Ollama Co-founder Jeffrey Morgan discusses the enterprise shift to open models, driven by cost, customization, and coding agents, leading to a 150X token growth. He highlights the dominance of Chinese models in cloud consumption, the hardware renaissance enabling local LLMs, and the rise of ultra-efficient "flash" models. Morgan also shares Ollama's journey, from its YC pivot to its monetization strategy, and reflects on lessons for building AI companies.
This episode explores the critical gap between AI investment and return, featuring insights from industry leaders. It covers why most organizations fail to see ROI from AI, the crucial role of mathematical optimization in agentic AI to ensure constraint adherence, practical strategies for structuring AI skills in tools like Claude for quality output, and a 10-20-70 budget framework emphasizing upskilling. Additionally, it delves into the growing importance of the semantic layer for analytics agents to ensure consistent data interpretation and efficiency in large enterprises.
Tanmai Gopal, co-founder of PromptQL and creator of Hasura GraphQL, shares insights on building secure and effective 'company brains' for AI agents. He highlights the surprising metric of a continuously rising edit count as a sign of health, dismisses common pitfalls like siloed knowledge, and proposes a human-in-the-loop, unified wiki approach with granular access control and dynamic credential injection to prevent leaks and foster collaborative knowledge creation.
Arm CEO Rene Haas discusses Arm's evolution from IP licensing to producing physical chips, the transformative role of AI in accelerating chip design and verification, and the critical supply chain bottlenecks impacting data center build-out. He also covers SoftBank's strategic capital deployment, the immense opportunities for Arm in the burgeoning robotics sector, and emphasizes why power-efficient CPUs remain fundamental to all AI workloads, from data centers to edge devices.
Tom McGrath, Chief Scientist at Goodfire, delves into neural network interpretability, intentional design, and the geometry of learned concepts. The discussion covers how models learn modularity and algorithmic structures like arithmetic, the limitations of current interpretability tools like sparse autoencoders, and the implications for alignment, reward hacking, and multi-agent systems.
Cerebras CTO Sean Lie discusses the