Text |Intelligent relativity.
Author |Shen Lang.
Perhaps, no one expected that future applications would come so quickly, and now traditional applications have entered an unprecedented stage of disruptive reform from development to experience.
Not long ago, the American artificial intelligence company OpenAI held a developer conference. On the spot, Sam Altma, the founder of the company, spent just 3 minutes to complete the construction of a project called Startup Mentor GPT app.
This demonstration shocked the industry.
Sam Altma, on GPT Builder, just typed a definition of GPT — he wanted an app that would help startup founders think about their business ideas and get advice. GPT Builder then automatically generates the app's name, icon, and features. Users only need to access this GPT, and they will automatically enter the conversation interface and ask them about entrepreneurship-related content.
Traditional applications are already facing disruption, and the era of AI agents is coming - this is the signal that OpenAI sends to the market. In China, technology companies represented by HUAWEI CLOUD are also proposing a similar trend"Applications will come in the form of AI agents, applications will have the ability to understand, plan decisions and perform complex tasks autonomously, and human-computer interaction will be as natural, immersive, and personal as human-to-human interactions. ”
For a long time, the industry has continued to explore the development of application modernization, and with the explosion of AI large models this year, perhaps its future evolution ideas and market forms will be clearer.
Future Applications:Extension of human cognition
What form will future applications take?Returning to the essence of application, what it actually represents is the extension and strengthening of human capabilities.
For example, the search, e-commerce, social and other applications that have emerged in the past represent an extension of human perception, through which people can obtain information, goods and interactions thousands of miles away.
When this perception is extended and strengthened to a certain height, the era of super apps has arrived. For example, WeChat, Alipay, etc., people only need a super app to obtain clothing, food, housing, transportation, shopping, social networking and other services, which is very powerful.
However, with the increasing number of super apps, the evolution of apps has entered a "bottleneck period" - no matter how many functions and services a super app integrates, the user experience is already very limited.
As a result, super apps began to evolve into specialization. Taking the Huawei Health app as an example, it integrates many sports and health-related functions, and can not only scientifically monitor and manage health with smart wearable devices such as Huawei Watch, but also intelligently improve users' sleep quality through multi-modal fusion algorithm models.
As a result, the evolutionary logic of the application changesOn the one hand, it is still an extension of perception, which corresponds to a super application, but it is more focused and professional. On the other hand, it is the extension of cognition, which corresponds to intelligent applications, and the service is more active and intelligent.
Not long ago, OpenAI teamed up with SKT, one of the world's largest telecommunications companies, to hold the "ChatGPT App Development Competition". Judging from the results of the competition, the new applications built at this stage are clearly manifested as the extension and reinforcement of cognition.
Among them, the Born to Be Prompters team, which won the first prize, developed a generative AI application called GLESOM that helps people with autism simplify complex sentences;Team Hyper has developed a generative AI app called Jikimi that analyzes emails, emails, and text messages to alert users to potential security risksThe app is trying to understand and think about human language and feed it back to the user.
In addition, the core of the service of Startup Mentor GPT, which Sam Altma demonstrated and created at the developer conference, is also to understand and think about the rules and content of human business society, and to provide guiding suggestions for users. - To put it simply, the app is learning to recognize the world, which is an extension of human cognition.
Therefore, the market has every reason to believe that the next direction of application modernization will most likely be the improvement of cognition, and more and more intelligent super applications will replace and subvert today's application ecology.
That is, the most talked about AI agent in the industry, the application will be presented to users as an expert assistant, with a highly integrated professional knowledge system, and the ability to independently understand and perform complex tasks to provide users with more intelligent and proactive services.
Application Modernization:An evolution that continues to push boundaries
The extension and strengthening of perception and cognition have laid the foundation for the evolution of application modernization, but what kind of future applications will actually appear in front of users?INTERVIEWER This question does not seem to have been answered clearly, that is, what changes will AI agents bring if they replace traditional applications?
Judging from the results presented by the industry so far, from super applications to AI agents, the commonality lies in the continuous breakthrough of boundaries. To put it simply, application modernization is a process that continuously breaks down boundaries, and it is manifested in three things.
One is the boundary between scenes.
Previously, Musk has repeatedly proposed to create a similar "X APP", or "Everything APP", to provide users with one-stop services such as clothing, food, housing, transportation, information and news, consumption and entertainment, and at the same time through the establishment of a huge and prosperous digital ecosystem, so that users can get a good service experience and great convenience.
In fact, about this vision, WeChat, Alipay and other super applications have been initially realized. In this process, super apps such as WeChat and Alipay have become difficult to locate in a single field or scenario. This means that super applications have made breakthroughs in scene boundaries, and in the future, future applications represented by AI agents will continue to break through scene boundaries along the path of super applications.
The second is the boundary between the characters.
In the traditional process, the developers and users of applications are often different groups of people, but with the evolution of application modernization, the development model represented by low and zero is emerging, and business personnel who do not understand the best also have the opportunity to complete the development of a light application. At present, both overseas OpenAI and local HUAWEI CLOUD are committed to simplifying the application development process and steps to help users quickly complete application development.
For example, CodeArts Snap, an intelligent programming assistant created by HUAWEI CLOUD, has Xi learned 76 billion rows, selected tens of thousands of classic technical documents, and more than 85 million open source warehousesThese capabilities go far beyond the lifetime of an average programmer, so that with the support of intelligent programming assistants, non-developers can easily develop their own applications, build and use them, and continue to break down the boundaries between roles.
The third is the boundary between interactions.
Although many applications are very powerful on mobile phones, once migrated to other terminal platforms such as smart TVs and car machine systems, it is very "chicken". In general, the boundaries of human-machine interaction are still relatively obvious, and applications are far from achieving the effect of switching between different terminals and providing seamless services.
However, this limitation is bound to be broken. At present, local vendors, represented by HUAWEI CLOUD, are continuously strengthening the cross-platform development and deployment of applications, and AIGC applications can be developed on terminal platforms such as automobiles, Internet of Things (IoT), and the Internet based on global serverless. It is conceivable that in the future, users may only need to wake up and call their own exclusive AI agent on different terminals such as TVs, cars, mobile phones, and computers, just like Iron Man calling "Jarvis".
In general, from the perspective of current applications, future applications have already had a preliminary prototype, and it may not matter what form and mode they will serve users in the future. The orientation of application modernization is to break down all kinds of boundaries, so that users can enjoy seamless services at any time, place, and terminal for the first time, just like the experience brought by today's mobile apps, but also more advanced, intelligent and proactive.
Back to the cloud and industryof the home field
Is this a wild imagination or a reality that has not yet arrived?The answer must be found in the present.
On the one hand, the intelligent performance brought by AI technology is becoming more and more important in the upgrading of applications, and intelligence has become the key direction of application evolution.
On the other hand, whether it is a scene boundary, a character boundary, or an interaction boundary, there are applications in the industry that are completing the initial leap, and the problem is more about how to further upgrade the functions and services of the application on the broken boundary.
In addition, the most critical underlying technology, cloud computing, and the cloud industry behind it are maturing. At this stage, cloud vendors are basically providing support for the massive data, huge computing power, and complete ecology required for the development of AI large models.
For example, HUAWEI CLOUD has built three AI cloud computing centers in Gui'an, Ulanqab, and Wuhu to provide enterprises with surging Ascend AI computing power, so that enterprises no longer need to rush to buy GPUs or worry about potential continuity challenges of GPUs in the future. At the same time, HUAWEI CLOUD has also launched the Ascend AI Cloud Service 100-mode Zone, where enterprises and developers can directly use mainstream open-source large models, such as LLAMA and GLM, to implement intelligent upgrades of their own applications.
In addition, based on industry practices and business requirements, HUAWEI CLOUD has proposed six key technologies for application modernization, including intelligent R&D, assembled delivery, digital intelligence-driven, service-oriented architecture, security, trustworthiness, and resilience, to comprehensively support the application modernization process.
At this stage, the digital transformation of enterprises has been steadily promoted, and most business applications are running on the cloud, which is the basic foundation for application modernization. On this trend, the comprehensive assistance of cloud vendors will accelerate the process of application modernization from the bottom.
There is no doubt that application modernization will well promote the "two-way rush" of the artificial intelligence industry and the cloud industry - based on future applications, the artificial intelligence industry and the cloud industry will be fully developed.
Of course, the premise is that future applications cannot be separated from the industry, otherwise it will be really fanciful.
Today, HUAWEI CLOUD, together with the China Software Industry Association (CSA) and a number of industry customers, software providers, and consulting organizations, has established the Application Modernization Industry Alliance, which aims to promote the efficient and high-quality implementation of application modernization in an industry-oriented manner.
Then, when the technical conditions tend to mature and the industrial environment becomes more rational, it is not difficult to predict the future trend of application modernization by the market and users - AI agent has been initially presented to the public, and the landing path of future applications has been clear and clear, just wait for relevant manufacturers to continue to make efforts.
All of this article is on the Internet.