From Agents to World Models: Why AI’s Next Leap Is Simulation

A co-hosted panel discussion and networking reception for AI founders, engineers, and investors.

Apr
30
2026

Date: April 30, 2026

Time: 4:00PM - 7:00PM (PST)

Location: The Cove Market Street San Francisco, CA

As artificial intelligence (AI) agents became more capable of using tools, coordinating workflows, and automating digital tasks, developers and founders began to consider what came next. The most challenging—and most valuable—problems in science and industry required systems that could simulate physics, test outcomes, and predict future states.

Basis Research Institute convened a conversation with operators, innovators, and Mintz attorneys to discuss where today’s agents reached their limits and what became possible when AI could model and simulate the world rather than simply navigate it.

Following the panel discussion, AI companies, venture capital funds, and investors from across the Bay Area connected over cocktails and hors d’oeuvres. The reception created space for networking between local investors and the founders and engineers advancing cutting‑edge AI work.

Who attended: AI founders, developers, engineers, chief technology officers, senior technology leaders, venture capital funds, and investors.

Discussion topics included:

  • Where agent systems faced limitations and how world models unlocked new capabilities
  • Why the next category‑defining AI companies might be built on prediction and simulation, not just automation
  • Data, intellectual property, and governance considerations for reasoning systems
  • Best practices for teams building and deploying these models
  • Insights from real‑world development and implementation

Speakers

Terri Shieh-Newton, PhD

Terri Shieh-Newton, PhD

Panelist

Terri Shieh-Newton, PhD, is a Mintz Member who guides clients on portfolio management and landscape analysis. She's a trained immunologist with a PhD in Cellular and Molecular Medicine, and she works with companies of all sizes in a wide range of life sciences and clean technology industries.
Dileep George

Dileep George

Panelist

Dileep George is an engineer, scientist, and entrepreneur focused on artificial intelligence, neuroscience, and robotics. He is currently at DeepMind following the acquisition of Vicarious, which he co founded, and previously co founded Numenta, where he helped develop Hierarchical Temporal Memory. He earned his PhD from Stanford University and has published influential work on brain inspired learning systems.
Zenna Tavares

Zenna Tavares

Panelist

Zenna Tavares is the co founder and president of the Basis Research Institute, where work focuses on building universal reasoning systems to address complex scientific and societal problems. Research centers on understanding human reasoning and developing computational tools for causal inference, probabilistic programming, and scientific model discovery, with applications across science and engineering. A PhD in Cognitive Science and Statistics was earned at MIT, followed by postdoctoral research at MIT’s Computer Science and Artificial Intelligence Laboratory and earlier work as an innovation scholar at Columbia University.