'Preparing for the Intelligence Explosion' argues AI-driven progress could compress a century into years

Philosophers William MacAskill and Fin Moorhouse, of the research group Forethought, published “Preparing for the Intelligence Explosion,” an essay arguing that if AI systems automate scientific and technological research, the resulting acceleration could compress roughly a century of further technological progress into a single decade.

Their argument rested on a scaling claim: once AI systems can substitute for human researchers, the total supply of research labour would no longer be capped by population growth, and could expand rapidly — by their estimate, conservatively around twelvefold a year and, on a more aggressive reading, fiftyfold. Applied across fields simultaneously, they argued this would force societies to confront, in years rather than generations, problems that historically arrived one at a time and gave institutions decades to adapt.

The essay listed nine “grand challenges” it said current governance had no process for handling: misaligned superintelligence; destructive technologies such as bioweapons; concentration of power via authoritarian control or coup; permanent ideological “lock-in”; the moral status of digital minds; space governance; competitive races to the bottom; disrupted shared epistemics; and the strains of sudden material abundance.

The piece was explicitly speculative rather than a forecast with a date attached; its authors argued the point was plausibility and stakes justifying preparation before the fact, not certainty. It became a frequently cited reference point in 2025’s debate over rapid, compounding AI-driven progress, alongside the parallel case made in Anthropic’s recommendations to the White House the same week.