Averi Kitsch, Staff Software Engineer at Google, outlines a four-step evolution for securing AI agents that access databases, moving from dangerous, model-controlled tools to a zero-trust architecture. Drawing on insights from over 20 million monthly tool calls, the talk provides a practical roadmap for preventing data leaks by separating identities, constraining actions, and removing credentials and PII from the agent's control.
Varant Zanoyan, original author of the Chronon feature platform, introduces 'agentic experimentation'—a pattern where AI agents improve high-stakes ML systems without making live decisions. He explains how Chronon solves key challenges like infrastructure sprawl, safety, and reproducibility through a semantic API, branch-based isolation, and compute reuse, enabling agents to safely create production-ready pipelines for human review.
A tour of MLX, the on-device AI framework for Apple Silicon. This talk explores real-world applications from real-time vision and multimodal omni models to sub-100ms speech synthesis and video generation, all running locally. It highlights breakthrough techniques like Turbo Quant for 1M context and showcases community projects in robotics and native apps, arguing for a future where powerful AI runs without the cloud.
Predictive AI forecasts what will happen next based on historical data, while Generative AI creates new content by asking what something could look like. This summary explores their fundamental differences in outputs, data types, underlying models like transformers and diffusion systems, and how they can be used together in enterprise applications.
Alexander Taubman, CEO of Long Lake Management, details their pioneering 'AI take-private' strategy, using a horizontal AI platform called Nexus to acquire and transform traditional service businesses. He explains why ownership trumps SaaS for driving real-world AI adoption, focusing on growth and creating a positive-sum flywheel for employees and customers, exemplified by their acquisition of Amex GBT.
This talk explores the architectural shift required to build durable, long-running AI agents. It argues against traditional replay-based durability, proposing a two-part solution: treating LLM context as a durable, append-only log and managing the compute state (memory, files, processes) with OS-level snapshot and restore, implemented efficiently using Firecracker microVMs.