Per reporter: Cai Ding Per editor: Gao Han
On Monday (June 5), Eastern time, the much-anticipated Apple (AAPL, stock price of $179.21, market value of $2.82 trillion) global developers conference opened, the theme of this conference is "Code new worlds".
During the day's keynote, Apple made a big presentation about the new notebook Macbook Pro, iOS 17, and the highly anticipated Vision Pro mixed reality headset. However, in the two-hour speech, only the hottest artificial intelligence (AI) was missing, and Apple used more words like "machine learning" at the conference.
This can't help but make investors wonder: Under the wave of ChatGPT, is Apple still silent? What is Apple's attitude towards AI?
Recruitment positions reveal mystery
At the end of April ~ early May this year, the US stock technology giants have announced the first quarter of this year's earnings, after experiencing big layoffs in 2022 and early 2023, Google (GOOGL, stock price $127.31, market value $756.3 billion), Microsoft (MSFT, stock price $333.68, market value $2.48 trillion), Amazon (AMZN, stock price $126.61, market value $1.3 trillion), Meta (META, With a stock price of $271.12 and a market capitalization of $694.8 billion), giants have told Wall Street a beautiful new story: all in AI while reducing costs and increasing efficiency.
However, in a clear difference from the above four giants, during Apple's earnings call, Cook only took a few minutes to answer a question about AI. "It's clear that we've made huge strides in integrating AI and machine learning across the ecosystem, and we've built it into our products and capabilities over the years." Cook said.
In the opening show of this developer conference, the keynote speech, Apple did not say a word about AI. Could it be that under the wave of AI, Apple really intends to "stand by and watch"?
Image source: Apple's official website
The "Daily Economic News" reporter noticed that in fact, at this developer conference, AI was hidden in some hardware and software updates. When introducing the latest iOS 17 upgrade, Craig Federighi, Apple's senior vice president of software engineering, announced the use of the Transformer language model for input and speech recognition text.
Federighi said that Apple Keyboard now uses a self-developed Transformer language model with the ability to automatically predict word suggestions and corrections. It should be noted that Transformer is also the underlying technology used by many generative AIs, including ChatGPT. Apple says that every time a key is pressed, the iPhone will run a Transformer language model under the powerful computing power of Apple's chips, which can complete a word or an entire sentence, and will also learn according to your writing style to guide its suggestions.
Image source: Apple's official website
Apple also said that speech recognition "employs a Transformer-based speech recognition model that uses a neural engine to make speech recognition more accurate." Other AI updates include Live Voicemail (live voicemail that supports real-time speech-to-text), Air Pods adaptive audio (which automatically adjusts the volume of media playback based on the environment), and more.
However, these updates can be said to be few compared to the new hardware and new systems that are overwhelmed at this conference.
Unlike Google and Microsoft in the field of AI, Apple seems to have maintained an "ambiguous" distance from AI, according to the New York Times report, Apple is still in the testing stage of large language models and other AI tools, and reporters noted that it did not open AI-related job postings until March this year.
According to Apple's official website, the company is currently actively looking for machine learning experts who are "keen to build extraordinary autonomous systems", covering integrated system experience, input experience NLP, machine learning research and development and technology development teams and other fields.
The "Daily Economic News" reporter combed and found that Apple is currently recruiting about 600 positions, 97 related to AI. More than one-third of them were newly created in May. In the first three weeks of May alone, Apple posted 28 AI-related jobs, including senior engineers, research scientists, special project managers and so on. Apple is also testing new natural language generation features for Siri, and the latest job listings include similar jobs.
By comparison, Google and Microsoft currently have 509 and 235 AI-related jobs open, respectively.
One of Apple's job postings is a job posting for a research engineer in visual generative modeling, which will be part of "a platform that will shape generative AI." In addition, while job descriptions vary, the vast majority contain expressions of "Siri's next-generation natural language understanding" and "a key shift in artificial intelligence technology leading the human-computer interaction revolution."
While it's unclear what Apple's open job descriptions mean, the job descriptions that are now open at least suggest that Apple has noticed the potential of generative AI and is taking action. In addition, the "next generation of Siri" is also one of the possibilities to look forward to.
Can Apple calm down?
Behind the "ambiguous" attitude towards AI, is Apple really lagging behind in this AI race, or is it waiting for the opportunity?
In this regard, Will Wong, senior research manager of technology consulting firm IDC in Singapore, pointed out in an interview with the "Daily Economic News" reporter, "It is too early to judge Apple's success or failure in the artificial intelligence race, especially because Apple always ensures that its products and technologies are ready to provide a good consumer experience before launch." Consumer experience has also been something Apple emphasized. So, with that in mind, Apple's current strategy may be to investigate whether generative AI can meet the most critical user needs and provide superior value to consumers. ”
Apple has always firmly articulated two main points of its AI strategy: efficiency and privacy.
Efficiency refers to machine learning algorithms and models executed locally that are more responsive and perform better. Privacy, as the name suggests, is privacy protection. Current generative AI seems to run counter to Apple's AI strategy.
First of all, generative AI similar to ChatGPT is completely dependent on the network, and even if it has been launched for more than half a year, the service is still not stable enough, and there will be downtime or errors from time to time. This does not allow such an unstable situation for a company like Apple that pursues user experience and efficiency.
Secondly, as early as the beginning of this year, some industry scholars and experts raised sharp questions that in terms of privacy and personal information security, ChatGPT may collect and process data and information without user authorization, or use it beyond the scope, resulting in personal information leakage. In addition, AI tools such as Midjourney and Stability AI have also been subject to infringement lawsuits by many websites and institutions, claiming that these companies illegally crawl copyrighted content for creation, which violates copyright law. Generative AI's lack of attention to privacy and copyright protection is also contrary to Apple, which has been focusing on privacy protection in recent years.
"Apple is a hardware company with chips and standalone operating system capabilities, which will give it an edge in deploying generative AI, especially since privacy and data security are several of the major issues facing generative AI today. The integration of hardware, software, and chips could allow Apple to better deploy tools on local devices, which in turn could help address issues such as privacy and data security. Will Wong added to every reporter.
According to Fast Company, a well-known US business magazine, while today's programmers mainly create AI applications that run in the cloud, they inevitably focus on building applications that run AI models locally on client devices such as smartphones and laptops. Against this backdrop, no company has benefited more from the developer ecosystem than Apple over the past decade. While Apple is not directly involved in generative AI, the company will support third-party generative AI applications on iOS and MacOS.
Avi Greengart, founder and chief analyst of market research firm Techsponential, also pointed out in an interview with the "Daily Economic News" reporter, "Although Apple seems to lag behind Microsoft and Google in terms of generative AI, Apple can afford to wait for generative AI technology to mature before integrating AI technology in a more meaningful place in its ecosystem." ”
Want to go your own path to AI?
Although Cook's answer to AI on last month's earnings call was extremely brief, he used a very interesting word to describe Apple's achievements in the field of AI: weave. Cook said that Apple is "weaving" artificial intelligence algorithms into its various products.
Over the past few decades, Apple has been developing and iterating on AI in its products, and the AI embedded in Apple products may be so much that users are unaware.
For example, Apple has always wanted Siri to be the most helpful AI assistant for Apple users, so Siri voice assistant has previously used natural language processing (NLP) technology from Apple devices to help users perform tasks. In addition, AI technology is also deeply integrated into the company's iOS software, with features such as Deep Fusion to improve image quality, a machine learning-assisted camera function that optimizes and improves the quality of photos pixel by pixel in seconds.
In this way, Apple is not not doing AI, but its approach is different from giants such as Microsoft and Google. There is an argument that Apple is unlikely to try to build an AI system with general-purpose large language models like ChatGPT or Google Bard.
With the layout of chips, hardware and software in the early years, Apple is not so much doing AI as doing the infrastructure for current generative AI deployment. Macs and MacBooks now run Apple's own M2 Pro and M2 Max chips, both with specialized 16-core neural network engines designed to accomplish artificial intelligence and machine learning tasks. Apple says the chips are 40 percent faster than their predecessors.
Apple's M2 Ultra is the world's most powerful chip ever built for personal computers. The Mac is also the most capable of running generative AI applications on the market — with an M2 processor capable of achieving an industry-leading 15.8 trillion operations per second (TOPS), and a powerful neural processing unit that could allow AI applications to run on the device without having to log into the cloud.
Image source: Apple's official website
Avi Greengart further analyzed to the "Daily Economic News" reporter that Apple's key advantage in using AI in future products is its reputation for vertical chip integration and focus on privacy. "Apple can build as much nervous system processing as possible on its M2 chip and then AI integration on the device, rather than relying on cloud builds."
Daniel Ives, managing director and senior equity analyst at U.S. brokerage Wedbush, said in an email to reporters, "We believe that Apple will eventually build its own artificial intelligence ecosystem in the iOS developer community, and the release of Vision Pro will bring more applications in the next few years." While investors' instinctive reaction will revolve around near-term expectations, we believe this will further create a developer moat for Apple, maintaining its solid foundation and Apple ecosystem. ”
"Overall, we believe Apple will continue to redouble its efforts on the software front, centering on the Developers Conference while releasing new product categories and key product updates that further illustrate the strength of the Apple ecosystem." We maintain our outperform rating and $205 price target," Ives added.
Cover image source: Daily Economic News Photo by Zhang Jian (data map)
Daily economic news