What use is prompting if there's ASI?

As artificial intelligence grows more capable, researchers grapple with a growing question: what future exists for human expertise once AI surpasses it? Per LessWrong, chess offers a troubling precedent—where superhuman machines left humans as 'intelligence consumers.' But evidence from programming and mathematics suggests experienced practitioners still matter as multipliers, guiding AI systems beyond what machines produce alone. To explore this, LessWrong contributors created a contest: write humanities essays on AI safety using large language models as collaborators. The results are encouraging. AI baseline essays achieved fifty-two percent of the evaluation rubric; human-guided entries reached seventy-eight percent. It suggests that human judgment might retain its role—not as the source of intellect, but as the hand that focuses AI toward its best outcomes.

Listen to this story

Hear this and more stories in a personalized audio briefing.

Open The Chonkerton