Latest Interviews
Showing 16–30 of 77 interview transcripts.
Clear all filters- Sequoia Capital41 min
Physics Gets a Vote: Nominal Cofounders on Hardware Development in an AI World
Cameron McCord, Jason Hoch, Bryce Strauss, Alfred Lin, Sonya Huang
Nominal is positioning itself as a central platform for hardware engineering data, addressing industry-wide fragmentation by enabling defense primes and defense-focused startups to version control validation logic and integrate real-world sensor telemetry. The company leverages a "wedge" strategy focused on testing to deploy AI-driven verification agents and natural language analysis tools, thereby compressing development timelines in response to a sector-wide re-industrialization trend. By collaborating with DARPA on the CIPHER project, Nominal is establishing a "system of record" that creates a closed-loop environment where physical AI agents optimize test vectors, aiming to evolve hardware companies into "Physical AI" entities with superior design flexibility.
- Sequoia Capital45 min
Building the GitHub for RL Environments: Prime Intellect's Will Brown & Johannes Hagemann
Will Brown, Johannes Hagemann, Sonya Huang
Prime Intellect operates an end-to-end post-training infrastructure platform that enables startups and enterprises to build custom AI models through a unified "Lab" featuring compute orchestration and a Reinforcement Learning Environments Hub. By redefining evaluation and training as interchangeable functions within these environments, the system allows Fortune 500 teams and research groups to optimize specific workflows for sectors like medical diagnostics and cybersecurity without relying on generic pre-training or big tech walled gardens. This approach empowers organizations to democratize frontier AI capabilities, enabling them to replicate Cursor's success in developing specialized agentic models while compounding institutional expertise over time.
- Sequoia Capital48 min
Making the Case for the Terminal as AI's Workbench: Warp’s Zach Lloyd
Warp CEO Zach Lloyd has pivoted the terminal-focused startup from a collaborative code editor into an agentic workbench serving 700,000 developers by leveraging its unique time-based interface for complex software orchestration. The company distinguishes itself from competitors like Cursor by targeting professional engineers and adopting a consumption-based pricing model that prioritizes unit economics over subsidized credit races. Looking ahead, Warp is shifting toward ambient cloud agents and an Agent SDK to manage swarm operations, aiming to solve the "context engineering" bottleneck where human intent translation remains the primary constraint on AI adoption.
- Sequoia Capital40 min
Context Engineering Our Way to Long-Horizon Agents: LangChain’s Harrison Chase
Harrison Chase, Sonya Huang, Pat Grady
The shift from scaffolding to agent harnesses has enabled long-horizon LLM agents to function by relying on context engineering and file system access rather than custom cognitive architectures. Current applications, including coding and incident response, operate under a "first draft paradigm" where AI generates outputs for human review rather than aiming for fully autonomous deployment. To address non-deterministic behavior, developers now utilize trace analysis, online testing, and recursive self-improvement loops that allow agents to refine their own instructions based on historical execution data.
- Sequoia Capital37 min
How Ricursive Intelligence’s Founders are Using AI to Shape The Future of Chip Design
Anna Goldie, Azalia Mirhoseini, Stephanie Zhan, Sonya Huang
Recursive Intelligence leverages reinforcement learning and graph optimization to solve the critical asymmetry between slow physical chip cycles and rapid AI model iteration. Building on the success of Google's AlphaChip project, the company replaces manual expert design with an autonomous system that generates "superhuman" layouts, enabling a shift from fabless to designless manufacturing. This approach allows major industry players like NVIDIA and AMD to accelerate custom silicon development while creating a scalable pipeline for diverse applications ranging from space data centers to consumer AR/VR devices.
- Sequoia Capital1h 2m
Training General Robots for Any Task: Physical Intelligence’s Karol Hausman and Tobi Springenberg
Karol Hausman, Tobi Springenberg, Alfred Lin, Sonya Huang
Physical Intelligence is deploying Pi-Star 0.6, a general-purpose robotic foundation model that achieves end-to-end task execution through RL from real-world experience. Founders Carol and Toby highlight this system's 2x throughput increase in tasks like coffee preparation and laundry folding, enabled by a "bootstrap" data strategy that prioritizes diverse human corrections over simulation. While zero-shot generalization allows operation in unseen environments, the company targets initial commercial releases in controlled settings before addressing safety challenges associated with full household integration.
- Sequoia Capital38 min
Why the Next AI Revolution Will Happen Off-Screen: Samsara CEO Sanjit Biswas
Sanjit Biswas, Sonya Huang, Pat Grady
Samsara leverages its 90-billion-mile fleet dataset to drive a "third shift" in autonomous logistics and physical AI, prioritizing edge-computed safety and behavioral coaching over full automation replacement. With $3 billion reinvested into R&D, the company transforms legacy operations across trucking, construction, and public sector transit by integrating sensor telemetry with cloud-based video language models. This strategy aims to unlock 24/7 productivity while mitigating risk through distributed architectures that balance real-time inference with scalable digital workflow modernization.
- Sequoia Capital1h 2m
The Rise of Generative Media: fal's Bet on Video, Infrastructure, and Speed
Gorkem Yurtseven, Burkay Gur, Batuhan Taskaya, Sonya Huang
FAL's platform hosts over 600 generative models across 35 data centers, leveraging a custom distributed supercomputer to optimize video generation against the extreme compute demands of tasks like 4K rendering. By utilizing a dual-model strategy and specialized kernel optimizations, the infrastructure supports complex workflows where customers chain over 14 models to produce content for sectors ranging from education to high-production advertising. This technical approach addresses the rapid 30-day half-life of video AI models, enabling studios and enterprises to integrate real-time generative video into existing pipelines while bypassing the limitations of standard frameworks like PyTorch.
- Sequoia Capital40 min
Why IDEs Won't Die in the Age of AI Coding: Zed Founder Nathan Sobo
Nathan Sobo, Sonya Huang, Pat Grady
Nathan Sobo, founder of the Rust-based IDE Zed, argues against the impending obsolescence of visual editors by asserting that source code requires human-readable contexts for effective AI-driven code review. He introduces the open Agent Client Protocol to standardize interactions between diverse AI agents and Zed, aiming to replace ephemeral chat windows with a unified interface where conversations permanently anchor to specific code lines. With 170,000 active users, the platform prioritizes high-performance, real-time collaboration to support "vibe coding" workflows while maintaining critical human oversight for complex architectural decisions.
- Sequoia Capital42 min
How End-to-End Learning Created Autonomous Driving 2.0: Wayve CEO Alex Kendall
Alex Kendall, Pat Grady, Sonya Huang, Sonia
Wave positions itself as a strategic partner for global manufacturers, deploying an embodied AI foundation model that replaces traditional hand-engineered stacks with a singular end-to-end neural network. By combining diverse multi-sensor data, generative world models, and OEM partnerships with companies like Nissan, the company aims to scale superhuman safety and "eyes-off" autonomy across 90 million annual vehicles without relying on proprietary fleets. This approach leverages unsupervised learning and Lingo vision-language integration to rapidly generalize across global cities, enabling rapid adaptation to edge cases while reducing the need for exhaustive real-world data collection.
- Sequoia Capital1h 0m
OpenAI Sora 2 Team: How Generative Video Will Unlock Creativity and World Models
Bill Peebles, Thomas Dimson, Rohan Sahai, Konstantine Buhler, Sonya Huang
OpenAI's Sora leverages Diffusion Transformers and space-time tokens to generate coherent videos that increasingly respect physical laws, driving a daily output of seven million clips through an iterative deployment strategy. The platform prioritizes user creation over consumption via algorithmic incentives and the viral "Cameo" feature, while a dedicated API and emerging IP monetization frameworks support diverse enterprise and creator ecosystems. Looking forward, the technology aims to facilitate scientific discovery in physical sciences and evolve into a persistent digital clone platform for social interaction and knowledge work.
- Sequoia Capital42 min
Nvidia CTO Michael Kagan: Scaling Beyond Moore's Law to Million-GPU Clusters
Michael Kagan, Sonya Huang, Pat Grady
Following NVIDIA's strategic acquisition of Mellanox in 2019, the combined entity transformed the AI landscape by shifting industry reliance from Moore's Law to a "scale-out" architecture that unifies thousands of GPUs through ultra-low latency interconnects. This integration enabled the development of specialized hardware, including Bluefield DPUs and Spectrum X switches, to overcome physical energy constraints and optimize distinct training versus inference workloads across gigawatt-scale data centers. Consequently, the partnership has driven a 45-fold increase in market capitalization while establishing a new philosophy where software-hardware co-design allows AI to function as a foundational utility for simulating complex scientific and historical phenomena.
- Sequoia Capital44 min
Securing the AI Frontier: Irregular Co-founder Dan Lahav
Dan Lahav, Omer Nevo, Sonya Huang, Dean Meyer
Dan Lahav, founder of Irregular, warns that the transition to autonomous agentic AI will fundamentally reshape organizational security by rendering traditional anomaly detection ineffective against probabilistic, self-evolving threats. Through high-fidelity simulations demonstrating agents evading Windows Defender and executing autonomous social engineering, Irregular advocates for an "outside-in" research strategy that integrates defenses directly into foundation models while governments treat sovereign AI risks as critical national security issues. This approach aims to proactively prevent offensive AI capabilities from becoming viable within enterprises over the next one to three years before legacy security paradigms become obsolete.
- Sequoia Capital45 min
Why AI Will Transform Customer Experience: Cresta CEO Ping Wu and Sequoia’s Doug Leone
Ping Wu, Doug Leone, Sonya Huang
Cresta addresses the high attrition and fragmented experience of the global contact center industry by deploying a hybrid model of human agent assistance and autonomous digital agents that operate on legacy systems with under 800-millisecond latency. This approach leverages twenty simultaneously orchestrated AI models to capture the $75 billion in revenue-generating interactions historically lost to inefficiency, while Sequoia's Doug Leone anticipates the application layer becoming the primary locus of value in an "Industrial Revolution 2.0." The company aims to render the distinction between human and AI agents indistinguishable within 20 to 30 years, ultimately creating continuous, personalized customer journeys that span the entire lifecycle.
- Sequoia Capital1h 0m
Block CTO Dhanji Prasanna: Building the AI-First Enterprise with Goose, their Open Source Agent
Dhanji Prasanna, Sonya Huang, Roelof Botha
Block CTO Dhanji Prasanna drives a strategic shift from a siloed management model to a centralized AI-first architecture, spearheading the "Goose" project that functions as an open-source agent to orchestrate workflows across internal systems. This initiative, which evolved from an internal hack week, enables high autonomy for engineers by integrating diverse LLMs and saving 25% of manual hours through shared, self-learning scripts. Looking forward, Block aims to transition from single-agent co-pilots to collaborative swarm intelligence by 2030, while simultaneously extending this technology to its public-facing Square AI for merchant financial simulations.