🎙️ Podcasts

20VC with Harry Stebbings

The video analyzes Nvidia's record-breaking $81.6

80,000 Hours Podcast (Rob Wiblin)

Contrary to fears of immediate job elimination, AI automation will likely create a temporary '

Interesting Times with Ross Douthat

Chris Brose argues that the US military remains trapped in outdated assumptions of short, decisive wars with technological superiority, leaving it unprepared for the attritional, drone-dominated reality of modern combat against peer adversaries like Russia and Iran.

Huberman Lab

Andrew Huberman explains that grief requires neurologically 'remapping' the three dimensions of attachment—space, time, and closeness—while preserving the emotional bond itself, offering specific tools to navigate loss without becoming trapped in counterfactual guilt loops.

Impact Theory

Simon Dixon argues that three industrial complexes—financial, military, and technological—operate above national governments to extract real assets through a debt-based monetary Ponzi scheme, deliberately cycling through empires as fiat currencies collapse and reset.

Stripe

French tech leaders reflect on the ecosystem's transformation from early 2000s corporate culture to today's AI-driven renaissance, highlighting how reduced capital barriers and improved infrastructure are reshaping entrepreneurship.

Stanford Online

This lecture explains why RLHF hits overoptimization limits with learned reward models, and how RLVR (Reinforcement Learning from Verifiable Rewards) enables unlimited compute scaling on verifiable tasks like math and coding through simpler algorithms like GRPO.

Stanford Online

This lecture explains how post-training transforms raw pre-trained models like GPT-3 into instruction-following systems like ChatGPT through supervised fine-tuning and reinforcement learning, emphasizing that high-quality data curation matters more than algorithmic sophistication.

Stanford Online

This lecture details the pre-training data pipeline, covering the transformation of raw HTML and PDFs into linear text and classifier-based filtering strategies to curate domain-specific datasets, while emphasizing the strategic trade-off between data quality and training duration.

Stanford Online

Sachin Katti, OpenAI's head of industrial compute, details the infrastructure economics driving the AI supercycle, explaining how the company plans to scale to 30 gigawatts by 2030 while navigating the shift from training to inference-heavy agentic workloads and managing massive energy and supply chain constraints.

Stanford Online

Google DeepMind researcher Vivek Natarajan discusses the development of Co-Scientist, an AI system designed to act as a collaborative partner for scientific discovery by moving beyond fast System 1 thinking to rigorous System 2 reasoning, emphasizing that true scientific AI requires the generality of human cognition rather than narrow specialization.

Stanford Online

Google's infrastructure lead Amin Vahdat argues that AI's critical constraint is not raw gigawatts but value delivered per unit of power, requiring a shift from five-nines reliability to throughput-optimized systems where 99.9% uptime is acceptable if it doubles usable capacity.