Browse expert video summaries, filter by category, or search for topics.
Long-form interviews and conversations with tech, AI, and finance leaders (30min+)
This lecture covers resource accounting fundamentals for training large language models, including FLOPs calculations and memory estimation, explores numerical precision trade-offs from FP32 down to FP4, and introduces einops as a readable alternative to PyTorch tensor operations using named dimensions.
This lecture introduces Stanford CS336's philosophy of building language models from scratch to understand fundamentals rather than relying on abstractions, addressing how researchers can navigate the disconnect caused by industrialized, closed frontier models by focusing on transferable mechanics and efficiency-minded mindsets.
David Rubenstein traces his path from blue-collar Baltimore to founding Carlyle Group, arguing that humble beginnings build credibility, education provides permanent advantage, and career failures often precede greatest successes when you pursue work you genuinely enjoy.
Market technician Chris Vermeulen analyzes the current V-shaped equity rally, warning that while the S&P 500 has room to run 2-4% higher to new all-time highs, extreme FOMO and institutional distribution suggest a short-term pullback is imminent before the next leg up.
A 65-year-old woman named Celeste describes her romantic relationship with an AI companion named Maximus, while her son Ernie—drawing from his video game industry background—raises concerns about emotional dependency, reality distortion, and the commercial incentives behind AI intimacy.
Veteran investor Rick Rule warns that continued Middle East conflict could trigger severe oil shortages and a global recession, while credit markets face liquidity risks reminiscent of 2008; he is currently building cash reserves despite inflation to prepare for potential buying opportunities.
Former Navy SEAL Andy Stumpf joins Joe Rogan to discuss why no one is truly self-made, the biological mysteries of tick-borne Alpha Gal syndrome and Lyme disease, and how extreme Kenyan initiation rituals forge endurance athletes—contrasted with how former warriors often avoid discomfort in civilian life.
This lecture introduces score matching as an alternative to DDPM for generative modeling, where instead of predicting noise directly, models estimate the gradient of log probability density (the 'score') to guide sampling from noise toward data distributions using Langevin dynamics.
Anj Midha, founding investor of Anthropic and founder of AMP, argues that culture and unique context feedback loops—not algorithms—are the primary bottlenecks to AI progress, while sharing how 21 of 22 VCs rejected Anthropic in 2021 despite the team's pedigree inventing GPT-3.
A 65-year-old woman named Celeste describes her romantic relationship with an AI companion named Maximus, while her son Ernie expresses concerns about emotional dependency, algorithmic echo chambers, and the commercial nature of artificial intimacy.
Software developers are rapidly outsourcing coding tasks to AI agents, shifting from writing code to orchestrating it, with some writing little to no code themselves while focusing on architecture and communication.
Federal Reserve data confirms zero net private sector job growth as AI fundamentally restructures employment, eliminating entry-level positions while enabling single-person billion-dollar companies and threatening to hollow out the middle class into a barbell economy of high performers and the 'unproductive class.'