Open models, closed power: AI trust bottleneck tightens

Global active cure-related trials sit at 21,311 while, in the last 24h, 57 new Wire signals and 23 new Cure research signals were reported—evidence that both discovery and the meta-systems accelerating it are moving. The AI picture is mixed but unmistakably urgent: Z.ai’s open-sourcing of GLM-5.3-Flash signals wider access, yet Nvidia’s reported $12.9B acquisition of Hugging Face points toward centralized control, and AI’s ability to generate code faster than teams can verify it is creating a trust bottleneck that is breaking the old “trust-by-inspection” workflow. For medicine, this matters because faster iteration only helps if verification and trust keep pace—otherwise speed outruns safety. Still, the Cure pipeline shows real momentum: new AI-driven antibiotic discovery tooling (ApexGO), an FDA-approved TEVIMBRA-based regimen for HER2+ disease, and multiple positive clinical updates across inflammatory and oncology indications. The window is still open.