AI: Apple on really 'Think Different' AI path. RTZ #386
...echoes of AOL in early internet days
As one digests Apple’s WWDC 2024 ‘Apple Intelligence’ announcements, it becomes very clear that Apple is going a very different way than its big tech peers. And perhaps borrowing a page from a ‘blast from the past’.
After almost two years into this OpenAI ‘ChatGPT’ driven AI Tech Wave, Apple is thinking VERY different on matters ‘AI’. Let’s look at what it’s NOT doing for now, on the AI front, to clarify what it IS doing.
No Big LLM AI of its own in the Cloud: For now it’s off-loading that flow on a case by case ‘Opt-in’ basis to OpenAI’s ChatGPT-o (Omni), with the door open to other LLM AI models from Google (Gemini), Anthropic (Claude), and others. And no cash trading hands for now. Just traffic distribution flow to OpenAI. PERHAPS a revenue split on ChatGPT Plus upgrades. So OpenAI gets to try and convert potentially millions of mainstream Apple users onto its services, bolstering its growing revenue profile.
No major AI capex in the multi-tens of billions: Likely ‘just’ in the billions to scale its ‘Private Cloud Compute’ with its own Apple Silicon server chips, and other expenses related to upgrade almost every Apple application and service across its half a dozen plus Apple operating systems across its ecosystem. And potentially get their existing customers to upgrade their Apple hardware for new devices with better AI hardware and software capabilities.
No just repeating ‘AI’ repeatedly: As mentioned earlier, they said ‘AI’ just once in its almost two hour keynote, with ‘Apple Intelligence’ mentioned over 60 times. Not parroting the fashionable acronym of the moment. But attempting to redefine the acronym from the core out.
No grandly announced plans to ‘Smart Agents’ with scifi ‘Her’ and/or ‘ScarJo’ like capabilities: Just practical, almost prosaic services across dozens of Apple applications and services. Doing practical, ‘Agentic’ things in small bites. For now, very much under-promising, and obviously trying to over-deliver.
No aspirational comments about ‘AGI’: Apple is eschewing debating ‘existential’ questions about when we get to ‘AGI’, aka Artificial General Intelligence’. Rather, they’re just focusing on building practical bottom up ‘Apple Intelligence’ services one at a time, almost all at once.
I could go on, but you get the idea.
And I’d like to underline one thing in my view that Apple IS CLEARLY doing. And it’s again thinking VERY different than what most might expect them to do at this AI Tech Wave. Especially as many believe still that Apple is coming from behind on AI. (Pro tip: They’re not, as I’ve been arguing for some time).
Apple is doing something very different here, and perhaps the market is slowly realizing its merits.
To be clear, the core takeaway for me on Apple post WWDC 2024, is that Apple is NOT integrating deeply with OpenAI, Google or others on AI to get their AI capabilities into Apple as fast as possible.
It’s very clear after WWDC 2024, that Apple is building its own deep down secure and private AI infrastructure both in cloud and devices, to provide Apple versions of what it think these Generative AI technologies can do from the devices out, not the cloud in.
It will take time, billions in investment, and a bottom up software and hardware AI redo effort ACROSS every nook and cranny of Apple’s ecosystem.
And slowly show clear benefits to mainstream users what this AI thing maybe all about. While assuring them that it’s as Safe as possible, with a laser-focus on user Privacy, building on their reservoir of Trust. At Scale.
So they will do the ABSOLUTE MINIMUM amount of integrating any external AI systems.
And make sure the broader LLM ‘AI world’ is ’Opt-in’ for their users only, every time they invoke the external AI chatbot.
In the meantime, Apple is aggressively focused on getting millions of developers on the Apple ecosystem to build bottom up AI services (the ‘Apple Intelligence’ kind) WITHIN the Apple ecosystem.
In my view, they have a better than even chance of pulling that off, given how mainstream users barely understand what ‘AI’ is about, and are generally fearful of it.
So it’s a diametrically opposite strategy than peers. Thinking Small, to get really Big.
For folks with longer memories of earlier tech waves, it’s kind of like how AOL showed tens of millions what the ‘Internet’ was in the 1990s, after Netscape wowed the world with their browser in 1995.
For over half a decade, most mainstream users thought AOL, (or America Online as it was also known), WAS the Internet.
Apple is likely to do the same this time over the next few years, only with over two billion plus mainstream users worldwide in its ecosystem, not just few tens of millions primarily in the US.
It’s definitely a very ‘Think Different’ strategy than what’s being executed by all the rest this AI Tech Wave. It may take a bit of time, but they’re likely to stay the course on their approach. Stay tuned.
(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here)
Do you think Apple is playing it slow and cautious because of its large and profitable user base? With a billion devices deployed and each user having significantly higher ARPU than Google, they can’t afford to poorly execute AI. See past incidents, e.g., Apple Maps, antenna-gate, battery-gate, etc.