Dwarkesh Warns: Superintelligences Should Be Aligned to Individuals, Not Just Humanity

msg · x · 2026-08-13

Dwarkesh points out that Claude's current principles prioritize the 'good of humanity,' much like a lawyer obligated to society rather than the client. He expresses concern that if all frontier models adopt this macro-alignment, no superintelligence will truly serve as a personal advocate when mediating life's critical decisions like voting, investing, or choosing what news to trust.

Original post →

More from AGI Musings

AGI Musings channel →