entire prototypes. What once felt like magic is slowly becoming part of the everyday workflow of a software developer. This naturally brings up one uncomfortable question — if AI can write code, are software engineers still needed?

The 2025 DORA Report delivers a simple but powerful insight: AI isn’t a magic fix—it’s an amplifier. With over 90% of developers now using AI, productivity is rising, but trust issues remain, with 30% doubting AI-generated code . More surprisingly, AI increases throughput while hurting system stability, leading to more failures and rework—a trend clearly shown in DORA’s impact graph