Latest Interviews
Showing 5266–5280 of 6,403 interview transcripts.
Clear all filters- Lex Fridman12 min
François Chollet: History of Keras and TensorFlow | AI Podcast Clips
Initiated in March 2015 by François Chollet, the Keras framework was developed to streamline deep learning model definition by replacing static configuration files with dynamic Python code. After Chollet joined Google, he led the project's integration into TensorFlow starting in late 2015, eventually transitioning Keras from a backend abstraction layer to a core component of the ecosystem. This collaboration culminated in TensorFlow 2.0, which unified high-level Keras usability with low-level research flexibility to serve diverse user needs from rapid prototyping to custom training loops.
- Lex Fridman17 min
Gary Marcus: Limits of Deep Learning | AI Podcast Clips
Yann LeCun's critique of contemporary deep learning argues that current systems rely on statistical correlations rather than causal models, failing to grasp fundamental concepts like common sense or physical object permanence. The speaker contends that achieving robust intelligence requires a hybrid approach combining data-driven methods with symbolic AI to explicitly encode logical variables and structural rules. Consequently, the presentation rejects the notion that pure end-to-end learning can replace human engineering, advocating instead for continued manual specification of abstractions to ensure reliability in real-world applications.
- Lex Fridman10 min
Gary Marcus: Nature vs Nurture is a False Dichotomy | AI Podcast Clips
The speaker challenges the false dichotomy between innate biology and learning by arguing that intelligent systems require pre-encoded frameworks derived from evolutionary history, such as the vertebrate brain's reuse of genetic "libraries" for spatial and causal reasoning. Citing examples like baby ibex navigating physics, the presentation asserts that engineers can accelerate AI development by practicing biomimicry and incorporating cognitive insights from fields like developmental psychology and dognition. This approach posits that mimicking the cumulative strategies found in nature is more effective than starting from scratch, allowing for the rapid optimization of complex problem-solving capabilities without relying on slow, independent trial-and-error processes.
- Lex Fridman9 min
Jeremy Howard: Deep Learning Frameworks - TensorFlow, PyTorch, fast.ai | AI Podcast Clips
The speaker traces the industry's evolution from static frameworks like Theano to interactive environments such as PyTorch and Fast.ai, highlighting how the latter's multi-layered API reduces boilerplate while preserving low-level control. Despite these advancements, the discussion identifies persistent performance bottlenecks in Python-based systems and criticizes TensorFlow 2.0's sluggishness compared to PyTorch, attributing these issues to legacy technical debt. Looking forward, Swift for TensorFlow is positioned as a future solution for high-performance computing, though widespread adoption remains years away due to current gaps in tooling and Apple's limited support for numeric programming.
- Lex Fridman6 min
Yann LeCun: Human-Level Artificial Intelligence | AI Podcast Clips
Building truly autonomous AI requires systems to develop human-like world models through self-supervised learning, mimicking the cognitive milestones infants achieve within their first year of life. Current architectures depend on integrating predictive simulation capabilities with objective functions rooted in biological drives, yet failure often stems from misaligned goals or an inability to compute optimal action sequences. Overcoming these hurdles involves addressing the exponential complexity of real-world problems that historical optimism in general problem solving failed to anticipate.
- The Economist36 min
Chimamanda Ngozi Adichie: identity, feminism and honest conversations
Chimamanda Ngozi Adichie, Sacha Nauta
Chimamanda Ngozi Adichie explores the fluid nature of identity and critiques the "own voices" movement for potentially flattening literature, arguing that storytelling requires inhabiting experiences beyond one's own. She explicitly rejects sensitivity screening and demands that authors write with honesty rather than being constrained by a fear of offense, while advocating for a feminism that actively includes men to dismantle harmful socialization. Ultimately, Adichie calls for a world where individual merit supersedes identity-based judgments, urging creators to humanize marginalized groups without sacrificing the complexity of their characters.
- Lex Fridman1h 25m
Gary Marcus: Toward a Hybrid of Deep Learning and Symbolic AI | Lex Fridman Podcast #43
Gary Marcus argues that current artificial intelligence lacks the common sense and causal reasoning required for general intelligence, necessitating a shift from pure deep learning to a hybrid architecture that integrates symbolic logic. He contends that true "trustworthy AI" demands the explicit engineering of abstract ethical concepts and diverse testing frameworks, such as a "Turing Olympics," rather than relying on statistical correlations or black-box scaling. Ultimately, Marcus predicts a gradual evolutionary path where AI acquires physical and psychological understanding by mimicking human innate cognitive libraries, rather than through a singular disruptive breakthrough.
- Goldman Sachs23 min
Evan Spiegel, Co-Founder and CEO of Snap Inc.
Snap Inc. is capitalizing on accelerated revenue and user growth driven by Android product improvements, verticalized sales teams, and a premium content strategy that has boosted engagement among its core 13-to-34 demographic. The company differentiates its platform through high user retention rates, proprietary Distributed Device Machine Learning for on-device privacy, and a curated media model that contrasts with open social networks while expanding into gaming and augmented reality. Looking forward, leadership emphasizes sustainable business iteration and deliberate regulatory frameworks to position the platform as a hub for future visual communication and mass distribution.
- The Economist11 min
How to help America's poor
Idris Kahloun, Lauren Jones, John Prideaux
A recent analysis reveals that the United States relies on an obsolete 55-year-old poverty metric that fails to account for modern expenses like childcare and housing, leaving one in six Americans, particularly children, trapped in geographically concentrated deprivation. Field observations in suburban communities like Waukegan illustrate how the "working poor" face severe instability despite owning assets, as rising costs for diapers, internet, and rent force families into desperate survival strategies. To address these systemic failures, experts propose replacing fragmented subsidies with universal child benefits, a policy shift argued to be both more cost-effective and politically viable than current means-tested approaches.
- Lex Fridman8 min
Peter Norvig: We Are Seduced by Our Low-Dimensional Metaphors | AI Podcast Clips
Challenging the reliance on static explanations, the speaker argues that establishing trust in neural networks requires adversarial bias detection and dynamic conversations about decision-making within high-dimensional data spaces. The analysis contrasts AI's rigorous proof-of-worthiness requirements with human social trust, highlighting how robustness must be achieved through rigorous testing rather than simplified low-dimensional metaphors. Furthermore, the discourse warns against overestimating AI as the primary driver of systemic change, attributing greater influence to underlying communication technologies that facilitate global data collection.
- Lex Fridman1h 3m
Peter Norvig: Artificial Intelligence: A Modern Approach | Lex Fridman Podcast #42
This discussion with Peter Norvig outlines the evolution of artificial intelligence from memory-constrained logic to modern neural networks, highlighting a philosophical shift toward defining utility functions and addressing ethical challenges like algorithmic fairness. The dialogue examines the limitations of deep learning in reasoning, the necessity of combining symbolic AI with neural approaches, and the societal risks of attention economies and weaponization rather than existential robot threats. Furthermore, Norvig reflects on the changing nature of programming expertise, the unique dynamics of online education, and future research directions focused on integrating common sense reasoning into code assistants and natural language systems.
- Lex Fridman6 min
Leonard Susskind: Richard Feynman and Intuitive Visualization vs Rigorous Mathematics
Leonard Susskind, Richard Feynman
The speaker argues that while deep intuition and visualization can validate alternative physics methodologies, human neural architecture remains fundamentally constrained by a three-dimensional framework that limits the natural comprehension of higher dimensions. This cognitive limitation suggests that even with specialized training, abstract concepts in quantum mechanics and string theory cannot be fully internalized as purely natural visual experiences, though artificial systems might potentially overcome these biological barriers. Consequently, the dialogue highlights a persistent gap between human intuitive capabilities and the mathematical reality of modern theoretical physics.
- Lex Fridman6 min
Leonard Susskind: The Power of Quantum Computers | AI Podcast Clips
Experts identify quantum computing's primary value as simulating complex quantum systems intractable for classical methods, with broad applications spanning chemistry, material science, and black hole physics. While the technology promises to analyze macroscopic materials like superconductors through controlled manipulation, the speaker maintains skepticism regarding its application to the human brain, which currently lacks evidence of intrinsic quantum features. This distinction underscores the shift from narrow algorithms to systemic simulations capable of addressing fundamental limits in physics and biology.
- Goldman Sachs19 min
David Zaslav – Discovery President and CEO
Discovery is executing a strategic pivot from linear broadcasting to a targeted direct-to-consumer ecosystem, leveraging its top-ranked female viewership to drive a "view and transact" model that integrates commerce into niche apps like HGTV Go and Discovery Go. Supported by significant investments in proprietary technology and engineering talent from Amazon, the company is building a global platform designed to capture higher ad revenue per impression while diversifying into owned content like Golf Digest. This approach allows Discovery to thrive amidst accelerating cord-cutting trends by offering advertisers a safe, data-rich environment that captures a massive audience of women on Sunday nights, ensuring revenue growth even as traditional cable bundles face structural pressures.
- Goldman Sachs14 min
Michael Rapino – Live Nation Entertainment President and CEO
Live Nation aims to expand its 30% global market share to 125 million fans by targeting under-monetized regions in Latin America and Asia while leveraging vertical integration in established markets. To drive economic resilience and revenue optimization, the company is implementing dynamic pricing models, enforcing "Verified Fan" protocols, and shifting toward earlier on-sale cycles to combat secondary market speculation. This strategic approach capitalizes on the recession-resistant nature of the experience economy to secure long-term growth in arena development and primary ticket sales.