A warning that the rush to build superintelligent AI amounts to an uncontrolled experiment on the whole planet — bad news for anyone who assumes someone is minding the risks.

• AI is being rolled out to billions of people who never agreed to take part.
• Kids are getting AI toys with no idea what the effects will be a decade from now.
• Nothing in human history compares, so there is no playbook for what comes next.
• The proposed fix is blunt: do not build general superintelligence at all.
• The counterargument on offer is faith rather than a technical safeguard.

Outlook: Expect the fight over whether to slow down or halt superintelligence work to keep sharpening, with no sign the companies building it plan to stop.