So. Breathlessly anticipated by everyone after months of careful leaking, the details of Apple’s AI offering were finally dropped at last week’s World Wide Developer Conference 2024. It was Apple’s come-from-behind entry into the AI race, called, unsurprisingly, Apple Intelligence. Within minutes of the release of their slick one-hour-and-43-minute video about it, metaphorical forests were cut down all over the world to make space for endless analysis and prognosis.
Your correspondent joins the fray.
But first, let’s get some things out of the way. A number of commentators have noted darkly that Apple is not really an innovator, more an improver of existing technologies. The Apple II was an improvement on the first general purpose PC (the Altair 8080), Apple’s graphical interface was copied from Xerox, the iPhone stood on the shoulders of Blackberry, and the Mac Air was simply an existing laptop with a sleek design.
This has often been referred to as the “second mouse” strategy – which is to seek out pioneering innovations in the market, copy them or buy them, make a few nifty tweaks here and there, add a bang-up industrial design, user experience and the best marketing money can buy before presenting it as the next and newest big thing.
When it comes to Apple, this description is neither fair nor accurate. Apple has indeed done some of the above, but they are as innovative a company as has ever existed. From the first iPod to the iPhone/App Store ecosystem to the wildly superior MacOS, they have been pioneers in all the ways which count. They have always understood what users want, how much they will pay, and how to deliver it in a perfect package.
But in the volcanic world of AI they have indeed been a laggard. Microsoft, Google, Meta and a host of newcomers have raced ahead, leaving Apple aficionados scratching their heads and wondering what happened.
The announcement on 10 June was their way of saying: “We are here.”
First, let’s look at some of the specific AI capabilities that they announced before we get to the real underlying story.
You can make custom emojis on your iPhone. Er, okay.
You can transcribe voice to text. Big deal, lots of apps do that.
You can edit photos on your device in real time. Not earth shattering.
You can summarise text. Yawn.
You can generate images from text. Yawn again.
You can use ChatGPT for free – they’ve done a deal with OpenAI. Ouch, that must hurt.
You can use an improved version of Siri. About time, it was awful.
You can design and keep 3-D pets on your iPhone. Ho hum.
You get the picture. Every individual AI capability announced last week (many more than this list above) falls into the “Oh, okay” category at best, or into the snore category at worst.
So what was the big news that caused Apple stock prices to spike by 10% last week? It is the fact that Apple is embedding AI capabilities deep into all of their devices, operating systems and applications, so deep that AI smarts are now being gene-spliced into the DNA of every Apple product, from the hardware layer upwards.
This is, by all techie accounts, an impressive achievement. Because the compute requirements of most kinds of AI (particularly generative AI) are famously brutal, Apple set out to build a fancy multilayered architecture to distribute compute requirements to where they are most appropriate.
On the devices themselves (iPhones, Macs, etc), devilishly clever new AI software will be deployed. These are like “mini” AI modules for specific tasks, which will be swapped in and out depending on the user’s needs, thereby using the relatively modest compute capabilities of the device efficiently. And, in the future, specialised on-device AI chips will find their way into new devices.
Read more in Daily Maverick: Apple Hits Record After Unveiling ‘AI for the Rest of Us’ Plan
For those heavyweight tasks needing computational heft, user interactions and requests will be pushed out of the device and across the internet to mega data centres for processing. And, to tackle those AI tasks for which they have no current solution (such as large language models), Apple has surrendered to partners like the aforementioned ChatGPT.
All very clever, but it has caused a bit of a kerfuffle. Apple has traditionally been very serious about user privacy. But, if they are sending our data and prompts and other AI interactions willy-nilly over the internet to their own (or even to their competitors’) server farms, how will they ensure our privacy? (An amusing security-related kerfuffletjie had Elon Musk threatening, somewhat seriously, to lock his employees’ iPhones in radio-proof Faraday cages while they are at work should Apple go ahead with the ChatGPT partnership.)
Read more in Daily Maverick: When AI takes all of the jobs, who will pay our taxes?
Apple has responded. They have promised that these requests will go through all manner of encryptions and a “Private iCloud” and anonymity shields, so nobody should be worried, their privacy will be protected. Yes, well. Perhaps.
But let’s look on the bright side here. Apple has a history of product genius and strategic nous. Their “AI everywhere” approach is ambitious and serious and has been carefully thought through by some of the smartest engineers, designers and developers in the world.
Even if you have never lain awake at night dreaming of an emoji-maker, I am confident that the exciting stuff is still to come. Underestimating Apple has never been a smart move.
Especially when they come from behind. DM
Steven Boykey Sidley is a professor of practice at JBS, University of Johannesburg. His new book It’s Mine: How the Crypto Industry is Redefining Ownership is published by Maverick451 in South Africa and the Legend Times Group in the UK/EU, available now.
