Step into the world of driving measurable engineering improvements with the GitHub Engineering System Success Playbook (ESSP). This episode explores how focusing on quality, velocity, and developer happiness in unison can unlock better business outcomes, guided by a practical three-step process. We'll dive into identifying barriers, evaluating solutions, and implementing changes to help your teams thrive.

Unlock the secrets to building robust AI agents with the seven node blueprint, a powerful mental model that breaks down complexity. Discover how thinking of agents as graphs with specific components like LLM, Tool, and Guardrail nodes makes designing even advanced agents straightforward. Learn to combine these essential building blocks to create sophisticated and reliable AI workflows.

Join us as we explore how AI is fundamentally changing the landscape of software business models with Bret Taylor, co-founder of Sierra. We discuss the rise of AI agents as the new digital interface, the potential for outcomes-based pricing, and the strategic advantages for startups in this transformative era. Discover how selling value and outcomes is replacing traditional productivity measures.

Join us for insights from the Sequoia AI Ascent 2025 Keynote, where Sequoia partners share their perspectives on the world of AI. They discuss the market opportunity, the rapid pace of adoption, and where value is being created in the stack. The episode delves into current trends like vertical agents and looks ahead to future developments such as agent swarms and the emerging agent economy.

Tired of traditional tools missing complex threats and AI facing privacy hurdles? This episode dives into LSAST, a novel approach integrating traditional SAST scanners with Large Language Models (LLMs) and real-time knowledge retrieval to enhance vulnerability detection. Discover how LSAST offers improved detection accuracy and maintains data privacy, setting a new standard for cybersecurity analysis.

This episode explores how large language models extend their capabilities by interacting with external tools and data via APIs. We introduce the Model Context Protocol (MCP), a new open standard designed to streamline how AI applications access context and use tools. Discover why the relationship isn't "MCP vs API," but how they work together, with MCP often acting as an AI-friendly layer consuming traditional APIs.

The rise of AI agents is changing industries, but they face a critical challenge: the lack of standardized communication protocols. This fragmentation limits agents' ability to collaborate, access tools, and scale effectively to solve complex problems. Discover how unified protocols are essential for building a connected network of intelligence and unlocking the potential of collective AI.

Tune in as we discuss Mem0 and Mem0g, innovative memory architectures designed to tackle the fundamental challenge of Large Language Models forgetting past conversation context. Learn how these systems dynamically extract, consolidate, and retrieve salient information, allowing AI agents to maintain coherent and contextually rich interactions over extended periods. We'll delve into how Mem0 achieves state-of-the-art performance with significantly reduced computational overhead, paving the way for more practical and reliable AI collaborators.

Explore how VibeCoding acknowledges programming as a deeply human activity influenced by psychology, environment, and emotional state. Learn how vibe coders act as conductors, directing AI to transform creative vision into functional code and translating aesthetic feelings into technical specifications. This episode delves into establishing effective collaboration workflow with AI as a creative partner to compose complex systems and curate quality AI-generated code.

Welcome to the '12 Factor Agents' podcast! We're exploring how to build robust, production-ready AI agents by rethinking them as software components rather than simple "loop until you solve it" systems. Discover how factors like owning your control flow, context window, and prompts help overcome limitations like agents getting lost in long contexts. We'll dive into building small, focused micro agents integrated into deterministic workflows for better reliability and performance.