📣 Send us your press release
Site updates every 15 minutes
Technology

DoorDash CEO: AI programming alone won't significantly boost engineer productivity

DoorDash CEO Tony Xu (徐迅) stated that AI programming represents only a small fraction of an engineer's work and thus doesn't significantly increase overall productivity.

29 July 2026
DoorDash CEO: AI programming alone won't significantly boost engineer productivity
Image is an AI-generated illustration

DoorDash CEO Tony Xu (徐迅) has stated that while AI is valuable, AI-driven programming alone is insufficient to substantially increase software engineer productivity. He explained that coding represents only a portion of an engineer's daily tasks, with current AI models capable of assisting primarily in this specific area.

"Software engineers spend only a part of their day writing code," Xu told Business Insider. "While existing models can assist with programming, which is valuable, they primarily improve efficiency in that specific fraction of work." He estimated that the time dedicated to code generation might range from 25% to 50% of an engineer's total work time.

Xu highlighted that engineers also allocate significant time to product reviews, design meetings, and aligning with various business teams. These crucial aspects of the workflow require AI integration for broader efficiency gains. Without corresponding changes and integration across other parts of the workflow, companies will struggle to achieve expected productivity boosts from AI. The industry as a whole is exploring how to streamline processes so that AI becomes integral to operations, not just code development.

DoorDash itself is already leveraging AI extensively. Xu previously disclosed that "well over half," potentially approaching two-thirds, of DoorDash's code is AI-generated. However, he emphasized that the sheer volume of AI-produced code does not dictate how workflows and team architectures should be adjusted for optimal AI utilization.

Original source: ithome.com