An engineering leader details the journey of self-hosting LLMs at enterprise scale, covering how his team slashed latency by 70% with TensorRT-LLM, optimized GPU costs through counterintuitive scaling, and built a verticalized AI platform for HR tech. The summary explores practical solutions for cold starts, KV cache optimization, and managing the cultural adoption of AI coding agents in engineering teams.
Cursor's team discusses their latest Cloud Agents launch, which gives agents full cloud VMs to test changes, record demo videos, and provide remote access. We explore parallel model swarms, bug reproduction workflows, and the future of agentic coding where throughput and new bottlenecks in review and CI/CD take center stage.
This talk explores the history of high-assurance software, from the near-perfect code of the NASA Space Shuttle to the Airbus A320. It argues that while these rigorous engineering and formal verification practices were once too expensive for mainstream use, the economic shift brought by AI and agentic coding can make aerospace-level reliability a practical reality for a wide range of applications, solving the key quality limitations of current AI-generated code.
Anthropic's Boris Cherny, creator of Claude Code, discusses the paradigm shift from traditional coding to agentic workflows. He covers the co-evolution of models and their "harnesses," the importance of "hackability" via features like slash commands, and provides practical tips for leveraging coding agents for tasks of varying complexity.