AI safety researcher argues funders should allocate $100M+ grants specifically for automated AI safety work to address short-timeline AI risks.

Author advocates for $1-50B annual AI safety funding across the ecosystem over the next 2-3 years, citing concerns about short timelines to advanced AI

Current AI safety funding levels and spending practices are viewed as insufficient and overly conservative in the past 5 years

Proposes that while traditional spending increases (salaries, organization growth) should continue, large-scale compute and API budget allocation for automated safety labor is necessary for differential impact