In a recent instance of whether AI is a boon or bane for employees in software engineering firms, Uber CEO Dara Khosrowshahi has revealed that the ride-hailing giant is seeing a massive surge in productivity, thanks to generative AI. The CEO claimed that the technology is effectively turning the company’s engineers into “superhumans.”
In a recent interview ‘On with Kara Swisher’ podcast, Khosrowshahi revealed how Uber has integrated AI into its core development workflow, allowing engineers to focus on complex problem-solving rather than repetitive coding tasks.
AI makes a 40% productivity jump
According to Khosrowshahi, the adoption of AI-powered coding assistants—such as GitHub Copilot and internal custom-built LLMs—has led to a significant “velocity gain” within Uber’s tech teams.
“Our engineers are roughly 40% more productive than they were a year ago,” Khosrowshahi stated. He explained that AI is now handling the “grunt work” of software engineering, including writing boilerplate code, debugging, and generating documentation. This shift allows human developers to act as “editors and architects” rather than just “writers of code.”
AI for everything
While the “superhuman” impact is most visible in engineering, the CEO noted that AI is being woven into every layer of Uber’s operations:
Customer support: AI agents are now resolving complex user queries without human intervention, leading to faster response times for riders and drivers.
Logistics and mapping: Uber uses machine learning to refine “Estimated Time of Arrival” (ETA) accuracy and optimize the matching of riders to drivers in real-time.
Driver experience: The company is testing AI tools to help drivers navigate high-demand areas more efficiently, maximizing their earnings.
The CEO’s comments come at a time when the tech industry is debating whether AI productivity gains will lead to smaller workforces. Khosrowshahi highlighted that for Uber, the goal is “doing more with more,” rather than simply cutting costs. By empowering existing staff with AI, the company aims to accelerate its product roadmap, launching features at a speed that was previously impossible.
AI as an expert helper, not a mandate
While Uber’s Dara Khosrowshahi highlights the “superhuman” velocity of his teams, Zoho co-founder Sridhar Vembu offers a more nuanced, “human-first” philosophy. Vembu recently revealed that AI is enabling Zoho engineers to deliver weeks of work in a single day—citing a veteran developer who completed three weeks of complex UI tasks in just 24 hours. However, unlike many Silicon Valley counterparts, Vembu explicitly refuses to “force-feed” or mandate AI tools. He insists that the decision to use AI must rest with experienced engineers, who act as the essential architects providing the structure while the AI simply “fleshes out the details.”
Vembu remains cautious about the quality of machine-generated code, warning that unchecked AI output can lead to “technical debt” and security vulnerabilities.
