Is Apple playing Second Mouse in AI?

Disclosure: I've held Apple since 2012, and have added to the position multiple times up until 2018.

I've been thinking a lot about Apple recently, and I wanted to examine the question, "are they holding off on diving into AI until use-cases reveal themselves".

Because, to put it lightly, they are FAR BEHIND the result of the hyperscalers on AI offerings.

I think this could be a feature, not a bug.

This "second mouse gets the cheese" idea, plus their advances in security/privacy, really make me think they are positioning the company for longterm success. I believe privacy will be a symbol of tech-luxury as large data models converge with AI-use cases.

So, is Apple just old and crumudgeonly and unable to innovate, or are they doubling down on their luxury branding.

Let's see!

## Apple and AI today

• Apple has been delaying key AI features (notably Siri / “more personal Siri”) — some were promised early in iOS updates and later postponed. The Siri revamp is often cited as late or buggy. 

• Apple’s AI strategy emphasizes on-device processing, privacy, and selective feature rollouts, rather than full-blown cloud / large model dominance. 

  • Many analysts feel Apple is behind competitors in foundational AI / large language model development. 

  • Internally, Apple is experimenting (e.g. “Veritas” chatbot internally) but not releasing broadly. 

  • Apple has pulled back some marketing claims; e.g., removing “available now” from its AI page under pressure from NAD (advertising oversight). 

• Apple reportedly reprioritized development: it halted a Vision Pro overhaul to shift resources to smart glasses and AI focus. 

  • Unrelated, but I think the "liquid glass" UI update is preparing for a mixed-reality headset (esp after seeing the early success of Meta's Glasses).

• The difficulties of compressing large models to run efficiently on device is nontrivial. A recent research paper on model compression and on-device ML includes interview study with Apple experts (link) 

This wait and see attitude isn't necessarily bad imo, first-mover advantages are often talked about, but almost no one talks about first mover-disadvantages (high R&D costs, risk of misreading the market, architecture obsolences, locking in a sub-optimal design, etc)

For what it's worth, my thoughts came after listening to this podcast (below). The title is: Apple's Memory Integrity Enforcement is a big deal. Be forewarned, it's hella geeky, but super insightful.

What do y'all think of Apple today?