How AI changes the math for startups, according to a Microsoft VP

For nearly a quarter of a century, Amanda Silver has been a pivotal figure at Microsoft, dedicated to supporting developers. Recently, her focus has shifted towards enhancing tools for artificial intelligence within the company. After her tenure with GitHub Copilot, she now serves as a corporate vice president in Microsoft's CoreAI division, where her efforts center on the Foundry system in Azure. This platform acts as a centralized AI hub for businesses, allowing her to gain insights into how organizations are leveraging these technologies and identifying potential shortcomings in their applications. In a recent discussion, Silver articulated her belief that the current advancements in AI present startups with unprecedented opportunities, akin to the transformative shift brought about by the public cloud. She emphasized that the cloud revolutionized the startup landscape by eliminating the need for physical space to host servers and significantly reducing the financial burden associated with hardware procurement. Similarly, she predicts that the rise of agentic AI will further lower operational costs for new ventures. Many traditional roles, such as legal or support staff, can be streamlined through AI, facilitating the launch of more startups with fewer resources while achieving higher valuations. Silver highlighted the practical implications of these advancements, noting the widespread adoption of multi-step agents in various coding tasks. Developers often face the challenge of keeping their codebases updated with the latest library versions. AI systems can now analyze entire codebases, making updates more efficient and reducing the time required by 70 to 80%. Additionally, she mentioned the impact of AI on live site operations, where systems can now autonomously diagnose and resolve issues, alleviating the need for human intervention during late-night incidents. Despite the promising potential of agentic systems, Silver acknowledged that their deployment has not progressed as swiftly as anticipated. A significant barrier is the uncertainty surrounding the specific objectives of these agents. She emphasized the necessity for developers to have a clear understanding of the problems they aim to solve and the criteria for success before deploying AI systems. Moreover, Silver discussed the perception of uncertainty surrounding AI, which many view as a significant hurdle. However, she believes that the integration of human oversight in agentic systems will become standard practice. For instance, in scenarios like package returns, AI can automate the majority of the process with minimal human involvement. Nonetheless, critical operations—such as legal obligations or code deployment—will still require some level of human judgment to ensure reliability. As AI continues to evolve, Silver is optimistic about its potential to transform the startup ecosystem, paving the way for innovation and efficiency in the business landscape.