Self-improving AI experiments demonstrate frontier labs no longer hold exclusive advantage.

Experiments in using AI systems to build and improve other AI systems suggest that the future of AI development may not belong exclusively to major frontier labs. The article explores open-source and smaller teams achieving self-improvement capabilities previously thought to require massive resources.