The world’s first true AI PC is here. It integrates computing power, sensing, intelligence, security and personalization, and generates three-dimensional empowerment of video, music, graphics and text. The disruptive changes triggered by artificial intelligence have begun to detonate the creator economy.
At the Tech World 2024 technology ceremony held in Shanghai on April 18 with the theme of “AI for All, filling the world with AI”, Lenovo Group released a new species of AI-AI PC series products with built-in personal intelligence “Lenovo Xiaotian” At the same time, it also released innovative applications and solutions such as enterprise intelligence based on the core propositions, strategic vision and latest technological achievements of hybrid artificial intelligence, as well as Lenovo’s all-in-one heterogeneous intelligent computing platform. But the most eye-catching thing at the scene was the conversation between Lenovo Group CEO Yang Yuanqing and Lu Chuan, a sixth-generation director with a sense of technology geek.
Yang Yuanqing believes that AI will make it easier to transform human imagination into reality. AI is Artificial Intelligence artificial intelligence, which may be more appropriately called Augmented Intelligence. Smart devices such as AI PCs enhance people, support creation, and deeply empower creativity.
Director Lu Chuan said that when faced with artificial intelligence technologies such as multi-modal large models, he initially felt shocked and anxious. Later, after some exploration and attempts, he discovered that AI is actually a powerful weapon for creators. At the Tech World site, he also shared the wonderful clips generated by the team based on creative use of AI. The short clips did not stop at special effects. Lu Chuan believes that AI can greatly improve the speed of visualizing ideas and help thousands of young people with movie dreams to cross the financial threshold and directly present the great movie in their minds. It is said that Lu Chuan will also work with Lenovo to build an “AI imaging laboratory” to open up a new world for the application of AI in the film industry.
AI PC detonates the creator economy at just the right time: Zhike is stirring up China’s economy
Filmmakers, musicians, designers, media people, programmers and even scientists like Lu Chuan who use artificial intelligence to create and explore have now become a broad group that is beginning to emerge. In the changing Chinese economy, a phenomenon-level development hotspot has begun to emerge: the creator economy driven by AI productivity. Not only emerging but also rising rapidly, the new concept group of “intelligenters” has begun to stir up the industrial market and is becoming a new vitality for China’s economic development. The creator economy, redefined by strong intelligence, involves tens of millions of people and has an output value of trillions. It is a new growth pole that cannot be ignored.
Previously, the main body of the creator economy was considered to be content creators on various online platforms, which usually generated income through platform incentives for image, text, video, and music creation, user rewards, advertising sharing, or knowledge payment. But now the definition of creators and the creator economy is being completely changed by AI. The field, scope and scale are far beyond those before. In addition, smart devices with sharply increasing computing power and other AI-related performance, as well as various large, medium and small-scale generative models on the cloud side and on the device side, not only bring about “changes in tools”, but also lead to complete changes in the form itself.
This is a picture of an era in which people create new intelligence and intelligence in turn enhances people.
The intelligent customer group behind the creator economy are the first to make full use of strong intelligence to carry out creative and innovative activities. They appear in 9 areas:
Area 1: From film and television to short videos, video smart people are emerging, and visual content creators are changing in phase and intelligence.
Creators of movies, TV shows, and short videos have begun to apply AI and multi-modal large models to support their daily work. There are both efficiency improvements in the instrumental sense, and intellectual change in the sense of phase change – new forms of content creation have become a reality. The former is an intellectual change, and the latter is a phase change driven by intellectual change. In addition to conventional operations such as automatic subtitle generation, green screen keying replacement, and subject recognition tracking, it is not uncommon to use Pika, Runway, Stable Diffusion, etc. to create micro-movies and short plays. SORA’s targeted cooperation with the film industry, Adobe, etc. before its release is also accelerating this process. There are more than 2.1 million active UP owners at station B alone. There are more than 100,000 micro-short drama creators on the Kuaishou platform. There are more than 470,000 Internet celebrities with more than 100,000 fans on Douyin platform, and the number of knowledge creators is close to one million.
Video smart users are not the same as these groups. Even if they do not use SORA, film and television practitioners who can make good use of the AI capabilities of Adobe AE, Premier, and Davinci will have different competitiveness. Fully AI-generated videos actually lack viewership in the short term, but the practice of using AI to generate effects and scenes for some clips, or using AI to generate video content to avoid copyright issues is quickly becoming common. A one-person, one-player drama with personalized plots must rely on multi-modal AI. It will also become common for ordinary individual users to generate interesting videos and share them in the content market to gain revenue.
Area 2: Music Intelligence, AI-powered music creation has shown a high level from the beginning.
The feeling generated by the video to the user may be fake, but the music generated by AI gives the user a feeling that is real. This is a wonderful feeling, but it is also consistent with the human-machine experience. Because users naturally think that the content of the video should be produced in reality, unless it is animation, and music is originally created out of nothing, artificially created to express feelings and joy. A considerable part of the music generated by user lyrics in Suno, Udio, etc. has shown a relatively audible level, and AI musicians have emerged.
Previously, the number of independent musicians who settled on NetEase Cloud Music exceeded 400,000, and the number of musicians who settled on Tencent’s platform was approximately 300,000. However, now that Suno has enabled everyone to create music, the scope of music geeks is no longer limited to traditional musicians. The momentum of domestic and foreign AI music competitions seems to be even more natural than that of AI short drama competitions. However, although Udio has a Google background, it is obviously much worse than Suno. Currently, there are not many AI music creation platforms to choose from, and the community has not yet formed a climate.
Area Three: Design Smart Guests, art and industrial design are all embarking on the journey of AI creation and intelligent design.
Designers and editors have always needed high-performance computing equipment, including displays, that far exceeds the needs of ordinary users. Nowadays, designers in the sense of design geeks have more than just the realization of simple needs such as Dall-E, Midjourney, Wenxinyanyan, and Tongyi Qianwen, such as Vincentian pictures, Vincentian paintings, and Vincentian LOGO. Tools such as Adobe Firefly, which can quickly generate vector graphics and text effects from text, are advanced but not the end point. CAD (computer-aided design) and CAE (computer-aided engineering) under the integrated architecture of sensing, computing, data, and intelligence make both art design and industrial design even more powerful. AI computing power chips have been fully introduced into the design process, not to mention the daily life of art. The German artist Beuys’s slogan “Everyone is an artist” was so inspiring half a century ago, but today it has finally begun to become a reality.
Area 4: Avatar smart guest, generate yourself, digital people allow podcasters, media, and self-media to have intelligent avatars.
Regardless of live broadcasts or recorded broadcasts, the application of digital people by broadcasters, media, and self-media is accelerating. For example, the digital people of Liu Qiangdong, the founder of JD.com, have recently attracted attention. The industry that provides digital human services and the field that applies digital humans are growing side by side. The market size of the digital human industry will exceed 20 billion yuan in 2023, driving the market size of related industries to exceed 200 billion yuan. It is expected to reach 48.06 billion yuan and 640.27 billion yuan respectively in 2025. The live broadcast conversion rate of virtual digital people is one-third that of real people, but the cost is low and the time is long-lasting, especially small, medium and micro enterprises tend to choose it. However, digital people and intelligent avatars based on real people are more popular with users than virtual idol digital people. “Multi-modal large model + digital human” is closer to the ability needs and future form of intelligent clones than just being a digital human.
For C-side users, digital humans and smart avatars are also beginning to be used in meetings, social networking and other scenarios.
Area 5: AI empowers the entire online shopping process, e-commerce smart customers are becoming more and more dynamic, and traditional online stores are thinking about changes.
The total number of domestic online store registrations exceeds 100 million, and there are tens of millions of active stores, including more than 1.2 million Taobao stores that have been established for more than 10 years (more than 5 million new stores were added last year). There are more than 8 million video authors on the Douyin platform, and these are the market base. AI e-commerce is not just a new way of playing, but is becoming the new gene of “e-commerce smart customers” throughout the entire process. Including C2B, which all parties have long expected, can only be truly realized in the strong intelligence stage. The value of e-commerce traffic will gradually complete the transfer from traditional recommendation algorithms such as advertising to AI e-commerce. That is, AI e-commerce will form a value closed loop of purchase conversion through full personalized perception, understanding, interaction and communication with users. Relevant research predicts that 3.13% of the total online retail sales of 32 trillion yuan in 2030, which is 1.01 trillion yuan, will be the return generated by the platform traffic value from the perspective of AI e-commerce.
E-commerce traffic value AI e-commerce will gradually complete the transfer of recommendations from advertising and traditional algorithms to AI. AI will recommend products that best meet expectations to consumers in the form of dialogue. The indicator for measuring its value capability is also called TR. In the short term, it will mainly come from the inherent TR of the e-commerce platform; in the long term, it will bring about natural growth as AI matures and further penetrates. It is expected that by 2030, the monetization rate of AI traffic will reach 3.13%. Corresponding to the estimated total online retail sales of 32 trillion yuan, AI will contribute 1.01 trillion yuan in e-commerce traffic value.
Platforms are promoting AI e-commerce, and companies themselves are also taking action. The waiters of enterprises and merchants behind every online store are undergoing three changes. First, customer service, especially nighttime customer service, is introducing intelligent robots, including corresponding customer service and customer service management. Second, the introduction of 24-hour “personnel” with intelligent blessings. “In the live broadcast room, thirdly, in terms of marketing, accurate content and even precise videos are generated by intelligent dynamics, which is more effective than behaviorally targeted advertising in the past.
Area 6: GPT smart guests who specialize in value sharing move from specific knowledge and capabilities to characteristic intelligent scenarios.
Share your own GPTs to large model stores or markets, or provide apps with different functions for other users. The personalized customization of large models allows each user to generate GPTs that solve different problems in different fields, and at the same time, they can share the GPTs they think are good with other users to gain profits. The framework, knowledge and data submitted by users are equivalent to retraining, and personalized GPTs that mount knowledge bases in specific fields are more popular. Various Apps with different functions are one step closer, just like the rich Apps or small programs in the mobile Internet era. It took Apple App Store 15 years to reach 1.78 million Apps, and it only took two months for Open AI’s GPTs to reach 3 million. Opening Store is becoming a common practice for general large models. This aspect has started relatively quickly in China, and the development of third-party plug-ins and tools is gradually gaining momentum. Open AI has gathered more than 2 million Apps developers, and there are tens of millions of users from different industries and professional backgrounds who can potentially contribute GPTs.
Area 7: LoRA Zhike, deeply customizes models for users, with fine-tuning and meticulous attention to detail.
The full name of LoRA in English is Low-Rank Adaptation of Large Language Models, which is literally translated as low-level adaptation of large language models. In layman’s terms, it means fine-tuning model parameters at low cost to make the model perform better in some aspects, such as many LoRA models of Stable Diffusion. , let the model learn a specified style or character, so that consumer-level computer users can easily generate specific content, and obtain various prompts without achieving the effect. Although it is called fine-tuning, for users, it is equivalent to in-depth customization that is far better than the native model. It can be said that fine-tuning is not fine-tuning, and the requirements are sufficiently differentiated. The demand for meticulous fine-tuning also in turn increases model fine-tuning to A segment with a large enough base. LoRA Zhike, also known as model fine-tuner, is leading the popularity in the community of various open source models, and obtains paid rewards from users who download and use it. Although not everyone can do fine-tuning, compared with developing and training models, fine-tuning is much less difficult and its application scenarios are much wider.
Area 8: New paradigm of IT development, program smart people innovate to build a technology-driven world.
Big models are replacing programmers without ideas, but at the same time are enhancing programmers with ideas. Program generation capabilities empower programmers, improving development efficiency while allowing programmers to focus more on innovative architectures. There are nearly 8 million programmers in our country who use strong intelligence to develop efficiently. This potential group has nearly 8 million people. Although a program can be generated in one sentence, and future programming will be “natural language”, good programs still require programming experts who have a deep understanding of the program itself and architectural capabilities. The complexity and user experience also require more professional technology. personnel to grasp. Software and algorithms are even more so.
Area 9: Scientific intellectuals, intelligent analysis of scientific research data, development and application of intelligent models for scientific research in various fields.
If technology is the primary productive force, AI is becoming the primary productive force of technology – the force of innovation. AI for Science can be called a historical opportunity for scientific and technological innovation. Using intelligence to comprehensively change the process from scientific research to industrial implementation, what makes all this a reality are these “scientific wise people” who actively use intelligence to make full use of it: use GPT to read global papers to absorb cutting-edge wisdom, and use calculations to Powerful equipment analyzes scientific research data, uses professional models to analyze and predict protein structures, researches new molecular drugs based on massive data and algorithms, predicts and creates new materials, introduces strong intelligence into clean and sustainable energy research, and introduces Transformer into meteorological, climate, geographical and geological analysis… …
The creator economy, as a new growth pole for a trillion-dollar industry with tens of millions of people, has value not only in creation, but also in fundamental changes in the relationship between individuals and ecology, man-machine, and man-intelligence. The essence of the creator economy is technology-driven, innovation explosion, intelligent inclusiveness and human enhancement. It is not only the development engine of the smart economy, but also a symbol of the rise of the smart economy.