Latest Interviews
Showing 1–15 of 44 transcripts.
Clear all filters- Sequoia Capital26 min
RL Environments Explained: How AI Agents Learn Real-World Work | Brendan Foody, Mercor
Mercore has expanded its revenue run rate to $2 billion by transitioning the AI data market from basic crowdsourcing to high-skilled "agentic data" services that enable frontier labs to build complex reinforcement learning environments. The company leverages expert networks of lawyers, engineers, and doctors to create realistic simulated worlds with precise human-verified rubrics, demonstrating a fivefold increase in model performance on specific legal tasks during recent training. As the primary data vendor for major application layer companies, Mercore addresses the industry's need for ultra-long horizon tasks and social dynamics evaluations that synthetic models cannot yet self-generate.
- Sequoia Capital29 min
How Harvey Built a Research Lab on a Budget | Gabe Pereyra
Harvey, Gabe Pereyra, Brendan, Julio, Ross, Brock
Harvey differentiates itself from well-funded frontier labs by leveraging an application-layer strategy that combines synthetic data generation guided by domain experts with post-training on open-source models. The company builds specialized legal benchmarks and utilizes infrastructure partnerships to train agents on complex tasks like contract negotiation without exposing sensitive client information. By deploying these capabilities across multiple vendors and product surfaces, Harvey aims to solve organizational productivity challenges while mitigating the performance gaps inherent in current long-context environments.
- Sequoia Capital45 min
Memory and Continual Learning: Engram's Dan Biderman and Jessy Lin
Dan Biderman, Jessy Lin, Sonya Huang, Shaun Maguire
Ngram addresses the scalability and cognitive limitations of current retrieval-augmented generation by training custom, continually learning models directly within workspace environments. This approach utilizes adapter fine-tuning to internalize organizational knowledge into model weights, reducing inference token consumption by a factor of 100 while enabling true intuition rather than static fact retrieval. By shifting the focus from pre-training generic AI to perpetual, private adaptation, the platform aims to create personalized neural interfaces that evolve alongside a team's data.
- Sequoia Capital39 min
Simulating Humans at Scale: Simile's Joon Sung Park
Founded by Stanford alum Jun with co-founders Percy Liang and Michael Bernstein, Simile is an applied AI lab that has transitioned from the experimental "Smallville" project to a commercial platform capable of simulating complex human societies. By combining proprietary behavioral data with foundation models, the company validates its SaaS service with an 85% prediction accuracy against real-world self-reports, allowing enterprises like CVS to test strategic decisions without the cost of live field trials. This approach leverages specific statistical metrics and reinforcement learning to bridge the gap between attitudinal data and actual behavior, offering a scalable alternative to traditional polling for forecasting long-term market and societal impacts.
- Sequoia Capital40 min
Knowing What Your Customers Want, All the Time: Listen Labs' Alfred Wahlforss
Listen Labs leverages its database of 30 million participants to transition from AI-conducted interviews to generative agent simulations that predict individual customer preferences with up to 95% accuracy. Founder Alfred Walforce directs the platform to serve Fortune 500 clients like Microsoft and Chubbies by identifying specific product friction points and integrating insights directly into autonomous coding workflows. This approach creates a proprietary feedback loop where behavioral data refines agent models, allowing businesses to validate product decisions without the churn and cost of traditional market research.
- Sequoia Capital25 min
Neuralink's DJ Seo: Inside the Race to Connect Brains and AI
DJ Seo, Shaun Maguire, Sean McGuire
Neuralink is advancing its brain-computer interface technology through the deployment of two therapeutic products, Telepathy and Blindsight, to restore motor function and vision for patients with conditions like ALS and total blindness. The company has implanted its devices in over 20 human participants and reports successful trials where users regained computer control or visual perception, supported by a strategy of vertical integration and an "all-green light" engineering framework. Founder J.D. McGuire envisions these medical breakthroughs as a precursor to creating an "exocortex" that merges human intent with AI, pending regulatory approval for broader non-medical augmentation.
- Sequoia Capital35 min
Suno's Mikey Shulman: Everyone Can Make Music Now
Suno, founded by quantum computing PhD Mikey Shulman, has disrupted the music industry by modeling audio as continuous sound waves rather than discrete musical notes, enabling the generation of full songs with custom lyrics and vocals. The platform has achieved significant market traction by prioritizing user creativity over passive consumption, evidenced by a 90% creator user base, a landmark partnership with Warner Music, and chart-topping commercial successes. Future developments aim to deepen this position through social co-creation tools, voice cloning features, and interactive concert technologies designed to seamlessly integrate AI into professional music production.
- Sequoia Capital26 min
ElevenLabs' Mati Staniszewski: How Voice Becomes the Interface for AI
Mati Staniszewski, Andrew Reed, Alfred Fuller, Gabriel Sanchez, Konstantin
Established in 2022 by high-school friends Gabriel Sanchez and Piotr to address monotone localization in Poland, Eleven Labs has rapidly scaled to generate over $400 million in revenue with a flat, title-less organization of 400 remote employees. The company distinguishes itself by prioritizing immediate monetization and emotional nuance in audio synthesis, evolving its technology from text-to-speech to complex voice agents deployed by governments, enterprises, and educational platforms. With a roadmap focused on emotional intelligence and the integration of audio as the primary interface for robotics, the firm maintains defensibility through proprietary data and a vast ecosystem of user-contributed voices.
- Sequoia Capital27 min
Waymo's Dmitri Dolgov: 20 Million Rides and the Road to Full Autonomy
Dmitri Dolgov, Konstantine Buhler
Dmitry Dolgov, Waymo's co-founder, outlines a 21-year evolution from DARPA challenges to a sixth-generation autonomous fleet that has achieved 200 million fully autonomous miles and delivered over 20 million rides with a safety profile 13 times superior to human drivers. The company's architecture utilizes a multimodal "Foundation Model" integrating physics reasoning and spatial understanding to power a structured, end-to-end system across 11 cities, now accelerating toward global expansion in London and Tokyo through the newly launched Waymo Origin vehicle. This mission-driven approach bypasses incremental improvements to rapidly scale commercial deployment, treating full autonomy as an existential imperative to eliminate the 13 million annual road fatalities while transitioning the core technology into a primary family transportation tool.
- Sequoia Capital28 min
OpenAI's Greg Brockman: Why Human Attention Is the New Bottleneck
OpenAI executives estimate they are 80% toward achieving AGI, a milestone driven by aggressive compute acquisition strategies and architectural innovations that currently enable AI to autonomously engineer software kernels. The organization is pivoting toward an enterprise and consumer model where a single AGI entity handles complex goals, urging startups to leverage agentic coding tools for massive productivity gains while navigating a landscape where human attention has become the primary bottleneck. This shift is underpinned by anticipated scientific breakthroughs in physics and biology, alongside internal reforms designed to manage the risks of autonomous agents within a future where humans oversee teams of AI rather than writing code manually.
- Sequoia Capital27 min
Demis Hassabis: We're Three Quarters of the Way to AGI
Demis Hassabis, Konstantine Buhler, Dennis
Demis Hassabis founded DeepMind with the strategic intent of leveraging game development revenue to fund the pursuit of artificial general intelligence, a goal he predicts will be realized by 2030. Following a pivotal breakthrough with AlphaFold that solved the protein folding problem, the organization expanded into Isomorphic Labs to accelerate drug discovery and is now pioneering the application of AI for scientific simulation in fields like biology and economics. Hassabis maintains that classical neural networks are sufficient to model complex biological systems and views information as the fundamental building block of the universe, positioning AI as the ultimate tool for understanding reality.
- Sequoia Capital44 min
Greetings, Earthlings: Philip Johnston of Starcloud on Data Centers in Space
Philip Johnston, Sonya Huang, Pat Grady
StarCloud founder Philip Johnson is deploying a Low Earth Orbit constellation designed to host data centers that will eventually outpace terrestrial alternatives in cost and scale, aiming for 50% of new AI compute capacity to be space-based within a decade. Leveraging SpaceX Starship launches, the company plans to deliver 10-megawatt satellite clusters by 2028 that prioritize thermal management over radiation shielding to support high-density inference workloads for hyperscalers and government clients. This infrastructure model targets critical market gaps in edge computing and downlink bandwidth while positioning space-based hardware as a neutral, scalable alternative to traditional cloud providers.
- 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 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.