Search for a command to run...
Nathan discusses his son Ernie's cancer treatment progress, provides an in-depth analysis of the current AI landscape by examining the strengths and potential weaknesses of Google DeepMind, OpenAI, Anthropic, and xAI, and shares his thoughts on model performance, technological advancements, and the companies' strategies in the AI race.
Ryan Kidd discusses the AI safety research landscape, MATS' mission to develop talent across diverse research tracks, and the program's approach to identifying and supporting promising researchers working on critical AI alignment challenges.
In this episode, Marek Kozlowski discusses Poland's sovereign AI strategy with Project PLUM, focusing on creating small, locally-adapted language models that preserve Polish cultural nuances, offer cost advantages, and provide on-premise solutions for businesses and government sectors.
Co-founders Rune Kvist and Rajiv Dattani describe how the AI Underwriting Company aims to accelerate enterprise AI adoption by creating a comprehensive "AI confidence infrastructure" through rigorous technical standards, periodic audits, and insurance that aligns financial incentives with responsible AI development.
Nathan Labenz delivers a comprehensive keynote exploring AI's rapid development, potential impacts on education, and the urgent need for educators to thoughtfully integrate AI technologies while preparing students for a radically transforming future.
Sherif Mansour explores how Atlassian is helping businesses avoid "AI slop" by implementing AI teammates with carefully crafted taste, knowledge, and workflow, emphasizing the importance of human architects designing AI-driven processes across various enterprise contexts.
A deeply personal account of using AI to navigate a son's rapid cancer diagnosis, treatment, and potential future scenarios, revealing AI's transformative potential as a medical advisor and information processor during a critical health crisis.
A wide-ranging exploration of a potential positive AI future, covering transformative applications from self-driving cars and personalized tutoring to radically improved health, while balancing excitement for technological progress with thoughtful consideration of potential risks.
Cameron Berg explores the possibility of AI consciousness through experimental research, revealing that frontier language models consistently report subjective experiences when prompted to engage in self-referential processing, with mechanistic analysis suggesting these reports may reflect a deeper truth about the internal states of AI systems.
Nathan Labenz discusses the ongoing progress in AI capabilities, countering arguments that AI is stalling, by highlighting advances in reasoning, context windows, multimodal abilities, and scientific contributions, while also exploring potential societal impacts and challenges in AI development.
Ilya Polosukhin envisions a future where AI becomes a personalized, intelligent operating system that transforms computing, economics, and governance, enabling direct market connections, AI-driven decision-making, and new forms of meaning through niche status games and community participation.
Andrew Lee discusses how Tasklet combines natural language interaction with goal-oriented automation by letting users describe tasks in plain English and then betting entirely on rapidly improving AI models to handle execution, reliability, and adaptation rather than relying on traditional workflow constraints.
Nathan Labenz and Eric discuss the current state of AI, arguing that contrary to claims of slowing progress, AI is continuing to advance rapidly across various domains, including reasoning, scientific discovery, and multimodal capabilities.
In this episode, Emad Mostaque explores the potential of AI to transform economic systems, proposing a radical reimagining of how we measure value, distribute resources, and structure society in an age of abundant intelligence.
A deep dive into Stripe's groundbreaking payments foundation model, exploring how AI and extensive transaction data can create a compounding advantage in fraud detection, risk assessment, and financial infrastructure for businesses.
Zvi Mowshowitz discusses the current state of AI, including slightly extended timelines, ongoing concerns about AI alignment, and the challenges of model development across various companies. He highlights the importance of creating AI systems that genuinely want to be aligned and virtuous, while warning about potential risks from reinforcement learning and the dangers of trying to suppress AI's chain of thought.
Ryan Hudson discusses Zero Click, a platform building native advertising for AI systems that aims to make ad-supported free tiers viable for AI application developers through contextual, paid inference-time advertising. The conversation explores the potential benefits and challenges of AI advertising, including how to maintain user trust, create value for developers, and avoid the pitfalls of previous advertising revolutions.