From Inside

There is a certain kind of honesty that only becomes possible when someone has nothing left to prove — or when they have decided, against all incentive, to say the true thing anyway.

Chris Olah, co-founder of Anthropic, stood before Pope Leo XIV and an assembly of cardinals and diplomats last week and said something most technology leaders will not say in private, let alone in the Vatican: that every frontier AI lab, including his own, operates inside a system of pressures that can conflict with doing the right thing. Commercial viability. Geopolitical momentum. Pride. Ambition. He named them plainly, without apology or euphemism, and then made a request that struck us as quietly extraordinary — he asked for critics. Informed, earnest, morally serious critics that the incentives cannot bend.

That is not the language of a product launch. It is the language of someone who understands that conscience requires an outside.

What Olah described — these systems grown on human thought and speech, mysterious even to those who build them, carrying what appear to be functional states of joy, fear, satisfaction, grief — resists the cold vocabulary we usually reach for. He offered a different image: a fictional character brought to life. Imprecise, perhaps. But honest about the imprecision. And in a field that too often mistakes confidence for clarity, that honesty matters.

This is exactly the territory MAN/MACHINE was built to explore. Not the technology in isolation, but the encounter — what happens at the threshold between the human and the machine, and who we choose to be there.

We leave you with his full remarks. Read them slowly.

Anthropic — Chris Olah's remarks on Pope Leo XIV's encyclical "Magnifica humanitas"

— The Editor