Editor's note: 2023 is the hottest year in human history on record. This year, after 1,191 days, the pandemic no longer constituted a "public health emergency of international concern", and the global order began to be restored and reshaped at an accelerated pace. However, the world is not peaceful, the smoke of gunpowder between Russia and Ukraine has not gone away, the Palestinian-Israeli conflict has restarted, and the financial market is full of rumors and opportunities.
When the inflation situation eased, the pace of global central bank interest rate hikes began to reverse, and the emergence of ChatGPT demonstrated the infinite possibilities in the field of natural language processing. A new year is coming, is the world ready for the next "pandemic"?
2023 is known as the first year of generative AI. From ChatGPT sweeping the whole network at the beginning of the year to the hundreds of schools of thought competing with various large language models since then, people have witnessed the huge opportunities and challenges brought by artificial intelligence to business and society. However, in the penultimate month of this year, OpenAI, the "ancestor" of ChatGPT, has experienced a reversal and reversal and reversal of the "palace fighting drama", which seems to be just a matter of CEO Sam Altman's stay, but behind it is related to the business interests, scientific ethics, regulation and security of artificial intelligence.
ChatGPT爆火
ChatGPT, a chatbot that came out in November last year, quickly became popular online this year, with OpenAI having more than 100 million active users just two months after its release. In January, Microsoft announced that it would make a significant investment in OpenAI to accelerate its technological breakthroughs in AI.
After a series of tentative discussions, the charging plan of the AI chatbot ChatGPT has surfaced. On February 1, the artificial intelligence laboratory Open AI announced on its official website that it would launch a paid subscription version of "ChatGPT Plus". ChatGPT, which has been free for two months but has 100 million monthly active users, has finally embarked on its own "way to make money", and as a result, the discussion on the commercialization of AIGC has also heated up sharply.
On February 8, Microsoft launched the latest version of the artificial intelligence search engine Bing (Bing) and Edge browser powered by ChatGPT, and Microsoft, Google, and Baidu are fighting fiercely in the traditional business and AI track. According to the development trajectory of ChatGPT and the overlap of the three companies' businesses, it is likely that the chatbot will be implanted in the search engine.
On March 15, OpenAI officially released the upgraded GPT-4. ChatGPT has ripped off the "Chat" label and evolved again, and GPT-4 not only demonstrated stronger language comprehension skills, but also was able to process image content, and even scored more than 90% of humans on the test.
100 model wars
ChatGPT is like a depth charge in the AI circle, and the shock caused by it has gradually spread to all walks of life, and has set off a "100-model war" in China.
In March, Baidu took the lead in launching the "Wenxin Yiyan" model, in April, Alibaba Cloud launched the "Tongyi Qianwen" model, in May, iFLYTEK released the "Spark Cognition" model, in July, Huawei Cloud Pangu Model 3.0 and JD Cloud Yanxi Model were launched one after another, and in September, Tencent's hybrid model was released...... Not only Internet companies, including the three major operators of China Mobile, China Unicom and Telecom, home appliance manufacturers such as Changhong, but also scientific research institutes and universities such as Tsinghua University, Fudan and the Chinese Academy of Sciences have also released their own large models. In such a prosperous atmosphere, it took a year for the domestic AI circle to quickly build hundreds of large models. Compared with the original "100 Group War" in the takeaway industry, the "100 Model War" has a more demanding demand for funds. Therefore, every participant in the "100 Model War" needs to continue to think about how to find the direction of application landing as soon as possible while narrowing the gap with ChatGPT to achieve self-hematopoiesis.
OpenAI宫斗
On November 17, Altman was abruptly dismissed from OpenAI's former board of directors, and then his ally on the board and former chairman of the board, Brockman, also announced his resignation. On November 18, OpenAI's investors pressured the board of directors to reverse the decision to remove Altman.
On November 20, more than 95% of OpenAI employees signed a joint protest letter, saying that if OpenAI did not dissolve the company's board of directors, reappoint Altman as CEO, and return Brockman, they would resign and join Microsoft. Ilya Sutskevi, chief scientist at OpenAI, who is believed to be the mastermind of the incident, also signed the joint letter and issued a statement saying that he deeply regretted his previous participation in the board of directors.
Late at night on November 21, the American artificial intelligence company "Open Artificial Intelligence Research Center" (OpenAI) announced on the social media platform X (formerly Twitter) that it had reached an agreement in principle that Sam Altman, the company's co-founder and former chief executive officer (CEO), known as the "father of ChatGPT", would return to the company as CEO. So far, this OpenAI "palace fighting drama" that stirred up the technology industry has ended in a dramatic way after 100 hours of dizzying reversals, reversals, and reversals.
"Shovel seller" Nvidia
Since the explosion of ChatGPT, Nvidia has been recognized as the biggest "shovel seller" in this global AI gold rush wave, and it is also the most discussed AI chip company in major media and social platforms. However, as the AI boom continues to heat up, more and more manufacturers have also begun to make efforts in the field of AI chips: before Intel, AMD and other semiconductor giants announced a new round of AI chip research and development plans, and then OpenAI, Microsoft and other downstream customers promoted self-developed chips to break NVIDIA's monopoly.
On November 15, local time, Microsoft launched two custom chips at the Ignite developer conference held in Seattle. Microsoft said that the newly released chips will not be sold and will only be used to support its own products. However, Microsoft's launch of self-developed AI chips is not intended to replace manufacturers such as Nvidia.
In the context of the shortage of chips and the surge in demand for AI applications, Microsoft's rapid advancement of the custom chip plan is widely interpreted by the outside world as the only choice for cloud computing giants with resources.
On the heels of Microsoft, Sapeon Inc., a semiconductor startup owned by South Korean wireless carrier SK Group, has announced that the company has unveiled its latest artificial intelligence chip, the X330.
Prior to this, tech giants such as Amazon and Google had also used self-developed chips and partially offered them to customers. AMD, another major chip manufacturer, also launched the latest accelerator card Instinct MI300X not long ago. At the press conference, a line of words was specially typed on the PPT - special for large language models, which was regarded by the industry as a direct declaration of war against NVIDIA.
It is reported that the MI300X's high-bandwidth memory (HBM) density can be up to 2.4 times that of the NVIDIA H100, and the high-bandwidth memory bandwidth can be up to 1.6 times that of the H100.
Regulation has arrived
Since the advent of ChatGPT last year, it has caused a huge social controversy. As AI capabilities far exceed expectations, critics are concerned about rapidly evolving AI technologies including privacy violations, bias, fraud, and the spread of rumors and disinformation. Industry technology leaders have also recently been calling for caution about the risks posed by AI.
On March 29, thousands of technologists signed an open letter calling on all AI labs to immediately suspend training on systems more powerful than GPT-4 until the associated risks and problems are addressed.
On October 30, U.S. President Joe Biden signed an executive order on artificial intelligence, which was called "the most significant action taken by the government to advance the field of artificial intelligence security" by the U.S. media.
Russian President Vladimir Putin announced on November 24 at the "2023 Artificial Intelligence Journey" International Artificial Intelligence and Machine Learning Xi Conference that he will approve the new version of the National Artificial Intelligence Development Strategy, and one of the key tasks is to expand basic and applied research in the field of generative AI and large language models.
On the evening of December 8, local time, the European Union reached an agreement on the "Artificial Intelligence Act". A prominent feature of the bill is its focus on risk-based regulatory regimes, which classify AI systems into four levels: minimum risk, limited risk, high risk, and unacceptable risk, so as to balance AI innovation and development with safety norms.
In addition, countries such as Canada, South Korea, and Japan have also begun to adopt regulatory and restrictive measures in the field of artificial intelligence.
Of course, as a major AI country, China has never been absent in global AI governance, and has been committed to improving the safety, reliability, controllability and fairness of AI technology.
>>> 展望2024
The hype has become a reality in an all-round way
Ethical considerations are prioritized
In 2023, the AI wave led by OpenAI's GPT model will sweep the business world at an unprecedented speed, and the market will see the power and limitations of AI. Some institutions predict that by 2025, the global artificial intelligence market is expected to reach a staggering $190.61 billion, with a compound annual growth rate of 36.62%. For 2024, the industry generally believes that it will be a year to fully implement the hype into reality.
At the industry level, Bryan Harris, CTO of SAS, said: "Generative AI technology can do a lot of things, but it's not a panacea. In 2024, companies will adjust their generative AI strategy from using it as a single technology to incorporating it into their industry AI strategy. ”
When predicting the development trend of the AI industry in 2024, AI experts from Forbes Biweekly and Nvidia also believe that 2024 will usher in a big explosion of enterprise generative AI applications, and many companies will deploy hundreds of customized AI models to be applied to different business domains and processes to solve specific problems and explore opportunities. These AI applications will be deeply integrated into business systems and fine-tuned based on the enterprise's own data to provide users with personalized services under the premise of security and control.
However, TechCrunch, a well-known technology media, predicts that many companies will abandon the large-scale application of AI in their businesses because the benefits of AI tools cannot justify their costs and risks.
Regardless of strategy, looking at the AI industry as a whole, many companies are trying to make AI products actually make money. In a report published by analyst firm CCS Insight, the generative AI space predicts that the field of generative AI will be tested for reality in 2024. The report notes that the growth of generative AI is about to slow as the technology hype fades, running costs are rising, and regulatory calls are rising.
When it comes to a specific company, TechCrunch predicts that after the "Gong Dou" and leadership changes in November 2023, OpenAI will become a company that is more focused on "launching products", and the impact of Sam Altman taking charge of the company more holistically will gradually be felt at all levels of the company, as exemplified by the GPT Store. In 2024, OpenAI may launch more blockbuster products similar to the GPT store.
TechCrunch also made predictions about Apple's progress in the field of AI. TechCrunch believes that Apple's approach has always been to wait, wait and see, learn from the failures of other companies, and then join in in a refined way to lead the technology trend. 2024 is a good time for Apple to show its skills in the field of artificial intelligence, and the time left for Apple to exert its product power has come, but it is difficult to say whether Apple's AI service will be an improved version of Siri or a completely new service.
In terms of ethics, legislation and regulation, Forbes biweekly and NVIDIA AI experts believe that the unprecedented development of AI has also brought countless ethical problems and become a major challenge for today's society. In 2024, there will be a growing interest in AI ethics, and how to ensure the healthy, sustainable and responsible development of AI will become an important topic, and ethical considerations in the field of AI research and development will also be prioritized.
As a result, the pace of AI legislation will be further accelerated. Currently, major countries, including China, the European Union, the United States, and India, are working to develop comprehensive AI policies. Countries have developed AI policies with three goals: to drive technological breakthroughs and attract global investment, while protecting their own populations from any undesirable effects of AI. Discussions within the industry suggest that potential international synergies suggest that global collaboration on AI benchmarks and specifications may soon be realized.
Big moves such as AI legislation come with the potential to change the way the industry operates, but these regulatory rules tend to be slow to take effect, and except for companies that voluntarily make changes, they may not have much effect for a long time, and companies will use lawsuits to challenge these laws. TechCrunch expects this to lead to a boom in the AI compliance industry: after all, companies that spend a lot of money on large model development and computing power will most likely not want to spend a lot of extra money on resolving highly uncertain risks such as litigation.
Comprehensive report by Beijing Business Daily