Forty years to the day since the first Macintosh appeared beside a youthful Steve Jobs and made the world swoon when the word “hello” appeared on its display, Apple is closer than ever to realizing the vision it had then.
The Mac, with its GUI, mouse, and keyboard, launched an industry and spawned imitators (including Windows). The crowd at the Jobs-hosted launch event (which IDG/Computerworld founder Pat McGovern once told me included him, though he quipped he’d actually first seen the Macintosh the previous year) went crazy.
To the crazy ones
We know what happened after that.
The company saw insanely great success, deep failure, came close to collapse and then rose all over again. The human-centered principles inherent to the Mac user interface were later carried over to Apple’s mobile products such as its notebook, iPod, iPhone, iPad, and most recently, Vision Pro. The OS continued to improve, too, and that on-screen “Hello” became Siri — which is itself poised to rise again.
Siri seems to have fallen behind rivals, something Apple doesn’t want to see. So, the company is making significant investment in artificial intelligence (AI) to put Siri on steroids. Citing Pitchbook data, a Financial Times report claims Apple has acquired 21 AI start-ups since the beginning of 2017 and seemingly confirms that one of the main aims of the research is to put AI on mobile devices, most likely with iOS 18 later this year.
Apple joins the AI arms race
Wedbush Securities analyst Daniel Ives expects more news, predicting Apple will make a sizeable acquisition in the space this year. “There’s an AI arms race going on, and Apple is not going to be on the outside looking in,” he told FT.
Evidence Apple is accelerating its efforts continues to percolate. Morgan Stanley analyst Erik Woodring described Apple’s “Edge AI” opportunity in an extensive note to clients. “We believe Apple’s efforts to bring these features to market are accelerating, increasingly the likelihood of an “AI iPhone” launch as soon as Fall 2024,” he wrote.
“For example, the mix of Apple job postings for Deep Learning and Natural Language Processing roles continue to increase, with Deep Learning appearing in 49% of Apple’s AI jobs postings, and NLP appearing in 23% of postings.”
Hello to the smart machines
The scalp Apple seems to want is to create an on-device generative AI solution. Existing systems, even on mobile devices, rely on cloud-based AI, which drives up the cost per query. Apple seemingly hopes to exploit the Neural Engine on its powerful existing chips alongside its own innovations (including MLX and “Ferret”) to enable on-device intelligence at least partially.
This, of course, extends to Siri and suggests that in the not-so-distant future, when you turn on your Mac, you’ll be greeted by a spoken “hello” from a computer that through a combination of GenAI LLM models and machine vision intelligence has something like automated consciousness. (Ferret is interesting because it translates what it sees into words, which can then be used to drive LLM models — perhaps one day a little like Georgie La Forge’s glasses.)
The answer to what question?
Don’t expect much visibility into the company’s plans for the next few months. We are still at that stage in Hitchhiker’s Guide to the Galaxy during which pundits and philosophers appear on chat shows to disagree about what they think is coming. Deep as these thoughts may be, all eyes will be on the company at June’s WWDC, as that is where it is most likely to discuss these plans, if they are to come later this year.
In a client note today, Morgan Stanley’s Woodring once again stressed that once Apple does improve the AI in its devices it will experience a strong “Edge AI refresh cycle,” warning these plans will remain a “closely guarded secret” for some time to come —just like the Mac four decades ago, and potentially equally as life-changing.
Please follow me on Mastodon, or join me in the AppleHolic’s bar & grill and Apple Discussions groups on MeWe.
Copyright © 2024 IDG Communications, Inc.
This story originally appeared on Computerworld