Beneficial AGI: The Path the Frontier Labs Are Missing

The dominant AGI narrative centers on a handful of frontier labs racing to scale ever-larger models. But as compute-to-capability gains flatten, a different path is emerging, one built on hybrid neural-symbolic-evolutionary architecture, decentralized infrastructure, and open governance rather than concentrated scale.

In this session, Ben Goertzel, an AGI researcher for over two decades and founder of SingularityNET and BGI Labs, discusses how the frontier labs' scale-first approach may be solving the wrong problem. He makes the case that genuine general intelligence requires multiple forms of cognition working together—reasoning, memory, and autonomous agency—not just bigger training runs.

The talk also explores why how AGI is governed may matter as much as how it's built, and why open, decentralized alternatives may offer a more resilient path forward. Attendees will leave with a sharper framework for evaluating competing AGI strategies.