A People's AI
Over the last year we’ve seen Microsoft, Google, and now Apple demonstrate what they are doing with AI. As a somewhat casual observer, it seems to me all of these announcements fall in what I will call a “me too” category. Microsoft added a ChatGPT bot to Bing so Google had to add Gemini chat to Search. Microsoft added AI features to Office and Google did the same to Gmail and Docs. The problem with this matching of functionality is that both companies are focused on each other rather than their customers.
The main challenges to broad use of the latest AI tools are how easy they are to use and explaining what specifically they do and why a person should want to use them. So far the most publicized uses are enhanced search, writing, and creating pictures. In a way not unlike how the handwriting recognition of the Apple Newton Messagepad was treated, there is more attention put on how these AI tools are doing things wrong and often times with hilarious results. Personally, my use of GPTs has been as an enhanced search, and I prefer to do my own writing.
How Microsoft and Google have handled AI as response to each other is not surprising because honestly that is how these companies operate as they are competing against each other for corporate customers. Microsoft and Google both have retail consumer products but these are layers on top of a corporate DNA, and frankly it shows. Apple’s DNA has consumers woven from top to bottom and we see and feel this in their actual products and how they talk about them, and frankly that makes them more relatable and trustworthy to consumers.
While I find it hokey, the “branding” of these features as “Apple Intelligence” is consistent with Apple’s opinionated approach to all technology. Apple’s message is clear, they are not going to use AI in the same way as others, so much so that they call it something similar to but not the same was what others call it. Apple is saying AI is not Artificial Intelligence, it’s Apple Intelligence.
Of course, Apple is doing the same thing as Microsoft and Google in adding AI features to their first party products, but there is a key differentiation between Apple and these others, Apple is a hardware company and AI might be the biggest gift to Apple’s long term prospects since the mobile phone. The biggest question at Microsoft and Google right now is, how can they make money from AI, and it turns out that for them the AI revenue flow looks the same as advertising and search, but that is not how AI generates revenue.
The hottest company on the planet right now is Nvidia, who manufacture the processing chips (GPUs) AI needs to quickly provide results. Ever since Apple announced they put their M1 processors in iPads and similar high end processors in the iPhone, there have been questions about why Apple is putting all this processing horse power in these products when it will never be fully used. The answer to that question should be obvious now, that processing power is needed to run these new AI tools with performance similar to Nvidia. People expect Apple to keep their data on their iPhones, iPads, and Macs and their processors allow them to do that.
I own the Macbook Air that has Apple’s first M-series processor, and its going to be three years old in November. The Macbook Air M1 does everything I normally do, web browsing, checking email, some writing, crunching numbers, so fast I can’t imagine when I will need to replace it for these purposes. Last week I installed Ollama on this Macbook Air so that I can run AI models like llama3 locally and for the first time since I’ve owned it the Macbook was slow! Before this I imagined the reason why I would replace this Macbook Air is that Apple no longer supported it in the latest release of OS X, which could be ten years from now, but now I know that if I want to run future models locally I will want a newer Macbook.
Cloud-based AI has it’s place in providing widely available information in a comprehensible manner, but when it comes to actually doing things for people, that AI processing will need to be on the edge and near the people. The people behind the Rabbit R1 and the Humane AI Pin recognize this, but both are physically small and thus constrained. The perfect AI edge device is the smartphone, and so AI fits “glove-in-hand” with Apple’s core consumer product, while neither Microsoft or Google have such product. For Apple AI is not the “thing” and it will never be the “thing,” their “thing” is Apple silicon put in iPhones, iPads, and Macs, and AI makes the best case for the need of that “thing” than anything that has come before.