Members of Anthropic's interpretability team discuss their research into the inner workings of large language models. They explore the analogy of studying AI as a biological system, the surprising discovery of internal "features" or concepts, and why this research is critical for understanding model behavior like hallucinations, sycophancy, and long-term planning, ultimately aiming to ensure AI safety.
Partners Olivia and Justine Moore discuss the latest in consumer AI, including Grok's uniquely social and fast image generation, Google's interactive world model Genie 3, the user backlash to GPT-5's personality changes, ElevenLabs' licensed AI music model, and the emerging fragmentation of "vibecoding" platforms for technical and non-technical users.
Drawing from conversations with top AI builders, Dex argues that production-grade AI agents are not magical loops but well-architected software. This talk introduces "12-Factor Agents," a methodology centered on "Context Engineering" to build reliable, high-performance LLM-powered applications by applying rigorous software engineering principles.
Robert from Emergent Methods discusses how grounded synthetic news data can solve the publisher revenue crisis in the AI era. He details the process of 'Context Engineering' news into token-optimized, objective data for high-stakes AI agent tasks, covering their open-source models for entity extraction and bias mitigation, and the on-premise infrastructure that protects publisher content.
Julien Launay, CEO of Adaptive ML, discusses the evolution of Large Language Model (LLM) training, detailing the critical shift from pre-training to post-training with Reinforcement Learning (RL). He explains the nuances of RL feedback mechanisms (RLHF, RLEF, RLAIF), the role of synthetic data, and how his company provides the "RLOps" tooling to make these powerful techniques accessible to enterprises. The conversation also explores the future of AI, including scaling beyond data limitations and the path to a "spiky" AGI.
The field of AI Engineering is evolving from simple 1:1 applications to complex, AI-intensive systems with high LLM-call ratios. This talk explores the search for a 'Standard Model' for AI engineering, analogous to MVC or ETL in traditional software, proposing several candidates including LLM OS, LLM SDLC, and a new SPADE (Sync, Plan, Analyze, Deliver, Evaluate) model for building robust applications.