Produced by |Tiger Sniff Technology Group.
Author |Wang Xin.
Edit |Wang Yipeng.
Header |Visual China.
Apple has abandoned one of its most ambitious projects, building a car, and is considering moving 2,000 employees to its artificial intelligence division. A person close to Apple revealed to Tiger Sniff that people from the Vision Pro and automotive project teams had previously transferred to artificial intelligence projects.
As Steve Jobs said, "Deciding what not to do is just as important as deciding what to do." ”
Even if ten years of hard work to build a car have come to naught, in the face of OpenAI, which thinks different, Apple has finally woken up from its dream and decided to fully join the AI battlefield.
The world says that Apple's full shift to AI is long overdue, but Apple is just "getting up early in the morning and catching up with the evening". From Steve Jobs' early attempts at Atari's Arkanoid game to Siri, who taught all voice assistants a lesson in 2016, Apple has written several important strokes in AI history.
Now, Apple can no longer tolerate the "double reed" performance of OpenAI and Google, and chooses to focus on reinvesting in AI. As probably the only technology company with full-stack capabilities such as computing power layer, middle layer, model layer, and application layer, will Apple's entry into the game break new ground in the AI industry? In this AI war that has swept the world, what are the unique advantages of latecomer Apple?
Looking at the global AI, Microsoft and OpenAI are waiting for the princes and wolves; Google is following OpenAI, but it is following suit; Domestic factories are touching openai to cross the river, but they have the benefits of the moat. This three-point world and the pattern is undecided, the vast world is promising, if time, then the hegemony can be achieved, Apple can be prosperous.
Source: Visual China.
An ancient AI evangelist.
How early did Apple get up in the field of AI? Tracing back to the source, the starting time is far ahead of any current AI giant.
In 1975, Joe's gang leader was unfinished and Apple was not established, and he was still upgrading and fighting monsters at Atari Games. At that time, Atari was in the gaming industry, comparable to Nintendo in the mid-to-late 80s and PlayStation in the 90s. 20-year-old Steve Jobs and Steve Wozniak (Apple co-founder) spent four nights designing the hardware for Arkanoid.
Steve Jobs and Apple co-founder Steve Wozniak Source: Visual China.
Arkanoid was well received by players as soon as it was launched, and it has become a game system for the DeepMind team to train deep learning.
Beginning with the chess programs written by Turing and Shannon, the founding fathers of artificial intelligence, early AI was committed to creating a program that could defeat humans in games.
So DeepMind combines reinforcement learning with deep neural networks to create systems that can learn Atari games. After thousands of training sessions, the average score of deep learning networks in Arkanoid is 10 times that of humans. DeepMind became famous, and a year later it was acquired by Google, and Alphago was born, which started the last wave of AI.
In the field of speech recognition, a major breakthrough in artificial intelligence - natural language processing, Apple has been ahead of any mobile phone company.
In 2010, Apple acquired Siri. In 2016, Apple launched Siri, a smart voice assistant on the iPhone, which significantly improved the performance of voice assistants compared to other mobile phones. "Sometimes, a performance improvement is so dramatic that you test again to make sure you haven't missed a decimal point, and Siri is one of those cases. Apple engineers commented.
Missing out on the new wave of AI?
Although Apple indirectly gave birth to the last wave of AI, under the new wave of AI set off by OpenAI, Apple, which was caught off guard, has always been cautious and low-key, although it has made frequent moves.
In the new era of AI, Apple only does three things, but each one fails to achieve the extreme.
First, acquire a number of AI startups. According to the latest report released by market research agency Stocklytics, Apple acquired a total of 32 AI companies in 2023 and applied the acquired AI technology to product improvement. For example, in 2020, Apple acquired Dublin-based speech AI technology company Voysis to improve the experience of Siri human-machine conversations.
Second, research and development of large models and AI conversational robots. In July 2023, Apple was revealed to be developing a large model AJAX and an internal chatbot, Apple GPT. The AJAX system is built on top of Google's machine learning framework, Google Jax, and serves as the basis for Apple GPT, an internal ChatGPT-style tool, but it is more of an internal driver of internal products and is not open to consumers. But there are also Apple employees who claim that it essentially replicates Bard, ChatGPT, and Bing AI, and does not contain any novel features or technologies.
Third, Siri, which is considered to be the most likely to be connected to large models. In 2018, Giannandrea, the former head of AI at Google, joined Apple to lead the company's artificial intelligence and machine learning teams, pulling over the leadership of Siri, Apple's intelligent assistant. Siri has been criticized for its stagnancy, with Siri co-founder Dag Kittlaus saying that Siri may not have reached its full potential after being acquired by Apple. After ChatGPT upended the personal assistant experience, Siri was even more threatened.
Source: Visual China.
It is not difficult to see that although Apple has spent a lot of time and energy on artificial intelligence, it has also been a little hesitant. In response, Cook has tactfully said that Apple will add artificial intelligence to more of its products, but it will be "deliberate".
Previously, when asked about Apple's work on generative AI, Cook replied "stay tuned." "It's true that Apple hasn't made a big splash in the AI space yet," said Brian Mulberry, an investment manager at Apple shareholders. In the era of the new wave of AI, we do not feel Apple's enthusiasm for AI like Microsoft, Google, and OpenAI.
Where Apple AI comes in.
The day after it was revealed that he had abandoned car building and switched to AI, Cook reversed his attitude and said at a shareholder meeting on February 29 that the company would "break new ground" in generative AI in 2024, stressing: "We believe this will bring transformative opportunities to users." ”
So, what are the advantages of latecomer Apple? Will it break new ground in the AI world, as Cook said?
Qiu Zhen, an overseas partner of Huaying Capital, believes that Apple's advantage is that it may beIt is the only technology company with full-stack capabilities such as computing power layer, middle layer, model layer, and application layerUnder the multi-layer cost reduction and efficiency improvement optimization superposition of its own artificial intelligence chips, cloud computing, algorithms and data, the optimization effect of the model system will be very significant.
Therefore, Apple, which has reinvested in AI, can rely on its unique advantages to embark on a path of Apple's characteristic AI and dominate the AI competition full of gunpowder.
Denmark, Apple's symbol in the European data center Source: Visual China.
Next, let's expand and take a look at Apple's at each of the above levelsAdvantages:
Computing power layer: Currently AITraining sideHighly dependent on Nvidia, Apple has its own data centers in China, Europe, and the United States, so it has hoarded a large number of NVIDIA GPUs, at least it will not face a shortage of computing power; InInference sideApple's continuous investment in the field of AI chip design and development and the full-stack tool chain has made it possibleApple is fully capable of launching an inference acceleration solution based on its self-developed chipsand can be adapted to the self-developed pedestal model. Google's TPU and startup Groq have already made achievements in this area to match Nvidia's.
Data layerSince Siri, which has been integrated into the iPhone 4S in 2011, Siri has accumulated user conversation data over a period of more than ten years.
Application layer: The existing iOS system provides a platform and audience for content generation services such as text and audio for C-end users.
Qiu Zhen believes that if Apple from the aboveLarge model of the base and self-developed chipsThe cost of combining software and hardware, and doing the two things that OpenAI and Nvidia focus on at the same time is still very high, so it is necessary to temporarily abandon the concentration of resources for car manufacturing. Apple's adjusted priorities are likely to become:MR large model unmanned vehicle embodied intelligence
With the above advantages, in which areas will Apple, which has fully switched to AI, bring to users?Transformative opportunities
Let's start with the hot ones nowadaysAI phonesSpeaking of which, although domestic competitors have already entered the end-side large model a year ago, Apple's continuous investment in chips and self-research capabilities can call device CPU, GPU, NPU and other computing resources more efficiently than other mobile phone manufacturers. In addition, the operation of large models on the device side has extremely high requirements for memory rate and memory resources, and Apple can use its strong R&D and chain integration capabilities to solve hardware technical problems to the greatest extent. However, according to Apple's previous promotional tone, it is likely that it will not mention AI mobile phones and device-side large models, but will focus on applications. What solution will Apple find that is different from its competitors? The next release of iOS 18 and iPhone 16 will reveal the mystery.
For the next generation of mobile terminal platforms that Apple has given great hope tovision pro, AIGC's cost reduction and efficiency increase for its 3D scene construction will be greatly reduced. The real-world and user data obtained through sensors and built-in cameras can also be combined with AIGC for big data analysis and modeling to obtain better real-time interaction and user experience feedback.
A person close to Apple revealed to Tiger Sniff that Apple's abandonment of car manufacturing is likely to be just a temporary postponement of the car manufacturing plan and waitingAutonomous drivingAfter the technological breakthrough, more mature driverless car products will be launched. SORA's ability to learn and expand the laws of real-world physics and judge the trajectory of objects will bring new ideas to autonomous driving in the predicament. Apple is not completely giving up on building cars, but there is no need to make cars without self-driving. To do autonomous driving, AI is the only way, which is why Apple gave up making cars and turned to AI in an all-round way.
People who are changing and want to change the world are all on Tiger Sniff app