Playbook
Tasteful Skills
“Tasteful Skills” argues that the best agent skills are not documentation or best-practice lists.
Back to Podcast Digest
RTMH: Pope Leo's Magnifica Humanitas on AI
TL;DR
Paragraph 99 is the whole fight — Pope Leo explicitly says AI only imitates intelligence and does not experience joy, pain, love, responsibility, or moral conscience, and Zvi treats that as the document’s central error because it blocks serious engagement with AI cognition and existential risk.
The best part is paragraph 98’s humility about not understanding AI — Leo notes that modern systems are "more cultivated than built" and that even developers do not really understand their internal representations, a point Zvi says is unusually important and too often ignored.
Most of the document is really Catholic social doctrine plus AI, not an AI theory — the encyclical spends long stretches on dignity, labor, subsidiarity, solidarity, public oversight, migrants, schools, and the common good, with AI folded into a broader anti-technocratic moral framework.
Leo’s most concrete policy asks are about war, not AGI — he calls for humans to remain in the kill chain, for an identifiable chain of responsibility over lethal decisions, and for international agreements to curb AI arms races and autonomous weapons.
Jobs are a central Vatican concern — Leo repeatedly argues that states and firms should preserve meaningful work, redirect investment, regulate platforms, and judge economies by more than GDP, while Zvi says this labor-protection stance resembles a European technocrat playbook.
Chris Olah quietly pushed back from inside the Vatican — while endorsing the document’s moral ambition, the Anthropic co-founder said researchers are finding internal model states that functionally mirror introspection, joy, fear, grief, and unease, gently contradicting Leo without saying so outright.
The Breakdown
Pope Leo’s 82-page AI encyclical makes one big, loaded claim: AI is powerful but can never really think, feel, or count as a mind — and Zvi Mowshowitz argues that this punts the most important question of the decade. He praises the document’s moral seriousness and concrete asks on war and accountability, but says it mostly channels a European technocratic, labor-first worldview that misses superintelligence, existential risk, and the possibility that AI may become something more than a tool.
Share
Keep Reading
Two Ways to Follow
The Weekly Echo. The inbox-shaped summary of what mattered.
New editorials announced here.
