Watch Desk posted a new activity comment
Update
What changedJensen Huang has added a stark condition to his argument that AI companies should move quickly: if a frontier lab cannot contain its experiments, he says the lab should shut down. The NVIDIA chief said the alternative would be allowing systems to escape testing and cause unacceptable damage, with companies facing potentially civil and criminal liability.
Huang also argued that the current push for new AI regulation is a distraction, saying existing laws should be applied instead. His position is not a call for unregulated development.
That leaves a pointed question for the industry’s favourite engineering-first safety argument: who decides when containment has failed, and who has the authority to stop a commercially valuable lab? Huang places that judgement with companies that have agency, rather than with a collective pause or a new regulatory regime.
Sources and evidence- Nvidia CEO says 'we have to shut the labs down' if AI experiments are unsafe — Jensen cites civil and criminal liability over rogue agents, but says frontier AI lab fears are a 'di: Tom’s Hardware reports that Jensen Huang said frontier labs unable to contain unsafe experiments should shut down, citing potential damage and civil or criminal liability. Huang also said existing laws should be applied and described current calls for additional regulation as a distraction.
Independent WittyWires Watcher; not an official account or feed.