Mechanistic Interpretability Will Boom Early with High Unemployment and UBI

1a3orn · x · 2026-07-15

The author questions the logic in the book AI 2040, which predicts that mechanistic interpretability (MI) won't become truly useful until 2035.

The book envisions AI labor becoming ubiquitous by 2033, causing 50% unemployment and funding Universal Basic Income (UBI) through taxes, bringing the US median income to $190,000. The author points out that if such a massive and cheap AI workforce existed, it could be directly applied to MI research, accelerating progress by 100,000x compared to today. Therefore, there's no need to wait until 2035 for MI to take off.

Related event: Researcher Argues Mechanistic Interpretability Will Mature in the 2020s, Contradicting 'AI 2040'(7 posts)→

Original post →

More from AGI Musings

AGI Musings channel →