If users prioritize big-data AI services, like those offered by Google, Amazon, Facebook, then Apple will be in a position similar to BlackBerry was in 2007
Before the iPhone, RIM's BlackBerry was the king of smartphones. They seemed unstoppable, because by most accounts … Tweets: @hunterwalk , @bryfitz , @craigmod , @benthompson , @benbajarin , @erickschonfeld , @mdudas and @marcoarment Tweets: Hunter Walk / @hunterwalk : “If they're right — and that's a big “if” — I'm worried for Apple.” http://marco.org/... by @marcoarment http://twitter.com/... Brian Fitzgerald / @bryfitz : BlackBerry was doomed when Apple changed what smartphones were for. What if big data AI is doing that to Apple now? http://marco.org/... Craig Mod / @craigmod : When the interface becomes invisible and data based, Apple dies: http://marco.org/... Ben Thompson / @benthompson : Yep. If anything the headline is too optimistic. http://marco.org/... Ben Bajarin / @benbajarin : Not so big of an IF in my opinion. Also, doesn't mean hardware is not relevant in this future. Also, Watson... http://twitter.com/... Erick Schonfeld / @erickschonfeld : “It's easier to conceal the development of an entire car than a major presence in AI and services.” https://marco.org/... Mike Dudas / @mdudas : “Where Apple suffers is big-data services and AI...Google — are far ahead of them, and the gap is only widening.” http://marco.org/... Marco Arment / @marcoarment : If Google is right, that's a BIG problem for Apple: http://marco.org/... Expand More For Next Unexpand More For Next
Context & Ripple Effects
Marco Arment's 2016 essay landed a year after Google's I/O keynote suggested it was closing the design gap with Apple, extending the worry from interface polish to the deeper question: whether smartphone value was shifting from the device to the big-data services behind it. Brian Fitzgerald's framing on Twitter made the analogy explicit — BlackBerry fell when Apple changed what smartphones were for, and big-data AI might be doing the same to Apple.
The subsequent record cuts both ways. Ben Thompson's [[a:923866|iPhone X analysis argued Apple's hardware-software integration fends off less integrated rivals even when their services are superior]], and eight years after the warning, Apple answered it directly with [[a:867095|Apple Intelligence, embedding text and image generation across iOS, iPadOS, and macOS as features rather than a standalone service]].
First-order effects
- Apple faces the exact fork Arment described: match Google, Amazon, and Facebook's data-hungry AI services or cede the AI layer to them — while its privacy stance, which Joan Westenberg argues positions it to set new AI data-usage standards, constrains the first path.
Second-order effects
- Apple's chosen answer — treating generative AI as an integrated OS feature per the WWDC demos — forces the comparison onto integration and trust rather than raw model scale, where Thompson's analysis suggests Apple's structure holds up against less integrated rivals like Google.
Third-order effects
- If Arment's premise holds across AI waves, platform power migrates from device makers to whoever controls the data and service layer; if the Apple Intelligence approach works, the industry splits into two durable models — big-data service platforms versus privacy-first integrated devices — instead of one winner.
The trend: Each AI wave reopens the same structural question Arment posed in 2016: whether value accrues to the device maker or to the big-data service providers behind it, with Apple's privacy-first integration now the counter-model being tested.