In a lengthy post on X, Microsoft's CEO said we can't keep treating AI as a stack of nested black boxes whose advice we simply accept, and that it's time to reassess the trust architecture.
Take: Coming from the guy who ships models to half the Fortune 500, this isn't philosophy — it's a market signal. Once the biggest distributor admits the goods are untrustworthy, 'verifiable AI' becomes the next thing you pay a premium for.