Citi Gen AI Summit 2025 Takeaways: The Impact of Gen AI on Software Development

Nowhere has Gen AI had a greater impact than software development. Coding co-pilots provided the first wave of innovation, increasing the speed that developers can write code. The second wave (agents) has opened up entirely new opportunities for experienced and novice developers alike.

A stellar panel gathered at this year’s Citi Gen AI Summit in New York to discuss this and other topics. Panelists included: Jason Warner, CEO & Co-Founder, poolside; Malte Ubl, CTO, Vercel; Russell Kaplan, President, Cognition; and Ioannis Antonoglou, Co-Founder & CTO, Reflection.

Panelists addressed the rise of software written by agentic AI, the evolving role of the software engineer, and the potential demise of the integrated development environment (IDE). Read on for key takeaways from the panel discussion.

Citi Stylus helped summarize the following discussion highlights.

The evolving role of the engineer and RIP IDE?

Gen AI is poised to fundamentally change the nature of software engineering. As AI agents take on more routine tasks, engineers will increasingly focus on managing AI agents, solving complex problems, and designing innovative solutions. The role of the engineer will evolve from a hands-on coder to a strategic thinker, orchestrating AI agents and resources.

The panelists speculated that the traditional integrated development environment (IDE) may become less central as AI agents take on more coding responsibilities. Instead of spending their time writing and debugging code, engineers may transition to managing their AI agents through task-oriented interfaces.

Software development will never be the same

AI agents are increasingly taking on tasks traditionally handled by human developers, such as bug fixing and code review. This automation frees developers to focus on more complex and creative projects, ultimately enhancing efficiency and fostering innovation within software teams. The panel agreed that Gen AI's role in software development will continue to expand, leading to significant changes in how humans (and AI) build applications.

The rise of agentic software development

Gen AI will be able to optimize more than just applications. “Agentic software development” has the potential to unlock previously unattainable levels of software functionality and customization. Examples include Gen AI systems that can automatically generate code, design user interfaces, and even optimize software performance. This approach could lead to a new era of highly adaptable and intelligent software applications.

AI agents will continue to get smarter

AI agents are constantly learning and improving through feedback loops, allowing them to adapt to new challenges and optimize their performance over time. This iterative process will drive rapid advancements in Gen AI capabilities within software development, leading to more sophisticated and efficient systems.