Open Source Wins, AGI Is Here, and Scorsese's AI Toolkit with CEOs of Cerebras & Black Forest Labs

Overview

This episode dives into the transformative impact of AI on infrastructure, reasoning models, and creative industries. Featuring Andrew Feldman (CEO of Cerebras) and Robin Rombach (CEO of Black Forest Labs), the discussion spans the unprecedented scale of AI data centers, breakthroughs in inference and reasoning, the rise of open-source AI, and the future of generative video in Hollywood and beyond.

Notable Quotes

- We're talking about data centers the size of football fields, consuming more power than mid-size cities. – Andrew Feldman, on the scale of AI infrastructure.

- The AI starts telling people, 'You're token maxing and need to get more focused here.' – Jason Calacanis, on how AI is reshaping human workflows.

- Getting the mental picture out of your head and into a visual form is one of the most powerful uses of this technology. – Robin Rombach, on AI's role in creative ideation.

🖥️ The Unprecedented Scale of AI Infrastructure

- Andrew Feldman describes the massive buildout of AI data centers globally, likening it to historical feats like the Great Wall of China.

- These centers are being constructed in unexpected regions, including Kazakhstan and the Nordics, and are projected to consume more power in the next few years than the last 50 years combined.

- Demand for AI compute is insatiable, with companies like OpenAI and Google pre-ordering chips years in advance, leading to a $25 billion backlog for Cerebras.

🧠 Reasoning Models and Breaking Moore’s Law

- AI has evolved from simple summarization to reasoning and intent understanding, with tools like OpenAI’s Fable and Hermes agents leading the charge.

- Andrew Feldman highlights how Cerebras’ chips have surpassed Moore’s Law, enabling faster and more efficient inference for reasoning models.

- Recursive learning, or loop maxing, is driving exponential improvements in AI outputs, raising questions about the limits of computational gains.

🌍 Open Source, AI Sovereignty, and Regulation

- Open-source AI is gaining traction as companies seek more control over their models, especially in regulated industries like healthcare and finance.

- Andrew Feldman and Jason Calacanis discuss the importance of domestic open-source models to counter reliance on foreign AI systems.

- The episode touches on the need for thoughtful regulation, with Feldman comparing AI oversight to pharmaceutical trials—ensuring safety without stifling innovation.

🎥 Generative Video and Creative Collaboration

- Robin Rombach explains how Black Forest Labs’ generative models are revolutionizing video production, enabling directors like Martin Scorsese to visualize scenes before filming.

- AI is being used in production workflows, such as replacing green screens with generative backdrops, significantly reducing costs for high-budget films.

- The technology is also empowering startups and independent creators to produce high-quality launch videos and content at a fraction of traditional costs.

🤖 The Future of AI in Robotics and Real-World Applications

- Black Forest Labs is developing multimodal models that combine image, video, and action prediction, paving the way for robotics applications.

- These models could enable robots to learn tasks from both real-world and synthetic data, with minimal fine-tuning required.

- Rombach envisions a future where robots can be prompted in natural language to perform complex tasks, bridging the gap between perception and action.