The New Stack Podcast

Jyoti Bansal, co-founder and CEO of Harness and a seasoned entrepreneur, dives into the complexities of AI in software development. He reveals how AI tools often create bottlenecks in testing and deployment rather than speeding up the process. Bansal discusses the importance of human oversight and the role of specialized agents within Harness AI that enhance pipeline efficiency. He also shares insights on the evolving landscape of AI, likening its current hype to the dot-com bubble, which will ultimately lead to meaningful innovations.

Lin Sun, Head of Open Source at Solo.io, dives into the challenges of secure connectivity in the agentic AI landscape. She explains how Agentgateway enhances communication governance between AI agents and tools, contrasting it with traditional API gateways. Lin highlights the integration of protocols like MCP and A2A, and how it provides fine-grained access control for tool exposure. Major tech players like AWS and Microsoft are already involved, and she discusses the framework-agnostic benefits of combining Agentgateway with Kagent for a resilient AI ecosystem.

David Cramer, Founder and Chief Product Officer of Sentry, shares insights on the limitations of current generative AI in replacing human engineers. He emphasizes the importance of human oversight in AI debugging tools, specifically discussing Sentry's AI-driven tool, Seer, which aids in identifying root causes of errors. Cramer also details Sentry's shift from open source to a fair-source licensing model to prevent exploitation while maintaining accessibility. His reflections on LLMs highlight their potential in coding, yet underscore that they are not ready for production use due to reliability issues.

In this engaging discussion, Andrew Milich from Cursor and Tom Moor from Linear dive into the synergy between coding agents and project management. They explore how Cursor's AI coding agent can streamline workflows by integrating directly with Linear’s new agent-specific API. Key insights include the importance of providing agents with clear, contextual tasks rather than vague commands. They also touch on the future of autonomous agents and the potential for more companies to adopt agent-specific APIs to enhance productivity.

Pat Casey, CTO and co-founder of ServiceNow, shares insights on running 90% of their workloads, including AI, on physical servers. He discusses the benefits of their in-house GPU hubs for low-latency operations and how their new AI coding assistant, Windsurf, has boosted productivity by 10%. Casey raises concerns about the future of junior developers as AI tools take on traditional tasks, though he remains optimistic about AI's transformative potential despite challenges ahead.

Christopher "Crob" Robinson, Head of the Open Source Security Foundation, brings critical insights on the EU's Cyber Resilience Act. He underscores how compliance burdens threaten lone developers who often manage popular open source projects with limited resources. The discussion reveals alarming statistics, such as 97% of commercial software utilizing open source, while compliance could impose penalties in the billions. Robinson emphasizes the urgent need for funding models and collaboration to support these developers in navigating new regulatory landscapes.

In this discussion, Zach Lloyd, Founder and CEO of Warp, reveals how the Warp terminal has transformed into an agentic development environment. He introduces Warp Code, showcasing AI agents that assist in coding, debugging, and project management. Transparency in AI-generated code is emphasized, allowing developers to maintain control. Lloyd also shares the story behind their unique Western-themed product launch, featuring the quirky "Code on Warp" branding that set them apart. This evolution paves the way for a new era in coding productivity.

Jim Zemlin, Executive Director of the Linux Foundation, delves into the fragmented world of open source AI. He discusses the shift from foundational to open-weight models and emphasizes the need for a unified structure while valuing innovation. Zemlin views the industry's challenges regarding model transparency and the costs of developing frontier AI models with a pragmatic lens. He also highlights the growing open-source culture in China and the impacts of global regulations on the open source community, underscoring the fine balance between scale and agility.

Raj Verma, CEO of SingleStore, dives into the evolving landscape of enterprise AI. He reveals that less than 1% of enterprise data is currently utilized for AI, stressing the importance of aligning data strategies with AI aspirations. Raj outlines a three-phase evolution for AI, from automating simple tasks to achieving the seemingly impossible. He critiques legacy systems and advocates for modern, unified data platforms to meet the high demands of agentic AI, emphasizing the urgent need for enterprises to transform their data infrastructure.

Tzvika Shneider, CEO and co-founder of Pynt, an innovative API security startup, dives into the pitfalls of the Model Context Protocol (MCP). He reveals how MCP, evolving from traditional APIs, is fraught with vulnerabilities as AI agents begin to connect. Shneider emphasizes that 72% of MCP plugins expose high-risk operations without proper validation. He points out that while MCP exploits are rare now, most organizations are ill-equipped to manage these security challenges. His insights underscore the urgent need to enhance security measures in this rapidly changing tech landscape.