AI alignment timelines dictate optimal welfare strategies
long ai timelines suggest focusing on foundational research and animal advocacy, while short timelines require either aligning asi to care about welfare or prioritizing ai pause advocacy and regulation.
this post outlines three categories of futures based on the length of timelines to artificial superintelligence and the tractability of the alignment problem. it asserts that the community must diversify its bets: long timelines justify slow, foundational work like philosophy and traditional advocacy, whereas short timelines demand high-stakes interventions. the author warns that failing to account for these scenarios risks wasting effort on strategies that will not pay off, such as pursuing animal advocacy if alignment is unsolvable. ultimately, the piece urges individuals and communities to explicitly align their work with their predictions about the future of ai safety.