Apple's new ceo built the neural engine in every mac and iphone. his ai bet is "compress intelligence into the chip" not "build a bigger model"

Apple has confirmed a major leadership change: Tim Cook is out, and John Ternus—the hardware engineering chief behind the Apple Silicon transition and the Neural Engine—is in as CEO. Ternus has spent the past decade building the hardware foundation for on-device AI, and his appointment signals a clear strategic bet: Apple will prioritize compressing intelligence into its chips rather than chasing ever-larger cloud models.

This approach stands in stark contrast to Google's cloud-scale models and API access, Microsoft's Copilot ecosystem, and OpenAI's bet on GPT as a platform. The M4 chip's Neural Engine can already run mid-size models locally, powering Apple Intelligence features with zero latency, no API costs, and no reliance on external uptime. For developers and power users, this challenges the current default architecture of cloud API calls for coding tools, research assistants, and agent frameworks. Tools like Verdent and Continue already allow switching between providers, but Ternus's hardware-first bet could make local inference genuinely competitive—routing tasks between local and cloud based on what each task actually needs.