Why a decade of doomsday warnings failed to slow the AI race
TL;DR
For more than a decade, scientists and tech leaders have warned that superintelligent AI could threaten humanity, from Stephen Hawking's 2014 warning that it could spell the end of the human race to Jacob Coxon's viral resignation from Anthropic last week. Coxon declared human extinction imminent when he left. Those fears have repeatedly shaken the industry, yet they have never stopped the race to build ever more powerful systems. The piece traces why the warnings failed to change the pace of development.
Nauti's Take
One advantage of the long history of warnings: AI safety now has a research field and a vocabulary that policymakers can draw on quickly. The risk is credibility.
A decade of apocalyptic predictions without concrete milestones makes it easy for critics to dismiss new alarms as noise. What moves the industry is evidence from specific incidents, so researchers who publish reproducible failures will carry more weight than forecasts of extinction dates.
Sources