Latent Space: The AI Engineer Podcast · Latent.Space

Bitter Lessons in Venture vs Growth: Anthropic vs OpenAI, Noam Shazeer, World Labs, Thinking Machines, Cursor, ASIC Economics — Martin Casado & Sarah Wang of a16z

·55 min·5 clips

Martin Casado reveals how AI models might raise enough capital to outspend and dominate all companies built on them.

Alessio from Kernel Alliance and Swyx, editor of Latent Space, host a live recording at the Andreessen Horowitz offices with Martin Casado, who has a background in software-defined networking and Netlify, and Sarah Wang, who has been called the best growth investor in the entire AI industry and has backed Noam Shazeer, Mira Murati, Ilya Sutskever, and Fei-Fei Li among others. The conversation opens with the question of what growth investing means in the current AI environment, where some of the most important companies are pre-revenue but raising at growth-scale check sizes. Casado and Wang describe this as a hybrid between venture and growth investing, where the fundamental bet is on the founder as in classic venture but the capital requirements are so large that growth-scale analysis is also required. The episode covers the Anthropic versus OpenAI question at length, with the panelists discussing what distinguishes the two companies' approaches to enterprise market penetration, research philosophy, and commercial strategy. Sarah Wang describes the framework she uses for evaluating large model companies, emphasizing that the primary bottleneck to improve models is how labs can effectively measure what success looks like, a point that connects to the broader evals conversation happening in the industry. Casado discusses his experience in infrastructure investing and how the lessons from networking and cloud apply to AI infrastructure. The panelists address the protein folding breakthrough by AlphaFold as an example of AI solving a problem that the scientific community had expected to require specialized hardware and decades of simulation, running instead on a standard GPU. The conversation covers what kinds of investments a16z is most excited about in the current cycle and what patterns they look for in exceptional founders building in AI. The tone is insider and detailed, with significant technical depth. The audience is AI engineers, founders, and investors following the leading edge of AI development and the capital flows that shape it.

Listen to the show on

Apple Podcasts Spotify

Play clips from this episode More from Latent Space: The AI Engineer Podcast