The future is unimaginable: AI can't contemplate, so good goals must be discovered by humans

sebkrier · x · 2026-09-09

curiousvii lays out his view on goal-setting in the AI era via three metaphysical claims: (1) the future is not only uncertain, unknowable and incalculable but actually unimaginable, making inputs from unbounded frontier exploration — scarce context, digital exhaust, embodied experience — essential to goal-setting; (2) good goals cannot be cast arbitrarily and must be discovered by contemplating that data, which is where he disagrees with synthetic-data and automated-research maximalists: AI models cannot contemplate, and can only serve as macroscopic lenses or sensor/actuator tissue, while only humans can do the final value judgment.

Original post →

More from AGI Musings

AGI Musings channel →