In the November 8 update of the Power On newsletter, Bloomberg's Mark Gurman reported that Apple's new CEO, John Ternus, is pushing software engineering chief Craig Federighi to speed up on AI and take a bolder approach, upgrading Siri into an AI agent that can autonomously do work for users. This urgency contrasts with Ternus's relaxed tone on last week's earnings call—where he praised the new Siri for finally fulfilling a vision years in the making and said he was "very pleased" with the feedback.
But behind closed doors, Ternus knows Apple is still far behind its rivals. At a closed-door meeting a few weeks ago, he bluntly expressed dissatisfaction that Federighi had not embraced AI as aggressively as peers, and demanded that development no longer settle for only necessary Siri upgrades.
An AI Skeptic's Tight Leash Caught between the CEO's expectations and the industry wave is Federighi's long-standing management habit. He is still seen as an AI skeptic, concerned about the impact on privacy, security, society, and jobs; when making decisions, he tends to get personally involved, debate fiercely with opponents, and use his familiarity with technical details to reject proposals he disagrees with, making some employees feel it is hard to influence product direction.
More troublesome is the mismatch in pace—he usually plans only 12 to 18 months ahead, while the hardware division is used to planning years in advance. The two rhythms are hard to mesh, slowing development and making first-generation products less bold. Apple wants to tighten the link between hardware and AI software, and how to reconcile the two planning cycles is exactly the hard nut Ternus has to crack. But Federighi is not without a track record: after Giannandrea failed to deliver on his Siri promises last year, it was Federighi who took over and, in a relatively short time, launched a new version of the assistant that actually works.
