Some well-known Apple bloggers conduct yearly evaluations of the hardware, software, services, and other aspects of Apple, and grade Apple in these categories. While they mention AI in some of their other blog posts about Apple, I decided to see if AI was a key consideration in their evaluations of the company.
Evaluations:
- Neil Cybart's grades: 2017, 2020, 2021.
- John Gruber's report cards: 2018, 2019, 2020, 2021, 2022, 2023.
- Jason Snell's Six Colors report cards: 2015, 2016, 2017, 2018, 2019, 2020, 2021, 2022, 2023.
- They involve surveys sent to numerous Apple watchers.
- Michael Tsai's commentary on Snell's report cards: 2016, 2017, 2018, 2019, 2020, 2021, 2022.
My assumption is that if these bloggers and (in Snell's case) the broader Apple community prioritizes AI development, then that will be a significant part of many evaluations. If I were to write an end-of-year evaluation of Intel, NVIDIA, or AMD, then a large chunk of it would be a discussion about their progress (or lack thereof) in AI over the past year. For example, my hypothetical review of NVIDIA's 2020 would compare the DGX A100 accelerator#A100_accelerator_and_DGX_A100) with its predecessor V100. The A100 delivered only a small increase in "traditional" single- and double-precision floating-point performance over the V100, but in contrast, its AI-specific integer and floating-point capabilities broadened and increased considerably.
Full disclosure: I didn't read most of the evaluations, but I searched each link for basic keywords such as "artificial," "[neural] network," "ML," etc. If AI was a significant part of an evaluation then one or more of those words would presumably appear, with appropriate context, in the text search.
Explicit mentions of AI in my keyword search, 2015–2022:
Neil Cybart, "Apple Questions for 2017":
The company is willing to be a bit less secretive in order to get better access to newer technologies. The $1B investment in Didi and allowing AI researchers to publish are two examples.
Neil Cybart, "Grading Apple's 2017":
Apple is relying on machine learning to provide Apple Watch wearers personalized information based on their daily routine.
Neil Cybart, "Apple Questions in 2019":
Last year, Apple expanded the team by one with John Giannandrea being promoted to SVP of Machine Learning and AI Strategy.
However, if you actually go and read the articles, there is not much elaboration about AI and its role that it plays—or should play—within Apple. Many evaluations discuss Siri but don't really go into the AI aspects, which are the most important part of a digital assistant.
Explicit mentions of AI in my keyword search, 2023:
Jason Snell, "Apple in 2023: The Six Colors report card":
Adam Engst:
If anything, Apple’s next task is going to be to figure out what new general capabilities (perhaps AI-driven?) the Mac can be given that would give most people reason to look beyond the low-end chips.
Leo Laporte:
Apple Silicon continues to lead the pack and Apple’s forethought on building in ML co-processing into ALL its devices puts them way ahead in the race for on-device AI. Microsoft, Qualcomm, and Intel are just now getting to the starting line.
Shelly Brisbin:
Speech Access leverages machine learning in an incredibly useful and very Apple way. Apple may not openly call Personal Voice an AI feature, but it is, and it’s legitimately a good thing, and a technologically interesting thing for Apple to do.
Andy Ihnatko:
And the Pixel’s AI features are the kind of useful ‘Why doesn’t MY phone do that?’ magic that used to be associated with the iPhone.
Federico Viticci:
As we close the book on 2023 and look forward to 2024, a big question looms over Apple: are they really ‘late’ to AI, or were they waiting to release a more useful application of AI with LLM-powered iOS features that are going to benefit hundreds of millions of people in ways that go beyond texting with a chatbot? I want to believe that Apple’s rumored rethinking of iOS 18 around artificial intelligence will see Shortcuts take a prominent role and allow users to control their devices and apps in ways that aren’t possible today. iOS updates have been pretty iterative and unsurprising for the past few years, and iOS 17 was no exception. That’s not a bad thing (interactive widgets are great!), but I’m ready for something new and different.
There appears to be more discussion on AI for 2023 than for all previous years back to 2015 combined. However, this commentary still seems to be reactive rather than a proactive "skate to where the puck goes." Even in 2022, we saw AI art becoming mainstream, quickly followed by ChatGPT exploding on the scene. Subsequently, some rumors last year pointed to iOS 18 being a major AI-focused update this year, which Viticci referred to in his concluding remarks.
But I think Viticci's conclusion (minus the rumor reference) really belonged in the 2022 Six Colors report card—or even earlier. That "big question" about AI and Apple was looming in my mind since 2017–2018. That timeframe was when NVIDIA announced the Volta GPU with Tensor Cores and Google announced Night Sight for the camera and the conversational AI Duplex.
It's likely that Apple, internally, shares many views and blind spots with the Apple commentariat. Note that the causality is mainly in the other direction—big fans of Apple are such because they like and agree with most of Apple's decisions—but we can still predict Apple's views from the diehard fanbase.
I kinda do the opposite.
I watch what I like, or at least what I want to see. However, I try to set my rating scale so that the mean is a 3/4 and the standard deviation is 0.5/4. (I use a 4 point scale.) So any rating of 4/4 or below a 2.5/4 should be very rare. Recently I've watched at least three movies that I'd give below a 2.5/4, which I think is an indicator that I should be more discerning….
Do you watch movies you know will be bad or average just to keep your ratings pyramid normally distributed? Do I have autism?
Letterboxd