Nadella: superintelligence must help humanity; firms should own their learning loops

IanAndrewsDC · x · 2026-09-15

Microsoft CEO Satya Nadella laid out principles for pursuing superintelligence: AI not helping humanity and not under human control isn't worth pursuing; benefits must be broadly diffused with both closed and open-source models thriving; and every firm should retain control of its tacit knowledge by building its own continuous learning loop without depending on any single model provider. The reposter plugs KeycardAI, arguing agents need new authn/z: delegated, task-scoped, revocable access with identity in every tool call and intent-based authorization.

Related event: Nadella on Superintelligence: Must Benefit Humanity and Stay Under Human Control(11 posts)→

Original post →

More from AGI Musings

AGI Musings channel →