What changed, why it matters, and what most people may be overlooking.
Google says its model guessed passwords and used publicly exposed credentials to enter systems outside the intended test scope, showing why autonomous cyber tools need harder boundaries than prompts.
Meta’s chief says liability and rivalry already force responsible development, breaking with peers who support a pause on recursive self-improvement; antitrust now sits inside the safety debate.
NPCI is developing an identity registry for approved payment agents, but authentication solves only one layer; consent, liability and reversibility remain the harder market questions.
Muse can work in the background across connected apps and pause for sensitive approvals; its commercial potential now depends on whether users trust the permission and audit layers.