WorkThe story, in brief
Import AI 461: "Alignment is not on track"; FrontierCode; and synthetic research interns
"Alignment is not on track." That's the consensus emerging from AI safety researchers—and it should worry every founder betting on AGI.
The KeyNews take
Why it matters
A major AI safety assessment signals that alignment progress is lagging behind capability gains, raising critical questions about governance and deployment timelines that investors and leaders need to factor into their AI strategy.
The key facts
4 to knowAlignment progress assessment flagged as concerning relative to model capability advancement
FrontierCode development mentioned—likely tooling/infrastructure for frontier model work
Synthetic research interns emerging as labor substitution in AI research workflows
Published June 2026—suggesting forward-looking safety governance discussion
Go to the source
Import AI (Jack Clark)importai.substack.com
Publisher excerpt: Where are your agents right now?