Andrew Feldman (Cerebras) and Robin Rombach (Black Forest Labs) join the All-In hosts to discuss the rapidly scaling AI infrastructure, including data centers rivaling cities, the shift from training to inference, open source dominance, AI sovereignty, and the road to AGI. Rombach explains the generative video breakthroughs behind Black Forest Labs, and the conversation covers Hollywood's adoption of AI tools from Scorsese to robots and IP concerns.
This summary was generated from show notes and public descriptions, not from a full transcript review. Details may contain inaccuracies.
Highlights
•
The AI Buildout: Datacenters Bigger Than Cities00:00
Andrew Feldman discussed the unprecedented scale of AI data centers, noting that the infrastructure needed for the next wave of AI training and inference rivals entire cities in footprint.
Reasoning, Inference, and Breaking Moore's Law01:50
Feldman explained how the industry is moving from training-centric AI to inference and reasoning workloads, and how Cerebras's architecture is outperforming traditional Moore's Law scaling.
Open Source, AI Sovereignty, and the Road to AGI16:28
The group debated the role of open source models, the increasing importance of national AI sovereignty, and the credibility of the path to artificial general intelligence.
Martin Scorsese, Robots, and the Future of Hollywood IP47:31
The hosts and Rombach discussed how AI tools are being adopted by Hollywood, with Martin Scorsese mentioned as a user, and what it means for intellectual property and creative work.