#精品长文创作季#
In the 2024 Hurun Global Rich List, half of the new wealth comes from AI. The direction is clear.
Among them, Musk is the richest man for a second time, and other people engaged in the semiconductor industry, such as Cambrian Chen Tianshi, MediaTek's Tsai Mingjie, and Dai Wenyuan from Beijing's AI company 4Paradigm, have also grown rapidly.
China and the United States are home to nearly half of the world's known billion-dollar entrepreneurs. China remains the world's largest billionaire entrepreneur with 814 entrepreneurs, with an average age of 60, and the world's largest number of self-made female billionaire entrepreneurs.
Have you decided to get into AI and embrace AI? But before you can do that, you need to figure out the following five things before you can grasp the turning point in your life of how to dance with AI.
Week after week, day after day, the news of AI is wave after wave, whether it is technological breakthroughs, product launches, industrial development, and even ethical discussions and social impacts, it is repeatedly stimulating the public's nerves!
Therefore, at the end of the first quarter, we will clean up the source in a timely manner, remove the chaff, and make a small inventory for you to help you grasp the latest cognition. Watch before you go!
First, the AI model bosses are polarized and reshuffled violently
The international giants of the large model, except for these four who are far ahead, have all been left behind.
The most popular Wenshengtu AI free and open source large model, stable Diffusion's team Stability AI, due to monetization problems, its technical core R&D team has collectively left, and SD is almost out. The application of some domestic shell SD will not be able to be upgraded in the future.
There is also a star AI startup Inflection, which has also had serious survival problems, because its CEO and most of the R&D personnel jointly invested in Microsoft's decision, this startup has been acquired by Microsoft in disguise.
Mistral, an open-source AI company comparable to OpenAI, has released a powerful flagship model, Mistral Large, which is currently closed-source, with only 7 billion parameters open-sourced.
After the reshuffle of competition in the underlying AI model, it may have entered a stage of integration, and there are currently only 4 companies left in the world that can make large models: OpenAI, Google, Meta (Facebook), and xAI. It is predicted that in the second half of this year, domestic AI large model companies will also enter the integration stage.
After 10 years of investing $20 billion in research and development, Apple announced that it would abandon new energy vehicles, not only switching its 2,000 employees to AI, Apple also announced that it is negotiating to bring Google's generative AI model Gemini and engine to the iPhone, and Apple has recently approached OpenAl and Baidu respectively and is considering using its model.
It is predicted that the leader OpenAI's ChatGPT5 will be released in the first half of the year.
Musk's large model chatbot Groc model, which challenges OpenAI, has open-sourced code, weights and architecture, with 314 billion parameters, making it the world's largest open-source model. This means that some small and medium-sized AI entrepreneurs in China can make their own large models without the need for tens of thousands of GPUs or hundreds of millions of US dollars.
OpenAI CEO Altman and NVIDIA CEO Jensen Huang jointly believe that the "computer that thinks like a human" AGI will arrive within five years, and Altman has predicted that AI will do 95% of the work of marketers, strategic planners and creative professionals. Hinton, the godfather of AI, pointed out in his speech that in 5-20 years, everyone has a 1/10 probability of being killed by AI. The benefits and risks brought about by the development of AI have become another acute issue in this round of technology wave.
When do you think the singularity of AI crushing human intelligence will come?
Second, the next stop in the AI game turned out to be electricity, photovoltaic and energy storage, and the ultimate PK is in energy infrastructure
The long-term, most direct, and shortest impact of the AI large model battle is the arms war of computing power. Masayoshi Son, the founder and CEO of SoftBank, is seeking to raise $100 billion to set up a chip company to compete with Nvidia and supply semiconductors necessary for artificial intelligence services. Previously, Nvidia had identified Huawei as its biggest competitor.
Chips are the core foundation of the artificial intelligence industry. AI NVIDIA's production capacity in 2024 has already been booked. The severe shortage of GPUs has forced openAI to slow down. Altman himself is also operating in the Middle East, planning to raise billions of dollars to build a global network of semiconductor fabs for OpenAI. The war between the global AI landscape and chip supply is about to break out.
AI is coming, chips are like gold, one of the direct relationships behind it is with everyone, and Musk reminded that silicon-based organisms consume huge amounts of electricity, and it is likely that the global power supply is insufficient.
Online AI computing is growing at a rate of 10 times every 6 months, and Musk predicts: humanity is on the verge of the biggest technological revolution ever, and artificial intelligence is facing a shortage, in addition to chips, the next shortage is electricity. Next year, humanity will face the dilemma of not being able to find enough electricity to run all the chips, and the world will face a power shortage.
Altman, CEO of OpenAI, and Jensen Huang, CEO of NVIDIA, also jointly said that the end of AI is photovoltaic and energy storage, and the ultimate PK is in energy infrastructure. OpenAI's ChatGPT consumes the equivalent of 100,000 homes for a day's lighting electricity every day, and it is expected that by 2030, extreme sets will consume as little as 8% of the world's electricity, including a minimum of 2% for data centers. During the "145" period, the mainland power grid plans to invest nearly 3 trillion yuan, and the current White House intends to invest 400 billion US dollars in research and development of clean energy technology.
Third, there are three things you must not know about humanoid robots
1. Musk believes that humanoid robots are more important than Tesla's car business, and that Optimus Prime robots are the key to Tesla's becoming a $10 trillion business. Nvidia CEO Jensen Huang said that embodied AI will be a new driving force to promote the next wave of AI, and FigureAI, a robotics company invested by Nvidia, Microsoft and Amazon, has raised nearly $700 million, claiming to make humanoid robots that walk and work independently within a year.
At the recent Global Science and Technology Conference GTC2024, NVIDIA showcased the GR00T Robotics Project, a bionic robot solution that is the world's first humanoid robot with language, video understanding, and human behavior mimicry capabilities, which can generate and execute actions in real time based on user instructions, enabling AI-powered robotics to be more deeply integrated into daily life. This reinforces NVIDIA's core position in the AI ecosystem.
2. Recently, Musk's flag company Neuralink released evidence of "fierce news". The first paralyzed patient to undergo Neuralink surgery will be able to control a computer and play chess with just his mind! This means that for the first time, human beings can control computers and play video games with their minds alone! In the future, brain-computer equipment + "thought" will not only solve many "disability" problems! It will also open the era of human superhumanism.
3. At present, the actual number of domestic robot applications is huge, 7 times more than the United States. The U.S. Information Technology and Innovation Foundation said in a report that the number of robots in China's workforce is 12.5 times the expected number, which the report argues shows that China is adopting robots at the fastest pace in the world, and that China is adopting automation far more than expected based on the wages of manufacturing workers.
On the other hand, the United States only uses 70% of its expected value of robots, the chairman of the American Information Technology and Innovation Foundation, the author of the report, said that China is already the world's largest industrial robot market, and its industrial robot installation in 2022 accounted for about 52% of the world's total, with 290,000 installed in China and 40,000 in the United States in 2022, a difference of more than 7 times.
Fourth, the United States accounts for 18% of the world's AI talents, and China has caught up with the United States in the cultivation of high-end AI talents
According to a recent news report by the New York Times, although the United States is a leader in the field of artificial intelligence innovation, China is already among the world's leading in measuring the cultivation of high-end AI talents.
From the comparison of CS Rankings AI single subject ranking, the number of excellent doctoral dissertations of the Chinese Society of Artificial Intelligence, CNUR artificial intelligence professional ranking and other indicators, the leading domestic universities in the field of artificial intelligence, comparing the data of Chinese artificial intelligence disciplines, combined with the comprehensive strength of universities, AI professional rankings, the list is Tsinghua, Nan University, Shanghai Jiaotong University, Peking University, China University of Science and Technology, Harbin Institute of Technology, Zhejiang University seven colleges and universities. (This ranking may change in the second half of the year...... ）
According to the Observer, the latest research report shows that China has surpassed the United States in exporting AI professionals, showing a significant advantage. According to data released by the McCropolo think tank, a subsidiary of the Paulson Foundation, nearly half of the world's top AI researchers come from higher education institutions in China, a proportion far higher than that of the United States, with only about 18% of researchers graduating from American universities. This trend shows that China's education system is playing an important role in nurturing AI high-tech talents.
According to the Securities Times, at the recently concluded Shengsi Artificial Intelligence Framework Annual Conference 2024, the director of the Science and Technology Development Division of the Science and Technology Department of the Ministry of Industry and Information Technology revealed that this year's artificial intelligence technology, especially in the field of large models, has shown significant progress, leading a new stage of rapid development of the industry. The mainland is standing in an unprecedented window of opportunity. According to statistics, by the end of 2023, the output value of the mainland's core artificial intelligence industry has approached the 600 billion yuan mark, building the three major industrial cluster effects of Beijing-Tianjin-Hebei, Yangtze River Delta and Pearl River Delta. The number of domestic core enterprises exceeded 4,400, ranking closely in the world, showing strong market strength.
Fifth, digital intelligence will eventually replace biological intelligence
Geoffrey Hinton, the father of deep learning and the father of neural networks, recently announced his departure from Google due to his fear of generative AI - "I regret my life's work. At the just-concluded closed-door forum of the Beijing International Dialogue on AI Security, Geoffrey Hinton argued that digital intelligence will eventually replace biological intelligence. AI trained on massive amounts of cyber information will be smarter than humans, and superintelligence may be able to manipulate humans with ease. Because the human brain can only share knowledge with each other through slow, inefficient distillation, different copies of a large model can share knowledge easily and quickly through weight or gradient sharing.
In fact, the transcendence of AI is not only the improvement of computing power, but also its advantages in overall organization and resource allocation. If multiple AIs can connect and collaborate with each other, they can surpass humans on a larger overall level, form a larger common whole, and emerge more powerful intelligence in higher dimensions. In addition, AI does not need to have independent "self-awareness" at the individual level, which makes it possible for them to form more optimal resource use and closer division of labor on a larger scale.
Yoshua Bengio, one of the three giants of deep learning, believes that the essence of AI is a machine with a purpose. A more powerful AI system is actually a machine that can achieve its goals. Who decides what the AI is targeting? It could be a malicious person or organization, or even the AI could be spontaneously formed. Therefore, Bengio believes that AI security and nuclear weapons issues are on the same scale. The more AGI projects there are around the world, the more dangerous it is for humans to come. "It's not a confrontation between nations, it's a struggle between humans and machines," he said. All of us should be in this together. ”
We need to build a global governance system to ensure that AI development benefits all of humanity and does not pose a threat to it. This means that we need a concerted effort from governments, international organizations, academia and civil society to develop appropriate regulations and standards to regulate the development and application of AI. Do you have any suggestions for this? Hurry up and see you in the comments section.