Training Data

Josh Woodward, VP of Google Labs, leads teams at the forefront of AI innovation. He discusses the rapid development of tools, with AI now writing 25% of Google’s code. Woodward emphasizes the need for taste and design in AI products and the importance of evolving coding practices this year. Key topics include transforming video generation through models like Veo 2, exploring human-computer interaction with Project Mariner, and the optimistic future where costs of AI solutions will dramatically decrease, enhancing creativity and productivity.

Winston Weinberg, Co-founder and CEO of Harvey, brings his expertise in legal AI to the forefront. He discusses the necessity of deep process knowledge for success in legal tech, beyond just powerful models. Weinberg details how Harvey fosters trust with law firms through tailored demos and focuses on complex tasks like international mergers. The conversation dives into the role of AI in reshaping legal practices, enhancing access to services, and the urgent need for innovation in the industry.

Daniel Nadler, CEO and co-founder of OpenEvidence, discusses his revolutionary AI-powered medical knowledge platform designed to aid doctors in real-time. He highlights how smaller, specialized AI models can outperform larger ones in healthcare applications. The conversation covers the app's rapid adoption—growing from 1,000 to over 100,000 users in a year—and its role in democratizing access to medical information. Nadler also emphasizes the significance of accurate, transparent AI in healthcare, and the beneficial partnerships with publishers to improve patient outcomes.

Isa Fulford and Josh Tobin, product leads at OpenAI, dive into the groundbreaking capabilities of the Deep Research agent. They discuss how this technology revolutionizes AI by training models end-to-end without traditional coding. The duo emphasizes the importance of high-quality training data and the o3 model's reasoning skills, enabling it to streamline complex tasks and enhance productivity. They explore how Deep Research can transform knowledge work and highlight the growing role of reinforcement learning in AI's future.

Nikesh Arora, CEO of Palo Alto Networks and former Google and SoftBank executive, dives into the complexities of AI and cybersecurity. He discusses the urgent need for better controls to handle AI threats and emphasizes the shift from prevention to real-time detection. Arora highlights the vulnerabilities of AI models and the importance of balancing innovation with responsibility in security. He also touches on the evolving role of CISOs and the challenges legacy companies face against more agile startups, all while navigating the exciting intersection of sports and business.

Sahir Azam, a product and growth leader at MongoDB, discusses the evolution of vector databases and their critical role in AI applications. He explores how the combination of vectors, graphs, and traditional data structures enhances software development and supports advanced AI capabilities. Azam shares insights from MongoDB’s cloud transformation and advocates for democratizing AI development, making sophisticated tools accessible to mainstream developers. He also highlights innovative applications of AI in robotics and the automotive and pharmaceutical industries.

Stef Corazza, who leads generative AI development at Roblox and has a background in Adobe's 3D and AR platforms, discusses the revolutionary impact of AI on game development. He highlights the creation of the Roblox Assistant, aimed at simplifying programming through natural language. Stef shares insights on maintaining civility within the community while empowering millions of creators. He also explores the potential of AI in enhancing gameplay with user-generated content and emphasizes the unique support Roblox provides to its developers.

Ioannis Antonoglou, a founding engineer at DeepMind and co-founder of ReflectionAI, shares his journey in developing groundbreaking AI technologies like AlphaGo. He discusses pivotal moments in the legendary match against Lee Sedol, the role of self-play in advancing AI, and the engineering challenges faced in creating AlphaGo. The conversation explores the evolution toward artificial general intelligence, including innovations like AlphaZero and MuZero, and highlights the importance of planning and scaling in tackling real-world problems.

Hema Raghavan, co-founder and head of engineering at Kumo AI, previously led AI initiatives at LinkedIn, including the iconic 'People You May Know' feature. In this conversation, she unveils how graph neural networks (GNNs) revolutionize automated machine learning by transitioning from CPU to GPU. Hema explains how GNNs enhance recommendation systems, simplifying feature engineering and adapting to user preferences. She also discusses the transformation of relational data into graph formats and the critical role of explainable AI in advancing technology.

Ion Stoica, co-founder of Databricks and Anyscale, is a renowned professor at UC Berkeley who revolutionized the AI landscape with Spark and Ray. He shares how strategic partnerships, especially with Microsoft, fuelled Databricks' growth. Stoica emphasizes the necessity of identifying future challenges and building solutions that matter tomorrow. He offers insights on cultivating innovation through academia-industry collaboration and the vital role of transparency and control in enterprise data needs.