Sapiens First launches to push AI safety politics

Sapiens First launched in May 2026 with a bold claim: AI safety won’t come from Silicon Valley alone—it’ll require a political movement. Founder rohantohab, a former Constellation employee and decade-long rationalist, warns that superintelligence is closer than benchmarks suggest, citing Mythos’ surprise release as evidence. The group’s manifesto argues that even well-intentioned politicians like Gavin Newsom or Donald Trump lack the urgency needed, pointing to the U.S. government’s COVID-19 response as a cautionary tale. Their solution? A coalition of mainstream AI reformers and grassroots activists, willing to set aside ideological discomfort to push for democratized AI governance.

The organization’s playbook draws from historical political movements, like the Civil Rights Act, which passed under a conservative president due to sustained public pressure. Sapiens First’s stance—'AI by the people, for the people'—frames governance as a bipartisan issue, though they acknowledge ideological friction is inevitable. Their Substack post invites debate on specific claims, signaling a data-driven, iterative approach to building consensus. For professionals in AI safety, this isn’t just another advocacy group—it’s a bet that the next decade’s existential risks will be solved in legislatures, not labs.