Hey Everyone,
At AI Supremacy on occasion, I ask some big questions about the future of AI.
I wanted to give you all a quick update on Apple’s status in Generative AI in early 2024.
Apple’s WWDC event is just weeks away, June 10-14, 2024. Apple is in a major crisis this year and will be counting on on-device Generative AI to boost its hardware and iPhone sales. I expect more AI updates and stunning new features in its WWDC event this year than usual.
It’s worth being excited about if you like me, are fascinating by AI devices and the future of what will be possible.
Apple’s on-device AI Needs to Boost Sentiment for iOS 18
Apple’s iPhone shipments in China have fallen again, government data show. Apple shipped roughly 2.4 million smartphones in February — a 33% drop from 2023, per Bloomberg. Apple’s stock is down 9% so far in 2024 and for Apple that’s a lot. Apple’s market cap only stands at $2.61 Billion now. I say only because it was the most valuable company in the world until relatively recently.
Apple will Likely Partner Heavily with Gemini and OpenAI
Apple is reportedly exploring a partnership with Google for Gemini-powered features and perhaps also OpenAI tech on iPhones. Apple GPT is the long-rumored AI chatbot from Apple that one day may arrive on all Apple devices. Meanwhile Apple’s overall revenue and earnings per share are projected to drop year-over-year for the first time since 2016, and it’s scrambling to improve Siri and add more on-device AI capabilities to its lineup.
Apple Needs to catch up with BigTech in AI
As Microsoft, Meta, Amazon and Google invest in Generative AI to a marked degree, Apple has a lot of catching up to do to integrate it into their devices, but a few papers they have released show some good results. Apple’s AI chief John Giannandrea is leading development on large language models within Apple.
Apple devices and features like Siri, Apple Maps, search, and even the mixed-reality headset Vision Pro have benefitted from the gradual introduction of machine learning and artificial intelligence. Apple needs to really impress at WWDC 24.
Apple is Overly reliant on iPhone Sales
Apple still makes well over 50% of its total revenue just from iPhones sales. The latest info shows the iPhone is Apple’s biggest moneymaker, generating $69.7 billion of revenue in the latest fiscal quarter (Q1 2024 ended Dec. 30), or 58% of the company’s total. Thus Apple needs to show major improvements in on-device Generative AI capabilities in its product features.
📶 From our sponsor: 📶
Perform Deep Market Research in Seconds 📊
Using up-to-date customer insights can contribute to a 175% increase in revenue, with higher quality leads and 72% faster lead conversion… However, market research projects can easily cost thousands of dollars or more and take weeks to complete.
Osum helps you perform deep market research in seconds. Just copy and paste a URL to instantly access detailed industry insights, SWOT analysis, buyer personas, sales prospect profiles, growth opportunities, and more for any product or business.
Stay ahead of your competition and discover new ways to unlock 10X growth.
Note: My readers gets 25% off on any plan for life. Use the code “MKS” at checkout.
Can iPhone Sales Rebound in China with iOS 18?
When Counterpoint Research, in its report released in early March, estimated that iPhone sales in China had dropped by 24% in the first six weeks of the year it’s clear Apple will become under pressure in China, where the U.S. is considering even tougher sanctions against Huawei. China pushing Huawei’s efforts also has led to a rejuvenation of its brand as a symbol of Chinese nationalism. Huawei’s net profit for 2023 rose by 144.5% to 87bn yuan, with the margin more than doubling on a year earlier to more than 12%.
Apple’s Great Pivot to Generative AI in Late 2023
Apple abandoning Project Titan its smart car EV effort also proves that Apple knows where it needs to improve, pivot and invest more of its significant cash pile. As of November, 2023 Apple had a staggering $162.1 billion in cash on hand. Apple CEO Tim Cook said his company “investing significantly” in artificial intelligence in late February but we’ll only find out more in June and Apple must be scrambling to catch up to BigTech in this crucial domain.
Recently Apple reveals ReALM
ReALM is Apple’s most interesting paper of the three I have reviewed in recent weeks. Along with MM1, relating to multimodal AI. Apple of course also wants its Generative AI to be on-device, that is, Apple wants AI to run directly on its hardware instead of in the cloud.
Tim Cook
“Later this year, I look forward to sharing with you the ways we will break new ground in generative AI, another technology we believe can redefine the future,” he said during the company’s annual shareholders meeting, which was held virtually.” – Tim Cook
If Apple poured in the area of $7 to $10 Billion into its Smart EV Car project in the past decade, it needs to do roughly the same effort for Generative AI upgrades, in roughly a three year span.
Apple has acquired dozens of AI startups in recent years, and embedded software in updated smartphones, computers, watches, apps, operating systems and Vision Pro VR headset. Now it needs to hire AI talent in Generative AI and focus more on product innovation in its next cycle.
What is ReALM?
Apple’s new AI models could let Siri remember your conversation history, understand what’s on your iPhone screen, and be aware of surrounding activities, such as recognizing the music playing in the background.
ReALM makes Siri a lot smarter about on-device intelligence.
Apple’s AI researchers this week published a research paper that may shed new light on Apple’s AI plans for Siri, maybe even in time for WWDC.
ReALM stands for “Reference Resolution As Language Modeling.”
Apple’s researchers said it outperforms GPT-4.
“We also benchmark against GPT-3.5 and GPT-4, with our smallest model achieving performance comparable to that of GPT-4, and our larger models substantially outperforming it.” – Apple paper.
Apple’s ability to make sense of on-screen data with Siri might truly feel like a substantial upgrade among other features.
Apple looks fairly confident in on-device LLM capabilities that are multi-modal and have reference resolution awareness. In early March, 2024 Apple referred to its new M3 MacBook Air as the ‘best consumer laptop for AI, raising questions about the company’s plans. In the past Apple would not use the term “AI”, but rather machine learning. Apple is setting a new tone regarding AI in 2024, as evidenced by its latest product releases, noted Fast Company.
This will make 2024’s edition of WWDC a lot more interesting than usual because Apple is under serious pressure to perform.
Apple Vision Pro’s recent upgrade of Spatial Personas is also fairly impressive. It’s clear that Apple has an opportunity to innovate on-device intelligence and new user experiences on a variety of its products powered by Generative AI and both LLMs and SLMs that are multi-modal.
As we have seen in recent years, its when companies are under pressure like Meta, Netflix, Google and now Apple that they actually begin to innovate and utilize Generative AI more significantly. Apple’s shareholders will be watching with increased scrutiny follow Apple’s China woes.
Apple is working on a large language model framework called “Ajax” as well as a chatbot known internally as “Apple GPT” but has a lot of work to do in a short-time to be ready for a great WWDC reveal.
Siri will be a Lot Smarter
As Amazon and Google rush to upgrade their personal voice assistants with Generative AI capabilities, I think Siri can make among the most radical upgrades.
Basically, the ReALM paper introduces Reference Resolution As Language Modeling (ReALM), a conversational AI system with a novel approach to improving reference resolution. The hope is that ReALM could improve Siri’s ability to understand context in a conversation, process onscreen content, and detect background activities.
Siri’s in-context learning and prediction will make it more attuned to the user and what the user is doing on the screen. Not just a ghost in the cloud, but a ghost in the machine and human-interface. And for mobile innovation, that might be a big deal for Apple and what we can expect from our mobile devices in the years to come.
Mobile Devices that are Self Aware?
To give you a more tangible idea of what makes the ReALM special for Apple: ReaLM can convert conversational, onscreen, and background processes into a text format that can then be processed by large language models (LLMs), leveraging their semantic understanding capabilities.
iPhones as the First Native-AI Devices
If this is true and they execute well, iPhones may become literally the first AI native devices, well before OpenAI is able to build such a device. Apple’s deep pockets and product quality excellence can captivate the world with a major upgrade in user experience and completely change our relationship with our devices.
This is the third paper I reviewed.
“Specialized Language Models with Cheap Inference from Limited Domain Data”
If the rumors about iOS 18 are true, Apple could finally take the current wave of AI development truly mainstream, as only Apple can, where around 1.5 Billion consumers use iPhones daily.
“As of 2023, there are an estimated 1.46 billion active iPhone users worldwide. This represents a market share of 21.67% of the world’s total smartphone-using population.”
Apple products and iPhones seem to be more popular with GenZ and the Alpha cohort (e.g. “teens”) Apple according to a Piper Sandler survey in 2023 I reviewed.
If iPhones can be seen as ubiquitous multi-modal AI devices with “in-context awareness”, we’ll be able to do things with them that were not possible even last year. It will make the experience feel fundamentally different.
Now nearing mid 2024, since small language models are becoming more efficient, on-device capabilities will begin to improve. There is surmounting pressure on Apple to catch up with competitors in the AI field such as OpenAI, Microsoft, Google and even Anthropic and yet for some product features it will likely partner with not just Google Gemini, but multiple of the above in my opinion.
Apple’s Multi-Pronged AI Strategy
It’s no secret that Apple is always acquiring AI startups, and it acquired DarwinAI in mid March, 2024. DarwinAI has developed AI technology for visually inspecting components during the manufacturing process and serves customers in a range of industries. But one of its core technologies is making artificial intelligence systems smaller and faster. Apple can do a mix of internal R&D, partnerships, acquisitions and heavily invest in applied product innovation and do relatively well in the Generative AI era, it will just take them more time since they have higher quality standards, they say a Microsoft or a Google. As we can clearly see today by how poorly Gemini or Copilot performs in general.
Apple Needs to Focus on High Performance Quality of its AI
Apple’s credibility and reputation is at stake here around quality. The iPhone maker needs to demonstrate innovation but those that are actually useful to its existing customers, among the Billions of consumers.
Historically we know that iPhone users are higher income earners, and also engage more on Apps than their Android counterparts. For example, the media iPhone app user earns $85,000 per year, which is 40% more than the median Android user who has an annual income of $61,000.
These more affluent customers have higher expectations for Apple. Along with progress in small language models in 2024, Apple is making headway in lowering the cost of language processing. Apple is also making progress in diffusion and animating photos and similar to what Google has done with VLOGGER. Apple’s machine learning research team also quietly releases a framework called MLX to build foundation models in December, 2023.
Apple traditionally a very secretive firm, is being noticeably more transparent about what it’s working on if you follow the industry news and read all of these papers. An educated analyst or consumer can have a fairly good guess of what to expect in iOS 18 a bit of which I’ve tried to outline in this article. Apple is speaking to the open-source innovation of the times in 2024.
🍎 High Hopes for iOS 18 as AI Features Shine
Siri will be able to do more in iOS 18 because it will understand more of what you are doing and asking. It will understand the user a lot better than in the past. As far as I see it, Apple’s iOS 18 update will likely be full of new AI features.
“But we’ve got some things that we are incredibly excited about that we’ll be talking about later this year,” Cook said in response to a question about generative AI on the company’s most recent earnings call.
Among the most anticipated June, 2024 revelations of Apple will be AI features around iWork. Apple has even ‘ramped up’ development of an AI-powered code completion tool that’s similar to Microsoft’s GitHub Copilot. Apple may even beat out Google for AI products that consumers actually want to use in 2024. While Google will have more products, Apple will have higher quality features. Google’s graveyard of poor product execution and failures is getting pretty significant at this point. Unfortunately this seems to be now even impacting Google Gemini related products.
Apple as the Default Leader for the Truly AI Native OS
Apple might still be the most logical pet to invent the first truly AI OS, or operating system. Around a quarter of folk in Western countries have exposure to iOS. However there’s no doubt Apple was slow out of the gate to work on Generative AI and traditionally is the least likely to hype what they do with machine learning. Apple could be a good come-from-behind story in Generative AI, since the reputations of Google, OpenAI and even Microsoft’s Copilots seems to be in relative decline as of April, 2024.
Apple has a history of being a fast follower rather than a first mover when it comes to major technology shifts. This time will be no different, Apple’s playbook is well-established and it’s focus on product-execution remains a global leader for a reason. WWDC is anticipated to kick off with a keynote speech — which traditionally unveils the new operating system for the iPhone, Mac, iPad, Apple Watch and TV. I’m curious to see what role Generative AI and iOS as an AI operating system features this year.
That Siri will be able to understand elements of what is on our screen might enable better personalization of products, features and even Ads. Apple is surprisingly good at Ads, when it decides to do them. Apple is expected to make around $10.34 billion this year on Ads and the growth is impressive. Apple’s ad revenues are growing about twice as fast as Google’s, at 21.2% compared to 10.3%.
Apple’s perfectionism means new kinds of efficiencies for LLMs and SLMs. New kinds of on-device multimodal performance that intersects with their various products and features. Their MM1 paper demonstrated a new method for training multimodal models using synthetic data including images and text. The researchers behind MM1 claim their new method speeds up performance and reduces the number of follow up prompts to get a desired result. This might bode well for AppleGPT.
Is Apple Bullish on Generative AI?
Apple sees “incredible breakthrough potential” for generative AI, Cook said, adding: “We believe it will unlock transformative opportunities for our users when it comes to productivity, problem solving and more.”
Apple’s “Spatial Personas” is very Minority Report-like I will note for the Apple Vision Pro. What might the future of Ads look like in such a feature I have to wonder?
Minority Report (2002) Personalized Ads of the Future
I haven’t seen this video in a few years, but Spatial Personas very much so!
One of the most important Generative AI trends of 2024 and 2025 will actually be how mobile devices improve their product with AI. In early February we learned that Huawei itself will focus on increasing the manufacturing of its AI chip, the Ascend 910B, at the expense of production of its Mate 60 phones in at least one facility. China scrambling for semiconductor independence might slow down how they adopt Generative AI into their devices thus giving Apple a temporary edge.
Subscribe to Semiconductor Things, for coverage of the semiconductor industry.
Apple will have its own Copilot, it just won’t charge an arm and a leg for each one like Microsoft has chosen to do. Apple has looked at potential uses for generative AI in consumer-facing products, like automatic playlist creation in Apple Music, slideshows in Keynote, or AI chatbot-like search features for Spotlight search.
Automatic playlist creation
Slideshows in Keynote
AI chatbot-like Search features for Spotlight Search
Ironically it may implement some of these with the help of Google Gemini, OpenAI, or both those companies.
iOS 18 will likely debut at the WWDC 2024 event, which we now know will be held between June 10 and 14.
Apple still makes the bulk of its revenue via hardware and iPhone sales. But to move to services and advertising it needs better AI. In the first quarter of Apple’s 2024 fiscal year, Apple services revenue accounted for just over 19 percent of the company’s total revenue. Apple needs to be a leader in Generative AI due to its vulnerability to iPhone sales. The iPhone will be 20 years old this June, it’s hard to believe. It’s also hard to believe how much Apple’s entire empire is dependent on it.
Apple’s Antitrust Woes Could Spiral
Apple’s increasing antitrust battles will also be very costly. The Department of Justice antitrust lawsuit against Apple could have major consequences for the iPhone maker’s business segments. Last month, in March, 2024, The European Commission, the European Union’s executive arm, hit Apple with a 1.8 billion euro ($1.95 billion) antitrust fine. The Commission said Apple abused its dominant position on the market for the distribution of music streaming apps.
The commission said it found that Apple had applied restrictions on app developers that prevented them from informing iOS users about alternative and cheaper music subscription services available outside of the app. Apple is a monopoly broker in multiple ways so these fines could continue.
Apple has been so dominant globally in the wearables category and the Apple Watch might also be infused with Generative AI capabilities and a smarter Siri might move the needle. However, even such a stunning success, Apple Watch revenues are pegged close to $20 billion, perhaps as high as $25 billion for 2024.
Can Apple Vision Pro Scale?
Apple reportedly has sold over 200,000 of its pricey Vision Pro AR headset in the first 10 days of pre-orders. Analysts are forecasting anywhere from 200,000 up to one million unit sales in 2024 for the $3,500 and up headset. So even a conservative estimate means they only really make under $2 Billion in sales in 2024 off of this incredible new and pricey device.
The Vision Pro might be even more dependent on flashy AI features to attract consumers. Apple needs to badly diversify more into Advertising, Apple’s services business, new product lines like its Spatial Computing and finally transition from hardware more into software and the Cloud. One thesis is a good execution of Generative AI products enables them to do this better.
Leverage AI for Better Revenue Diversification
Apple doesn’t need to have the best AI talent or do the best AI Research, it just needs to execute the bleeding edge of Generative AI into its products better than anyone else. Given the company’s history there’s a fair to good chance it will be able to accomplish this. It needs to integrate Generative AI into its iPhone and devices better than a Samsung, Huawei or even a Google for that matter.
Apple badly needs to augment its Advertising, iCloud, Music, TV+, Pay and Vision Pro device with Generative AI to finally provide the business with a greater variety of high-margin revenue sources.
Apple Advertising
Vision Pro (Spatial computing)
iCloud
Music
TV+
Apple Pay
Finally sometimes we forget how much money Apple really makes and how difficult it is to diversify where the revenue is coming from when you have reached that level of global domination in technology.
Apple revenue for the twelve months ending December 31, 2023 was $385.706B. The main use-case of Generative AI is product-enhancement for BigTech and making things like Advertising growth more efficient via new kinds of immersive Ads. Apple is well on its way to enhance existing product categories and innovate new use cases for a smarter iOS 18.
If you enjoyed this post subscribe to support the author or share this article.
Read More in AI Supremacy
