Iced latte

Edited by Fang Qi

Media|AI Large Model Workshop

When I woke up, the industry was swiped by Zhihuijun's "Robot Mobilization": On the morning of August 18, at the 2024 new product launch conference of Zhiyuan Robot, Zhihuijun released a total of five commercial humanoids of the two families of "Expedition" and "Lingxi", including interactive service robot expedition A2, flexible intelligent robot expedition A2-W, etc. At the press conference, a series of fancy operations such as the robot's "playing mahjong, dismantling express delivery, and lifting 30KG heavy objects" made netizens continue to swipe the screen??? To show shock and admiration.

Zhihuijun, the boss behind Zhiyuan Robot, is also a typical example of the industry's "technology boss leaving his job to start a business" - he joined Huawei through the "Genius Youth Program" in 2020, and left at the end of 2022 to found, and after its establishment, he received investment from BV Baidu Venture Capital and was listed as a unicorn company by Forbes magazine.

In the past year, Zhiyuan Robot has raised 6 rounds of financing, including Sequoia China, SAIC Investment, etc., and its valuation has soared to 7 billion yuan.

is not only Zhihuijun, at present, the resignation of the big man to start a business has become the most "city" behavior nowadays. Some time ago, Zhou Chang (nickname: Zhong Huang), the core technical backbone of Ali Tongyi's large model team, was revealed to have left his job to start a business. According to reports, a person close to Alibaba Cloud revealed that it was a normal resignation.

Like many bigwigs, Zhou Chang's resume can be called impeccable, both in terms of education and career experience:

Chang Zhou graduated from Fudan University with a bachelor's degree in computer science and technology in 2012 and a Ph.D. degree in computer software and theory from Peking University in 2017. After graduation, Zhou Chang joined Alibaba, and successively served as an algorithm expert at the Damo Academy and the head of the large model of Tongyi Qianwen, with his main research directions including graph learning and reasoning, recommendation systems, distributed graph computing, etc., and published more than 30 articles in international summits in related fields.

It is worth mentioning that in 2017, when Dr. Zhou Chang graduated, Alibaba DAMO Academy was just established, and since then, Zhou Chang has been in the workplace for seven years, and it is also seven years of rapid development of Ali's technological innovation. During this period, Zhou Chang led the team to develop the 10 trillion multimodal pre-training model M6 series, the general multimodal AI model OFA series, and the large-scale GNN training platform AliGraph, witnessing and participating in the whole process of the Ali large model family from scratch and from existence to strength.

Among them, the M6 model was jointly released by Alibaba and Tsinghua University in 2021, with a parameter scale of up to 100 billion, becoming the world's first 10 trillion parameter multimodal large model, and it is also the predecessor of the Alibaba Cloud Tongyi large model that is now well-known in the industry. Now, when the family matrix of the Tongyi model is fully formed and enters the high-speed development trajectory, Zhou Chang chooses to "retire after success" and pursue his dream in the entrepreneurial circle.

This is also in line with the consistent style and career trajectory of AI bigwigs in recent years: after completing the experience in a large factory, they will turn the rudder and leave the port and start their own business. made the same choice, as well as Zhou Chang's old colleague Yang Hongxia, who served as the president and AI scientist of the Dharma Research Institute, and led the team to develop artificial intelligence open source platforms and systems such as AliGraph, M6, and Luo Xi.

In May this year, it was reported that Yang Hongxia was preparing to start a business and anchor the AI Agent track; Recently, according to the latest reports, Yang Hongxia became a professor at the Hong Kong Polytechnic University, which makes people sigh that she has achieved both "academics and entrepreneurship".

Not only Ali bosses, but in the tide of AI entrepreneurship, more and more bigwigs have come out of big factories to start a new stove and take charge in multiple subdivisions. For the bigwigs, with the rapid development of AI technology, it is the best time to "transport heroes to freedom" and realize technological ideals. For the AI industry, with the rise of the "new forces" of AI startups, the industry landscape is also changing.

First, the big guy resigned, "scattered is full of stars": in a number of application fields

In the AI track, it is also an unbreakable truth that the times create heroes. After a hammer-making stage full of uncertainties, AGI finally ushered in a golden age as the large-scale model entered its first year of landing. The bigwigs are undoubtedly the first to hear the call of the times, and they are inevitably itchy.

For technology bosses, although they can fight side by side with the big team and master the front-line resources in the Internet giants, they are often subject to the company's strategic direction and resource allocation, and do not have so much autonomy. As Baidu CEO Robin Li once explained, in the era of large models, AI is a "top project". For technology bosses, leaving their jobs to start a business and becoming their own CEO can undoubtedly specialize in the field they are best at and are interested in, realize their technical ideals, and stand in the center of the stage.

At the same time, startups are more flexible and flexible, and can quickly adjust their strategies and seize the opportunities of the golden age of AI in the still-uncertain and rapidly changing AI application field.

So, what are the directions that the bigwigs are interested in? Looking at the entrepreneurial track of the bigwigs, they all point to the landing and application of AI.

In this wave of entrepreneurship, the ceiling-level existence is undoubtedly Li Feifei, the godmother of AI - became a tenured professor in the Department of Computer Science at Stanford at the age of 33, and became an academician of the United States National Academy of Engineering at the age of 44, Li Feifei can be said to be a legend in the AI industry. In September 2018, after leaving Google, he returned to Stanford to teach and plan to start his own business, and founded the AI company World Labs in April this year.

According to reports, the specific direction of World Labs is the "spatial intelligence" that Li Feifei has repeatedly proposed - that is, to push computers into the three-dimensional world, teaching computers how to perceive and act in the real three-dimensional world.

"We want more than just AI that can see and speak, we want AI that can do." Li Feifei wrote on social media that "spatial intelligence" is a key piece of the AI puzzle that can turn "vision into insight; Insight becomes understanding; Comprehension drives action. ”

In his keynote speech at TED in April, Li gave a concrete example to illustrate this idea – showing a photo of a cat stretching out its paws to push the glass to the edge of the table, explaining that the human brain can quickly assess the geometry of the glass, its position in the 3D world, its relationship to the table and the cat, and predict what will happen next and take action to stop it.

Obviously, Li Feifei's goal is to make artificial intelligence grow "human brains" in reality - once computers learn how to survive in these 3D worlds, it will be even more promising in the field of application. If you don't forget, there will be repercussions, and now, Li Feifei has finally started his own business.

is also interested in "injecting souls into AI and letting AI grow brains" and putting it into entrepreneurial practice, as well as Chen Junbo, the former head of autonomous driving at Ali Damo Academy and the founder of Youlu Robot after leaving his job, and his entrepreneurial projects are more down-to-earth.

According to reports, relying on the self-developed second-generation embodied intelligence technology LPLM model, Youlu Robotics is committed to making every professional equipment intelligent. When it comes to the original intention of entrepreneurship, Chen Junbo once said that embodied intelligence needs to obtain data in the real physical world to improve the world model like Tesla, but if you want to obtain data from the real physical world, you must first complete commercialization:

"[Only] if we can continue to put the product into the physical world, can we be the first to complete the world model together with the language model. Therefore, since its establishment, Youlu Robotics has taken product commercialization as the first task of research and development. ”

This also speaks to the common sentiment of the bigwigs - in the process of changing the world with technology, entrepreneurship may be the only way to the ideal highland.

Now, the bigwigs are gaining the favor of massive capital to accelerate the completion of the implementation of technical ideals.

Second, the influx of capital, the formation of the market, policy support: the time comes and the world are together

PitchBook data shows that in the past three months alone, investors have poured more than $27 billion into United States AI startups.

Recently, the United Kingdom Financial Times reported that Li Feifei's World Labs has conducted two rounds of financing within more than three years of its establishment, including the well-known Silicon Valley venture capital a16z, with the latest round of financing of about 100 million US dollars, and the current valuation has reached 1 billion US dollars (about 7 billion yuan).

The embodied intelligence track, which is also favored by capital, is also the Zhiyuan Robot introduced above, which was founded by Zhihuijun after his resignation. In July this year, iSoftStone Information Technology (Group) signed a strategic cooperation agreement with Zhiyuan Robotics. The two parties will carry out in-depth cooperation in the fields of general robot product innovation and scene demonstration, and work together to promote humanoid robots to become universal and serve thousands of industries.

When a strong capital meets a strong market demand, a strong landing force erupts. At present, the bigwigs are not "having technology but no market", but they are born at the right time - the popularization and application of AI technology has brought huge market demand. Whether it is consumer electronics, intelligent manufacturing, or medical healthcare, they all welcome AI technology. This demand provides a broad entrepreneurial space and market opportunities for technology bigwigs.

Taking the medical field as an example, according to IDC statistics, it is expected that by 2025, the total value of the global artificial intelligence application market will reach 127 billion US dollars, of which the medical industry will account for one-fifth of the total scale, and will be one of the fastest growing tracks in the next five years. Among them, combined with the estimation of AI-assisted new drug research and development and AI-assisted tumor diagnosis and treatment, the compound annual growth rate (CAGR) from 2020 to 2025 is expected to be 39.4%, and it will exceed 30 billion yuan in 2025.

In this track, Senyi Intelligence is one of the "nugget ceilings". Zhang Shaodian, the founder of Senyi Intelligence, is also a technology tycoon, with a bachelor's degree in AI and computer science from Shanghai Jiao Tong University and a doctorate in medical informatics from United States Colombia University.

At present, Senyi Intelligent is the core supplier of AI technology of CLP Data of the National Health and Medical Big Data Group, providing professional, safe and efficient overall medical big data solutions for medical institutions at all levels, scientific research institutions, government departments, etc. The stable and continuously expanding market has taken advantage of the technological outlet, making it the darling of capital:

A few days ago, Senyi Intelligent announced the successful completion of a new round of strategic financing, so far, Senyi Intelligent has successfully completed 9 rounds of financing in the 8 years since its establishment, totaling 1.315 billion yuan.

Not only that, in terms of policy, under the wave of new quality productivity, AI startups have also received more and more attention and support from the government and all walks of life.

For example, on July 23, in order to promote the high-quality development of Hangzhou's artificial intelligence industry, the Hangzhou Municipal People's Government issued the "Several Measures to Support the High-quality Development of the Whole Artificial Intelligence Industry Chain" from five aspects: "construction of computing facilities", "model open ecology", "empowerment of the real economy", "development of the whole industry chain" and "support of talent team".

When the heavens and the earth work together, the heroes are transported to freedom. Nowadays, with the continuous entry of bigwigs, new and old forces continue to collide, and the industry pattern has also changed.

3. After "heroes to freedom", the AI landscape has changed

Nowadays, as the bigwigs are working alone, AI talents are more scarce for large manufacturers, so they can only accelerate the cultivation of emerging forces and accelerate the battle for talent.

According to the United States Department of Labor, the average salary of an AI data scientist is $120,000 to $200,000 a year, and the average annual salary of a machine learning engineer is as high as $125,000 to $200,000. As the value of talents has skyrocketed and good generals are hard to find, large manufacturers have offered higher prices. According to reports, OpenAI offers an annual salary of up to $540,000 to $910,000 for AI software engineers, and up to $2.595 million for Google AI researchers.

In China, Lin Fan, the founder and CEO of Maimai, has publicly stated that talents in some technical positions continue to be scarce, and the supply-demand ratio of large-scale technical experts is only 0.25, which is equivalent to 4 posts competing for 1 person.

On the other hand, with the financing of the bigwigs and the research and development of the R&D, the AI ideal is in full swing, and some doubts have quietly emerged. From the perspective of the industry, technology bigwigs can get hundreds of millions of financing with a single shout, which undoubtedly accelerates the bubble of the AI track.

Recently, according to France's "Les Echos", the field of AI chips has been particularly turbulent recently. Massachusetts Institute of Technology professor and economist Daron · Ajemuoglu predicts that only a quarter of tasks automated by AI will be profitable in the next decade.

Goldman Sachs, an international investment bank, was similarly skeptical, publishing a 20-page report in June titled "Generative AI: Spend too much, gain too little?" pointed out that tech giants are spending $1 trillion on AI capital expenditures over the next few years, but there are few substantial, visible results to justify the investments.

This is also the cruel side of the technology bigwigs to achieve their ideals: the road to the ideal highland is inseparable from the coronation of capital, and the final result needs to be tested by time. In this "AI tide to the sea", some people will get everything they want, and some people will inevitably come and go home with excitement and failure.

This article is from Xinzhi self-media and does not represent the views and positions of Business Xinzhi.If there is any suspicion of infringement, please contact the administrator of the Business News Platform.Contact: system@shangyexinzhi.com