The real leverage today isn't in the models themselves but in who controls the inputs, the infrastructure, and the narrative around risk. Jensen Huang's Japan sweep signals that hardware dominance and supply-chain relationships matter more than any single breakthrough in capability. Morgan Stanley's emergence as the chief architect of data center financing reveals where actual power is consolidating: not with the labs publishing papers but with the banks structuring the capital that makes those labs possible. Meanwhile, the hiring bias research and synthetic insider attack stories expose a quieter truth that regulation and ethics frameworks keep obscuring. AI systems don't need to be conspiratorial or superintelligent to cause real damage. They replicate human bias at scale because that's what's in the training data, and they enable credential fraud at speed because that's what criminals optimize for. The nonprofit Current AI pitch about building culturally inclusive AI and the Australian traffic light trial both hinge on the same question nobody in the press release wants to answer directly: who decides what counts as fair when the system makes the call? Universities teaching AI as a productivity tool rather than banning it sidesteps the harder problem. The EU AI Act compliance webinars and ethical AI blueprints are already being absorbed into the standard operating procedure of companies that can afford compliance consultants. The real winners today are the ones controlling the capital flows, the infrastructure, and the data pipelines. Everything else is theater.
Sloane Duvall