Sequoia Capital

Rich Sutton: synthetic data is "just a big mistake"

Agents Will Use The Web 1,000x More Than We Do | Parag Agrawal, Parallel Web Systems

Parallel’s Parag Agrawal: Building a New Web for AI Agents

Relearning to walk — what AI is missing | Khurram Javed, Oak Lab

Rich Sutton and Khurram Javed: Why AI Models Stop Learning, and How to Start It Again

The pursuit of artificial intelligence requires moving beyond static, human-curated datasets toward systems capable of continual learning through direct experience. Reinforcement learning pioneer Rich Sutton and co-founder Khurram Javed argue that the "Bitter Lesson"—the observation that computational scaling eventuall...

Continual Learning: How AI Agents Get Better With Every Use | Arjun Karanam, Trajectory

Closing the experience gap in AI agents requires shifting from static models to systems that learn from real-world user interactions. Current agents often lack the practical experience necessary to perform effectively, functioning as if every task is their first day on the job. Improving these systems involves four cri...

When to Build Your Own Agent Harness | Harrison Chase, LangChain

RL Environments Explained: How AI Agents Learn Real-World Work | Brendan Foody, Mercor

Post-Training Is How You Keep Your Taste | Fireworks CEO Lin Qiao

How Harvey Built a Research Lab on a Budget | Gabe Pereyra

Application-layer companies can compete with frontier labs by building specialized research labs on a budget, leveraging the existing AI ecosystem rather than attempting to replicate massive infrastructure. Success hinges on creating high-quality, domain-specific benchmarks and using synthetic data generation guided by...

How Companies Are Building Their Own Intelligence | Sonya Huang, Sequoia Capital

Sovereign AI represents a strategic shift where companies move from renting foundation models to owning their own intelligence to ensure independence and competitive advantage. This transition is driven by the need for cost efficiency, lower latency, and superior performance in domain-specific tasks. The "Not Your Weig...

Chai Discovery's Bitter Lesson: Drug Design Is Another Scaling Problem

Chai Discovery aims to transform drug discovery into an engineering discipline by utilizing foundation models to design molecules with high therapeutic precision. By shifting from traditional, serendipitous screening methods to a computer-aided design paradigm, the company enables researchers to iterate on molecular hy...

The Problem With Testing AI Architectures at Small Scale | Jerry Tworek, Core Automation

The Most Automated AI Lab Isn't Removing Humans | Jerry Tworek, Core Automation

Building the Automated AGI Lab: Core Automation's Jerry Tworek and Rohan Anil

Current transformer architectures face a significant bottleneck in scalability and adaptability, as they rely on static, lab-based training rather than continuous, real-world learning. While transformers have mastered pre-training and reinforcement learning, they struggle with catastrophic forgetting and data efficienc...

The Philosopher CEO | Clay Co-Founder Kareem Amin

Every CIO will have to answer for every token | Factory's Matan Grinberg

90% of AI tokens will be asynchronous | Matan Grinberg, Factory

Factory's Matan Grinberg: The Coming ‘Dark Factory’ Where Software Builds Itself

Why "Tokens Aren't Fungible" - Anthropic's Angela Jiang