Four CEOs on the Future of AI: CoreWeave, Perplexity, Mistral, and IREN

The podcast features interviews with AI CEOs at Nvidia's GTC conference, focusing on the infrastructure, challenges, and future of AI. Michael Intrator of CoreWeave discusses transitioning from crypto mining to providing GPU compute for AI, emphasizing risk management and long-term contracts. Aravind Srinivas from Perplexity AI explains their approach to AI as an operating system, combining multiple models for accuracy and task automation, and their move towards personal computing with local data processing. Arthur Mensch, CEO of Mistral AI, highlights the importance of specialized, open-source models and data segregation for enterprise AI applications. Daniel Roberts from IREN shares insights on meeting the escalating demand for AI compute with sustainable energy sources and infrastructure development.

Outlines

Part 1: Infrastructure, Hardware, and Financing

Part 2: Market Constraints and AI Orchestration

Part 3: Product Philosophy and Future Impact

Part 4: Open Source, Enterprise, and Energy

Sign in to continue reading, translating and more.

Open full episode in Podwise

All-In with Chamath, Jason, Sacks & Friedberg

Part 1: Infrastructure, Hardware, and Financing

CoreWeave's Early Entry into GPUs and Expansion Beyond Crypto

From Language Models to GPU Lifespan and Depreciation

GPU Demand, Competition, and CoreWeave's Financing Innovation

Part 2: Market Constraints and AI Orchestration

AI Demand, Memory Constraints, and the Boom-Bust Cycle

Perplexity's Focus on Accuracy and the Evolution to Perplexity Computer

Local Models, AI as the Operating System, and Competition in the AI Space

Part 3: Product Philosophy and Future Impact

Model Council, Product Philosophy, and the Power of AI Agents

The Future of Autonomous Businesses and the Impact on Job Displacement

Part 4: Open Source, Enterprise, and Energy

Mistral AI's Open Source Approach and Verticalized Training

IREN's Transition from Bitcoin Mining to AI Compute and Sustainable Energy

Four CEOs on the Future of AI: CoreWeave, Perplexity, Mistral, and IREN

All-In with Chamath, Jason, Sacks & Friedberg

Part 1: Infrastructure, Hardware, and Financing

00:00CoreWeave's Early Entry into GPUs and Expansion Beyond Crypto

CoreWeave's Early Entry into GPUs and Expansion Beyond Crypto

07:03From Language Models to GPU Lifespan and Depreciation

From Language Models to GPU Lifespan and Depreciation

15:25GPU Demand, Competition, and CoreWeave's Financing Innovation

GPU Demand, Competition, and CoreWeave's Financing Innovation

Part 2: Market Constraints and AI Orchestration

23:44AI Demand, Memory Constraints, and the Boom-Bust Cycle

AI Demand, Memory Constraints, and the Boom-Bust Cycle

32:54Perplexity's Focus on Accuracy and the Evolution to Perplexity Computer

Perplexity's Focus on Accuracy and the Evolution to Perplexity Computer

38:51Local Models, AI as the Operating System, and Competition in the AI Space

Local Models, AI as the Operating System, and Competition in the AI Space

Part 3: Product Philosophy and Future Impact

47:55Model Council, Product Philosophy, and the Power of AI Agents

Model Council, Product Philosophy, and the Power of AI Agents

57:36The Future of Autonomous Businesses and the Impact on Job Displacement

The Future of Autonomous Businesses and the Impact on Job Displacement

Part 4: Open Source, Enterprise, and Energy

1:07:15Mistral AI's Open Source Approach and Verticalized Training

Mistral AI's Open Source Approach and Verticalized Training

1:18:55IREN's Transition from Bitcoin Mining to AI Compute and Sustainable Energy

IREN's Transition from Bitcoin Mining to AI Compute and Sustainable Energy