Latest Interviews
Showing 271–285 of 1,042 transcripts.
Clear all filters- a16z29 min
AI at the Intersection of Bio | Vijay Pande, Surya Ganguli & Bowen Liu
Vijay Pande, Surya Ganguli, Bowen Liu
This event details the transformative shift in drug design from theoretical potential to operational reality, leveraging deep learning and massive protein sequence datasets to navigate the vast chemical space while addressing historical validation bottlenecks. By integrating single-cell foundation models and latent space theories, the discussion outlines strategies to overcome distributional generalization failures and reduce the 90% clinical trial failure rate through personalized medicine and optimized patient selection. Ultimately, the presentation projects a decade-long roadmap toward "digital human" simulation systems that aim to slash development costs and expand FDA-approved target coverage despite significant regulatory and interpretability hurdles.
- a16z1h 17m
The Quest for Community-Trained Open Source AI Models
Bowen Peng, Jeffrey Quesnelle, Anjney Midha
News Research has unveiled the Distro method, a decentralized training framework that enables the creation of state-of-the-art "Hermes" language models using only standard internet connections and consumer-grade hardware. This breakthrough achieves an estimated 857-fold reduction in bandwidth requirements by allowing individual nodes to train independently and exchange only high-value insights rather than full model weights. By demonstrating that global AI development can be replicated without reliance on high-end data centers or single-entity resources, the project aims to democratize access to foundational models while preserving the neutrality and open nature of the technology.
- a16z35 min
Why Human Data is Key to AI: Alexandr Wang from Scale AI
Alexandr Wang, David George, Sarah Wing, Alex Wang, Max Wiethe, Dan Morehead, Mike Greenleaf, Alex Rosenberg
Scale AI is positioning itself as the essential data foundry for the AI industry by combining human expertise with algorithms to generate frontier data, addressing the impending "data wall" that public sources can no longer satisfy. While the current market phase focuses on engineering execution with abundant compute, the speaker predicts a future shift where algorithmic innovation and proprietary enterprise data will define competitive advantage as intelligence becomes a commodity. This strategy is supported by a hiring philosophy prioritizing maximum talent density and the belief that major tech firms must aggressively invest in AI to mitigate existential risks and capture value through workflow integration.
- a16z48 min
“The Future of AI is Here” — Fei-Fei Li Unveils the Next Frontier of AI
Fei-Fei Li, Justin Johnson, Martin Casado
Fei-Fei Li and Justin Johnson have founded World Labs to commercialize "spatial intelligence," a paradigm shift that moves beyond 2D language and image processing to enable machines to reason about 3D geometry, physics, and temporal dynamics. The venture leverages expertise from NeRF pioneer Ben Mildenhall and graphics expert Christoph Lassner to build intrinsic 3D representations capable of powering advanced applications in robotics, augmented reality, and synthetic world generation. This initiative marks a strategic evolution from the current multi-modal AI era toward a future where artificial systems possess a comprehensive "spatial brain" for interacting with the physical universe.
- a16z30 min
Luma's Dream Machine and Reasoning in Video Models
Luma released Dream Machine, a foundational video generative model that leverages massive 2D data scaling to achieve robust text-to-video and image-to-video synthesis with emergent 3D structural reasoning. The model implicitly simulates complex physical phenomena, such as depth perception, light transport, and causal character interactions, enabling consistent scene reconstruction from single inputs without native 3D priors. While currently classified as a research preview, the development roadmap aims to evolve the technology into a 4D spatiotemporal simulator and multimodal agent capable of handling intricate narrative and interactive requirements.
- a16z35 min
Implementation, Data, Impact of Healthcare AI with Julie and Vijay
Industry leaders project that AI agents will reduce healthcare costs by up to 50% through labor arbitrage and prevention-focused value-based care, while simultaneously compressing training times from months to days. Despite these economic advantages, widespread adoption faces significant hurdles including regulatory ambiguity over AI credentialing, the inelastic nature of healthcare demand, and the critical need to shift from retrospective records to prospective data generation. To overcome these barriers, organizations are reclassifying technology as labor, utilizing AI as a recruitment tool to reduce clinician burnout, and transitioning toward hybrid models where human oversight manages complex exceptions while automated systems handle routine phenotyping and triage.
- a16z6 min
Udio: From Text to Tune
Udeo is a generative music service designed to transition AI tools from niche experiments into standard components of both professional studio workflows and casual mobile creation. The platform's commercial viability was recently validated when a track generated by its system was sampled by producer Metro Boomin for a collaboration with Slayyyter and Drake. Future development aims to expand these capabilities, allowing users to remix AI-generated clips, add new instruments, and build entire musical pieces based on their creative intent.
- a16z6 min
Krea: Building the UI for AI Art
CREA, founded by CEO Victor and CTO Diego, empowers non-technical artists to generate high-fidelity imagery by combining text prompts with intuitive visual controls like shape placement and creativity sliders. The platform gained rapid traction by swiftly adapting to viral trends, such as a 48-hour spiral image tool, and operates on an agile model that releases specialized AI tools every six months. This approach aims to automate routine creative tasks while maintaining professional-level output, with the ultimate goal of merging the utility of Adobe and Autodesk with Pixar-quality rendering into a single ecosystem.
- a16z37 min
How Waymo Is Using GenAI to Build a Better Driver
Dmitri Dolgov, David George, Sarah Wang
Dimitri Dolgov traces the evolution of autonomous driving from the 2007 DARPA Urban Challenge to Waymo's current deployment of generalizable AI across four major US cities. By combining foundational machine learning breakthroughs like Transformers with massive synthetic simulation, Waymo has demonstrated a 76% reduction in property damage and operates 24/7 in diverse weather conditions. The company now focuses on scaling complex urban logistics while maintaining a hybrid market strategy that partners with Uber and prioritizes rigorous safety validation over short-term commercial shortcuts.
- a16z31 min
AI in Pharmaceutical R&D with Kim Branson
Kim Branson, GSK's Global Head of AI/ML, is driving a strategic initiative that creates a "sheltered bubble" to rapidly build AI capabilities before integrating them into the company's broader drug discovery workflow. His team currently leverages active learning loops and computational pathology to accelerate target identification and clinical trial optimization, achieving twentyfold reductions in screen times and enabling precise patient stratification. Branson emphasizes that the industry's primary bottleneck is access to longitudinal data with outcomes, positioning GSK to transition from simple algorithmic efficiency to a data-driven "juggernaut" defined by proprietary datasets and mechanistic biological integration.
- a16z7 min
Civitai: Open Sourcing AI Art
Justin Maier, Maxfield Hulker, Maxville Holker
Co-founded by Justin Mayer and Maxville Holker, the platform Civitai rapidly scaled to one million users by March 2023 to consolidate open-source image technology and host community-created models. The site differentiates itself through detailed metadata tracking, automated content safety via Safe Prompt Models, and integrated tools that allow users to train custom LoRAs from uploaded data. By balancing unrestricted creative expression with necessary moderation, the organization currently fosters a growing creator economy where developers can monetize their work without sacrificing open accessibility.
- a16z24 min
Mark Casey (Capital Group): Public Market and Macro Perspectives
Capital's long-term value investing strategy employs a four-to-eight-year horizon and a compensation structure weighted toward multi-year results to prioritize earnings convergence over short-term market noise. The firm actively manages portfolio construction by identifying unstoppable secular trends like cloud computing and e-commerce while criticizing deceptive metrics such as adjusted EBITDA in favor of free cash flow per share. Additionally, the approach emphasizes superior capital allocation by citing industry leaders like Warren Buffett and Hock Tan, while maintaining bullish positions on high-moat fintech and semiconductors tied to the AI revolution.
- a16z1h 31m
Trump Vs. Biden: Tech Policy
Trump, Biden, Marc Andreessen, Ben Horowitz
Marc Andreessen and Ben Horowitz publicly endorse Donald Trump over Joe Biden, arguing that the current administration's hostile regulatory stance toward cryptocurrency, artificial intelligence, and venture capital threatens the future of American innovation and national security. The speakers contrast the Biden administration's restrictive executive orders and proposed unrealized capital gains taxes with Trump's platform promising to repeal these measures, defend self-custody rights, and accelerate technological competition against China. Citing the recent assassination attempt on Trump as a demonstration of the unique leadership required for this era, the venture capitalists conclude that the "little tech" ecosystem faces an existential crisis under current policies that necessitates a 180-degree strategic shift.
- a16z1h 55m
Marc Andreessen on Inventing The Internet Browser (How We all Access the Internet)
Drawing from Mark Andreessen's journey from a rural Wisconsin farm to the University of Illinois, the event details the 1992-1993 creation of Mosaic, which defied the era's academic consensus by introducing a graphical web interface that prioritized openness over binary efficiency. Following a contentious split with the University of Illinois and its commercial licensee Spyglass, Jim Clark recruited Andreessen to found Netscape in 1994, catalyzing a legal battle and a Microsoft counter-offensive that ultimately collapsed Spyglass's business model while democratizing global internet access. The narrative extends beyond this history to warn of current risks, contrasting the successful "hyperstition" of early web pioneers with modern fears of regulatory capture and the incompetence of established experts in predicting technological evolution.
- a16z38 min
Grand Challenges in Healthcare AI with Vijay Pande and Julie Yoo
Amidst a healthcare labor crisis and provider burnout, AI adoption is accelerating through administrative efficiencies in revenue cycle management and the digitization of complex payer-provider contracts. Strategic startups are targeting near-term opportunities by embedding natural language interfaces into existing Electronic Health Records, automating prior authorization, and restructuring clinical trials through real-world data analysis. While full-stack AI clinicians remain a distant goal, the immediate landscape prioritizes behavioral change strategies, unbundled care roles, and value-based care alignment to optimize patient outcomes and reduce systemic waste.