AMD CEO Lisa Su Live on TBPN
This TBPN episode, live from AMD's Advancing AI event, features AMD CEO Lisa Su discussing the company's AI strategy and partnerships. Earlier, guests Oliver Cameron (Odyssey) delves into foundational world models, Muhammad (Ideogram) explores generative AI for enterprise design, and AJ (AMP CEO) champions 'output maxing' for AI infrastructure efficiency, which also leads into a discussion with Seth Cohen about nuclear energy powering future data centers. The hosts also cover recent tech news, including Google's AI monetization and job impact studies.
- Google's recent study, discussed by John Coogan on TBPN, directly challenges widespread fears of mass job displacement by AI, indicating it primarily aids workers. Read →
- Current AI infrastructure is shockingly inefficient: data centers hit less than 20% net flop utilization, and individual chips manage only 15% Model Flop Utilization (MFU). Read →
- Max Anderson, a founder who has spent over $500,000 monthly on Google ads, claims the company's Q1 and Q2 search revenue growth (19% and 17% respectively) is "artificial and extremely unhealthy." He argues this is due to AI cannibalizing legacy search. Read →
- Current AI image generators excel at novel, one-off visuals, but struggle profoundly with the core demands of enterprise marketing: consistent brand application, accurate typography, and deep integration of brand DNA. Muhammad, CEO of Ideogram, calls this the crucial gap in 'design generation.' Read →
- Oliver Cameron, CEO of Odyssey, argues that current AI systems, even in advanced fields like driverless cars, are like the "complex, hand-tuned" natural language processing (NLP) systems of the early 2010s. Read →