Apple leaps into AI with an array of upcoming iPhone features and a ChatGPT deal to smarten up Siri
- Thread starter bnew
- Start date
TECH
A marathon, not a sprint: Apple’s AI push faces big challenges in China
PUBLISHED THU, JUN 20 20246:38 AM EDTUPDATED THU, JUN 20 20247:08 AM EDTArjun Kharpal @ARJUNKHARPAL
SHAREShare Article via FacebookShare Article via TwitterShare Article via LinkedInShare Article via Email
KEY POINTS
- Apple Intelligence is Apple’s big play that aims to bring artificial intelligence across its devices.
- There is little talk on how the Cupertino giant will bring this tech to China. This is likely to do with China’s stringent rules on AI, analysts told CNBC, as Apple tries to figure out how to approach the complex market.
- Questions remain over which Chinese tech giant Apple will partner with for Siri and how Private Cloud Compute would work in China.
In this article
Follow your favorite stocksCREATE FREE ACCOUNT
The Apple Siri AI icon is being displayed on a smartphone, with Apple Intelligence in the background.
Jonathan Raa | Nurphoto | Getty Images
Apple’s big artificial intelligence push faces some big challenges in China — one of the iPhone maker’s most critical markets — as Beijing maintains strict rules around the buzzy technology.
The uncertain path in China comes at a time when Apple’s market share is being eroded in the world’s second largest economy by a resurgent Huawei and other local smartphones players, which are talking up their AI features.
Apple Intelligence is the Cupertino giant’s play that aims to bring AI across its devices. It features an improved version of Apple’s voice assistant Siri, as well as features that automatically organize your email or transcribe and summarize audio footage.
Apple said that Apple Intelligence will roll out in U.S. English this fall, with additional languages, features and platforms due to arrive over the course of next year. The company was, however, quiet on the product offering in China during the AI launch at its annual developers conference this month.
That’s likely to do with China’s stringent rules on AI, analysts told CNBC, as Apple tries to figure out how to approach the complex market.
“China is in another world when it comes to AI given the regulatory environment there, so China is a big asterisk on Apple’s big announcements last week,” Bryan Ma, vice president of devices research at IDC, told CNBC via email.
Beijing has enacted various regulations over the past few years focused on areas ranging from data protection to large language models — the massive sets of data that underpin applications like ChatGPT.
China’s AI market is heavily regulated. Some of the rules include requirements for LLM providers to get approval for the commercial use of their models. Generative AI providers are also responsible for taking down “illegal” content.
Navigating these rules will be tricky for Apple.
Firstly, some of the features of Apple Intelligence are based on Apple’s own language model, which runs on both the phone and on the company’s own servers.
Under Chinese rules, Apple would likely need to get its AI model approved by authorities.
Secondly, one of the biggest announcements this month was that Apple’s voice assistant Siri can tap into OpenAI’s ChatGPT for certain requests — but ChatGPT is banned in China, meaning Apple would have to find an equivalent domestic partner.
Baidu and Alibaba are among China’s technology giants that have their own LLMs and voice assistants, ranking them as companies with which Apple can potentially partner.
Meanwhile, China’s internet is heavily censored with regulators concerned about the potential for AI services to generate content, which may go against Beijing’s views or ideology.
The likelihood is that Apple will have to build an on-device AI model and a cloud-based AI model that complies with local regulations, Canalys analyst Nicole Peng told CNBC over email.
The other part of the equation on AI for Apple to be successful in China, according to CCS Insight Chief Analyst Ben Wood, is for the company to create a localized AI experience on its devices that appeals to Chinese users.
“Localising the Apple Intelligence experience will be a major challenge for Apple,” Wood told CNBC. “As with all technology deployments, there are nuances to the way the service is delivered to respect the specific customs, regulations and use cases in a particular country.”
A key part of Apple’s pitch during the AI launch was its focus on privacy. The company announced Private Cloud Compute, whereby AI is processed on servers owned by Apple. Apple said that data processed is not stored.
Whether the tech titan will be able to fully own its own servers is another question. Chinese iCloud data is stored inside servers located in China which are run by a third party.
This could mean Apple might require a similar partnership for its AI computing servers, opening the tech giant up to critcisms about how private the data actually is.
WATCH NOW
VIDEO04:55
Investors started to do the math on Apple’s AI news: Deepwater’s Munster on delayed stock boost
“Maintaining complete user privacy in an AI era in heavily regulated markets such as China will be the biggest test for Apple yet,” Neil Shah, partner at Counterpoint Research, told CNBC. “Its going to be challenging for Apple to have fully controlled own private compute servers in China.”
CCS Insights’ Wood said Apple’s focus on privacy could help introduce AI features to the market. China passed a major data protection law in 2021, which looks to limit how information is collected and stored.
“Apple’s on-going focus on privacy and security practices may help placate local regulators and Apple has not been afraid to make concessions when required,” Wood said.
CNBC has contacted Apple over Private Cloud Compute and the company’s AI ambitions in China. A spokesperson did not directly address those questions, but pointed CNBC to an interview in the Fast Company business magazine with Craig Federighi, Apple’s senior vice president of software engineering.
Federighi expressed the desire to bring Apple Intelligence to China.
“We certainly want to find a way to bring all of our best product capabilities to all of our customers,” he said in the Fast Company interview, adding that “in some regions of the world, there are regulations that need to be worked through.”
The Apple executive said the process was under way to introduce the AI products to China, but gave no timeline.
Smartphone makers globally are talking up their AI features as a way to sell high-end phones to consumers who want to hold onto their device for longer.
Apple has been facing a number of challenges in China, where its market share fell to 15% in the first quarter of 2024, versus 20% in the same period the year before, according to Canalys data. Huawei, whose smartphone business was crippled by U.S. sanctions, revived once more and is now the biggest smartphone player in China, where it competes with Apple with phones targeting the premium segment.
Apple’s lag behind domestic rivals in launching AI features in China is unlikely to be detrimental to iPhone sales.
“For Apple, deploying China-grade Apple Intelligence is going to be a marathon and not a sprint. It will be deployed in phases over the years until Apple is confident and until then it will have to face some competition,” Counterpoint Research’s Shah said.
Wood said Apple’s control of its hardware and software integration will allow it to deliver a different experience from that of its rivals.
“Apple has an uncanny ability to explain its services and features better than rivals, even if it is essentially delivering the same experience or a subset of what rivals can offer,” Wood said.
“Despite the current focus on AI by rival China-based smartphone makers, Apple should still be in a strong position.”
1/1
It seems nobody understands how all this works. Apple will offer a variety of AI models like ChatGPT, Anthropic Claude, Google Gemini, Meta AI, Perplexity etc. to answer "world knowledge" questions which Apple's own secure models cannot. If you don't trust any of these companies you can simply choose not to opt-in to using them. But if you do opt-in, your IP address will be obscured, no personal information is shared, and your query is not allowed to be stored.
I would personally steer away from Meta AI and use Claude or Gemini instead, but it is entirely in your hands.
Also, all of your personal context related queries and commands will use Apple Intelligence using on-device or Private Cloud Compute, along with any image-generation which is on-device.
To post tweets in this format, more info here: https://www.thecoli.com/threads/tips-and-tricks-for-posting-the-coli-megathread.984734/post-52211196
It seems nobody understands how all this works. Apple will offer a variety of AI models like ChatGPT, Anthropic Claude, Google Gemini, Meta AI, Perplexity etc. to answer "world knowledge" questions which Apple's own secure models cannot. If you don't trust any of these companies you can simply choose not to opt-in to using them. But if you do opt-in, your IP address will be obscured, no personal information is shared, and your query is not allowed to be stored.
I would personally steer away from Meta AI and use Claude or Gemini instead, but it is entirely in your hands.
Also, all of your personal context related queries and commands will use Apple Intelligence using on-device or Private Cloud Compute, along with any image-generation which is on-device.
To post tweets in this format, more info here: https://www.thecoli.com/threads/tips-and-tricks-for-posting-the-coli-megathread.984734/post-52211196
1/11
Actually, really liked the Apple Intelligence announcement. It must be a very exciting time at Apple as they layer AI on top of the entire OS. A few of the major themes.
Step 1 Multimodal I/O. Enable text/audio/image/video capability, both read and write. These are the native human APIs, so to speak.
Step 2 Agentic. Allow all parts of the OS and apps to inter-operate via "function calling"; kernel process LLM that can schedule and coordinate work across them given user queries.
Step 3 Frictionless. Fully integrate these features in a highly frictionless, fast, "always on", and contextual way. No going around copy pasting information, prompt engineering, or etc. Adapt the UI accordingly.
Step 4 Initiative. Don't perform a task given a prompt, anticipate the prompt, suggest, initiate.
Step 5 Delegation hierarchy. Move as much intelligence as you can on device (Apple Silicon very helpful and well-suited), but allow optional dispatch of work to cloud.
Step 6 Modularity. Allow the OS to access and support an entire and growing ecosystem of LLMs (e.g. ChatGPT announcement).
Step 7 Privacy. <3
We're quickly heading into a world where you can open up your phone and just say stuff. It talks back and it knows you. And it just works. Super exciting and as a user, quite looking forward to it.
2/11
Never believe product demos. Google Assistant was supposed to be managing all my appointments 6 years ago.
3/11
100% agree, "the proof is in the pudding". It has to actually work. I will say that I think the technology exists today to actually make it work at the needed threshold. Actually making it work is still difficult. But 6 years ago I would have said the technology does not exist.
4/11
So can we never turn it off?
5/11
Any thoughts on the new Siri summoning ChatGPT? I understand the importance and opportunity for modularity but I feel that this particular AI experience is unnecessary. Why do I need one assistance to access another assistance and add friction? I really don't get it. Maybe it's just me but curious what others think.
6/11
i’m very very curious on how Apple did context management and tool use over ~unbounded set of app intents!
if you come across anyone willing to share please send them our way, i’d love to get the engineering stories out of them
7/11
AirPods are a key conduit of LLMs & massively underestimated
Great products.
Great user- experience (physical/ digital/ handoff between devices)
Lots of companies talking about ‘owning the ears
’, but few do it as well as Apple
Her.
If AirPods were separated out as a business unit, revenue would be greater than Airbnb (2022)
Margins are likely more extreme
8/11
@readwise save thread
9/11
There is a race going on at Apple between how fast they can improve the hardware for local inference, and how secure they can make the idea of private cloud inference. The latter seems really hard to implement.
10/11
@daddyyjuul @iqanazmah Apple finally reigniting some interest after years of silence. The next iPhone OS is going to be
11/11
Calculator coming to iPad
but with this new feature called Math notes!!!! How cool..
To post tweets in this format, more info here: https://www.thecoli.com/threads/tips-and-tricks-for-posting-the-coli-megathread.984734/post-52211196
Actually, really liked the Apple Intelligence announcement. It must be a very exciting time at Apple as they layer AI on top of the entire OS. A few of the major themes.
Step 1 Multimodal I/O. Enable text/audio/image/video capability, both read and write. These are the native human APIs, so to speak.
Step 2 Agentic. Allow all parts of the OS and apps to inter-operate via "function calling"; kernel process LLM that can schedule and coordinate work across them given user queries.
Step 3 Frictionless. Fully integrate these features in a highly frictionless, fast, "always on", and contextual way. No going around copy pasting information, prompt engineering, or etc. Adapt the UI accordingly.
Step 4 Initiative. Don't perform a task given a prompt, anticipate the prompt, suggest, initiate.
Step 5 Delegation hierarchy. Move as much intelligence as you can on device (Apple Silicon very helpful and well-suited), but allow optional dispatch of work to cloud.
Step 6 Modularity. Allow the OS to access and support an entire and growing ecosystem of LLMs (e.g. ChatGPT announcement).
Step 7 Privacy. <3
We're quickly heading into a world where you can open up your phone and just say stuff. It talks back and it knows you. And it just works. Super exciting and as a user, quite looking forward to it.
2/11
Never believe product demos. Google Assistant was supposed to be managing all my appointments 6 years ago.
3/11
100% agree, "the proof is in the pudding". It has to actually work. I will say that I think the technology exists today to actually make it work at the needed threshold. Actually making it work is still difficult. But 6 years ago I would have said the technology does not exist.
4/11
So can we never turn it off?
5/11
Any thoughts on the new Siri summoning ChatGPT? I understand the importance and opportunity for modularity but I feel that this particular AI experience is unnecessary. Why do I need one assistance to access another assistance and add friction? I really don't get it. Maybe it's just me but curious what others think.
6/11
i’m very very curious on how Apple did context management and tool use over ~unbounded set of app intents!
if you come across anyone willing to share please send them our way, i’d love to get the engineering stories out of them
7/11
AirPods are a key conduit of LLMs & massively underestimated
Great products.
Great user- experience (physical/ digital/ handoff between devices)
Lots of companies talking about ‘owning the ears
’, but few do it as well as Apple Her.
If AirPods were separated out as a business unit, revenue would be greater than Airbnb (2022)
Margins are likely more extreme
8/11
@readwise save thread
9/11
There is a race going on at Apple between how fast they can improve the hardware for local inference, and how secure they can make the idea of private cloud inference. The latter seems really hard to implement.
10/11
@daddyyjuul @iqanazmah Apple finally reigniting some interest after years of silence. The next iPhone OS is going to be
11/11
Calculator coming to iPad
but with this new feature called Math notes!!!! How cool..To post tweets in this format, more info here: https://www.thecoli.com/threads/tips-and-tricks-for-posting-the-coli-megathread.984734/post-52211196
Apple Vision Pro U.S. Sales Are All But Dead, Market Analysts Say
Anyone in the U.S. who wanted a $3,500 Vision Pro may already have one. The only hope for increased sales is a rumored cheaper model in 2025.
gizmodo.com
Apple Vision Pro U.S. Sales Are All But Dead, Market Analysts Say
Anyone in the U.S. who wanted a $3,500 Vision Pro may already have one. The only hope for increased sales is a rumored cheaper model in 2025.By Kyle Barr Published July 11, 2024 | Comments (25)
𝕏
Those still holding on to their Apple Vision Pros may remain in a rather exclusive club throughout this year. Market research shows that sales for Apple’s first big, expensive headset will remain low in 2024. The latest reports from those keeping tabs on the Cupertino, California company say AVP will have dropped off 75% by the end of August. The true test for Apple’s spatial dreams may rest on the rumored (slightly) cheaper headset.
The market analyst firm IDC told Bloomberg the Apple Vision Pro has yet to sell 100,000 units. It’s an expensive headset, and Apple wasn’t expecting it to sell like an iPhone. Supply chain analysts have reported that Apple cut its sales expectations for its $3,500 “spatial computer”in April. But this latest report shows that sales will have dropped off a cliff in the U.S. in the third quarter of this year and will continue to slacken through the holidays.
Last month, Apple launched the Vision Pro in international markets, including Europe, the U.K., China, Japan, and Singapore. IDC expects the AVP’s interest in those markets to keep the headset’s sales afloat until next year. The real pick-me-up for Apple’s spatial dreams would be a new, less expensive headset. Those in the know have hinted Apple is working on a “budget” Vision device slated for the latter half of 2025.
Even if the next Vision device costs half the Pro model, it will still cost $1,750 and one of the most expensive consumer-end VR/AR headsets you can buy. Rumors hint that the next device could remove the pointless exterior display to save on manufacturing costs. It could also reduce the FOV and use a less-capable chip than the current M2. Bloomberg hinted that Apple was even considering tethering it to an iPhone or Mac for daily use, which would drastically reduce its portability.
We don’t have pure statistics on how many folks returned their Vision Pro after buying it during the initial hype rush, but analysts have noted that many who bought one were confused by its more complicated setup and what they were supposed to use it for in their daily lives.
Sales expectations will put even more pressure on Apple engineers to design something that can compete with devices like the $500 Meta Quest 3 while justifying the higher price tag. Fans of the Cupertino company are already used to paying an “Apple tax” on their products, but not when the cost is literally thousands of dollars more.
Apple is working on a visionOS update to improve the faux-3D spatial photos, add a few new gesture controls, and allow for a panoramic Mac screen mirroring. The latest version of visionOS won’t have a public beta, so we’ll have to wait and see if the changes will give the few on-the-fence customers a reason to pick up the ultra-expensive headset.
