It’s wild to me how Apple took every advantage in the world and basically lost the plot.

This is the most muted WWDC keynote reaction I’ve seen in years, but what caught my eye is their AI models on-device. This is great for the app ecosystem once this becomes commonplace (their text detection in images is otherworldly, as one example), but only to a point. The heavy lifting is all being sent off-device, and that’s not great for Apple.

At this point I think we can say the iPhone is pretty close to feature complete. I went from upgrading annually to “I’ll probably keep my 15 pro for a third year,” because there’s nothing super compelling that makes me eager to drop a couple of grand. Assuming this is the common case (always risky), what’s driving a hardware upgrade cycle? My thinking was that there’d be a big push for on-device AI inference. It makes sense for latency and privacy reasons, but also because these are basically supercomputers that spend 99% of their lives basically bored. Apple hasn’t invested in a huge infrastructure buildout themselves; it’s been an open secret for a while that they’re huge cloud customers; they’re beholden to third parties for cloud compute. If everything’s being pushed to central data centers, the phone itself becomes basically a thin client, albeit a very expensive one.

Combine that with Apple’s petulance around their anti-steering provisions, and I’ve little interest in doing more business with Apple for at least this cycle. It sounds ridiculous, but Android hasn’t been this interesting to me in over a decade.