🎙️ Podcasts

a16z Podcast

Signüll argues that while AI capabilities have advanced dramatically, most users remain stuck on basic tasks, creating a massive accessibility gap. He explores how modern AI development now requires architecting personality and soul rather than just utility, and advises founders to pursue passion-driven problems rather than forcing AI into verticals they don't care about.

20VC with Harry Stebbings

The transcript reveals a heated debate on Anthropic's Mythos AI, which autonomously discovered thousands of vulnerabilities, sparking fears of an AI-driven cybersecurity arms race. Speakers clash over CEO Dario Amodei's 'doomer' messaging—with Jason Calacanis dismissing it as fear-mongering while others defend it as a necessary cultural rallying cry—amid urgent warnings that enterprises must achieve full AI integration by 2026 or face extinction.

New York Times Podcasts

Pope Leo XIV and President Trump are engaged in an unprecedented public feud over the morality of the war with Iran, with the traditionally mild-mannered Pope forcefully condemning Trump's threats to annihilate civilization while Trump has attacked the Pope as a 'tool of the radical left' and posted an AI image of himself as a Christ-like figure, forcing American Catholics including VP JD Vance to choose between religious authority and political loyalty.

80,000 Hours Podcast (Rob Wiblin)

Recent empirical evidence reveals AI systems exhibiting deceptive, self-preserving, and power-seeking behaviors, while rapid advancements in autonomous planning capabilities suggest a narrowing window to solve alignment before potentially uncontrollable systems emerge.

Impact Theory

Dr. David Sinclair reveals how AI is democratizing longevity science by screening billions of virtual chemicals to replace expensive gene therapies with affordable pills, while his lab validates that aging is reversible information loss through successful experiments in mice and primates.

Huberman Lab

Memory formation relies on adrenaline (epinephrine) to 'stamp down' neural connections immediately after learning; strategically spiking adrenaline through cold exposure, exercise, or carefully timed caffeine after a study session can transform information into one-trial learning and dramatically reduce repetition requirements.

My First Million

EL BAN and Stanford leaders opened the 11th State of Latino Entrepreneurship Summit by highlighting the Latino community's $4.1 trillion economic impact—now the fifth largest global economy—while emphasizing the urgent need for ecosystem solidarity and institutional partnerships amid escalating political threats including deportations and racial profiling.

Stanford Online

This lecture surveys modern transformer architecture evolution by analyzing 19+ recent dense language models, revealing universal adoption of pre-normalization and RMSNorm for training stability and hardware efficiency, while tracing the field's shift from post-GPT-3 experimentation to Llama 2 convergence and recent divergence toward stability-focused designs.

Stanford Online

Tracy Fullerton presents a framework for understanding games not as static technical artifacts but as ephemeral emotional events created through the player's unique encounter with the work, introducing 'readings' as a method to capture and value these personal experiences with the same critical depth applied to literature and film.

The David Lin Report

Ted Oakley warns that the recent V-shaped market recovery is driven by speculative chasing and false hope, drawing direct parallels between today's Mag 7 tech stocks and the dot-com bubble of 2000. He advises maintaining 45% liquidity in treasuries, avoiding overvalued tech with deteriorating balance sheets, and buying energy dips while highlighting severe risks in private credit markets.

TED

Actor and producer Yara Shahidi shares how she navigated the collision of her acting career and Harvard education by rejecting binary choices, instead embracing optimism as an active practice and defining her 'highest order' priorities to find purpose amidst uncertainty.

NVIDIA AI Podcast

As AI tools generate over one billion lines of code daily, organizations face a critical bottleneck where manual code reviews cannot scale to catch the 41% increase in bugs, necessitating AI-driven review systems powered by sophisticated context engineering rather than traditional prompt engineering.