This podcast episode explores the launch of Llama 3, an innovative model developed by META and introduced by Zuckerberg, which has disrupted the world of model building with its smaller size, compatibility, and faster performance. The episode also discusses the introduction of Llama 370B, an open and inexpensive model that poses a challenge to closed competitors. The conversation touches on the importance of differentiation, CapEx, and the impact of open source and closed models in the model marketplace. It explores the adoption of enterprise AI, trust in open source, challenges in data integration, and the significance of agentic thinking in business. The future of GPT-5 and consumer AI, as well as the improvements in token per second achieved by Grok, are also discussed. The section concludes with insights into the VC landscape and the considerations surrounding startups' decision to go public or stay private.
