Less than half a year after ChatGPT came out, this industry has been AI-based.
It is of great significance that AI has entered the productivity phase.
On Monday, the artificial intelligence drawing tool firefly presented by Adobe and Avida detonated the circle of friends.
This is a major upgrade of the Firefly. Although the ability to edit videos has not yet been opened, this tool has been able to realize the creation of text images, sketches, and one-click modification of the content of the picture. It can also modify, layered, and output ultra-high resolution automatically generated images, initially reaching the productivity standard.
The product launched by Adobe is the answer given by the software giant to the impact of AI drawing-since the diffusion model series shocked the painting field last year, people are expecting artificial intelligence technology to do some work at hand, and the demo of some products has made us eager to try.
In addition to images, the natural language processing direction of artificial intelligence has undergone an unprecedented breakthrough. Since the end of last November, the ChatGPT launched by OpenAI has become a hot topic. Facing the new situation, domestic and foreign technology companies have followed up at an unprecedented speed.Technology giants are trying to build their own versions of ChatGPT and hope to realize real generic intelligence (AGI) through larger-scale models.
However, during the primary activity of the Ministry of Science and Engineering of the Ministry of Science and Technology (MIT), OpenAI Chief Executive Officer Yamamu/Sam Altman (Sam Altman) was retired. Other warning names, the life ChatGPT-like research ploy, and the future of the future.At present, OpenAI has no training GPT-5 and will not be trained in the short term.
Is this the end of the big model era?
In addition to adding more parameters to the model, there are many other ways to make the big language model (LLM) better and more useful. On the other hand, many companies and institutions are exploring the application of big models. From the perspective of practical technology, our imagination may not keep pace with the pace of AI innovation.
ChatGPT is coming, what will the world be like? After the launch of GPT-4, OpenAI has also studied it seriously. A paper submitted by OpenAI points out that ChatGPT may affect 80% of jobs, and jobs with higher salaries will face more potential impacts from AI models.
The beginning of productivity reform
In other words, after the breakthrough of AI technology, many jobs have a lot of space to improve efficiency. If you give orders and AI is responsible for doing things, it will be a brand new work form. Is the good day for workers coming? We can see the clue of change in some things.
The first is the computer field itself. Based on the GPT-4 of language big model GitHub Copilot, a chat mode centered on software program code is built, covering all processes in the development life cycle. By integrating chat and voice functions, developers can naturally communicate with AI.In addition to automatically writing code, Copilot are also integrated into pull requests, command lines, and documents to provide instant answers to project-related questions.
According to statistics, AI has written 46% code in the current work using GitHub Copilot, which can help developers improve their work speed by 55%.
The Office series for most Office scenarios are also undergoing AI. On March 17th, Microsoft officially announced the Microsoft of 365 Copilot, which covers all components of Office buckets.
Based on Copilot, GPT-4 can call other software materials in the user’s Word, and directly make Excel tables and PPT according to the Word content to realize the conversion of content forms. In addition, it can also change the writing style according to users’ needs and realize personalized writing,Automatically generate Excel analysis reports.
Microphone CEO’s proposal/Display, Copilot-like output: one-piece Monument this is a kind of creative construction method.
As more and more productivity tools based on big model technology are gradually launched, we may be entering a new era. The outbreak of artificial intelligence is described as the fourth industrial revolution-the first three have greatly improved productivity, but at the same time provided more advanced production tools, but the fourth industrial revolution is somewhat different,”Finished products” can appear automatically. You only need to think about many things you want to do, and you no longer need to do them.
However, just as code generation tools are used for programming, in order to automate in many vertical fields, specific optimized AI tools are required. After all, as Zhang Yong, CEO of Alibaba, said, “All products are worth redoing with AI models.” what other businesses have been automated? Recently, an automation tool based on the big language model (LLM),It has made a good start for UI design.
Subvert a vertical industry
On March 28th, a UI-oriented AI product, which enables AI to generate editable UI design drafts for “real-time AI” open internal testing applications, quickly attracted the attention of the industry and received tens of thousands of applications within 24 hours. So far, more than 30000 people have participated in the internal test,And more users are waiting for the release of the internal test qualification.
According to the inventory report released by the authoritative user experience platform UI China, there are currently three products around the world that can implement the function of AI to generate editable UI, namely, real-time AI, Galileo AI, and Uizard. Among them, two foreign products are still in the small-scale testing stage.Currently, “real-time AI” is the world’s first product that allows users to experience the function of “AI generates editable UI design draft” on a large scale.
Real-Time AI is based on the online collaborative design platform real-time design. It uses a large language model to enable people to quickly generate editable UI design drafts through natural language descriptions on the web page.
With instant AI, AI can automatically generate a UI interface template for you with a simple description:
If you provide a detailed description of the interface requirements and write the requirements with limiting statements like in AI drawing, then AI can complete the task as required.
As a practical productivity tool, real-time AI has the component recognition capability. It can recognize not only a single layer, but also most components on the page, such as buttons, title bars, and bottom bars. With component-level recognition capabilities, designers can use this tool to generate design drafts that conform to different design specifications.
The design draft generated by AI includes vector layers and icons, and the hierarchy is automatically organized to support manual secondary editing.
Now, designers only need to input text instructions to quickly obtain the UI design draft that initially meets the requirements, is editable, and has a clear layer structure. In the conventional UI design process, a page takes about 30 minutes for the designer on average, while the time for using “real-time AI” to generate a page is only 1 minute for you to write a prompt and wait for the big model to respond,And the generation time will be compressed again as the version is updated. The introduction of AI technology can give you more opportunities to sort out ideas and explore different designs.
Of course, this set of tools is not intelligent enough to replace designers. The design generated by it needs many modifications before it can be used, and the scope of generation is limited. The evaluation used by many people is that the generation level is similar to the high-fidelity prototype map of the product manager or the level of the primary designer. If you reduce the limit description,Let AI play freely. At present, it will be a good brainstorming tool.
Designers said that after the emergence of AI generation tools, the shape of UI design will change. First of all, after people give their needs, designers can provide party a with a variety of design schemes for selection in a short period of time, and quickly feedback and adjust further requirements. On the other hand, if you encounter relatively simple and previous requirements,You can also quickly complete tasks in a short time.
This tool was updated again this week, increasing the number of drawings produced each time. Due to the large number of people applying for internal testing, an additional 20,000 internal testing places were added.
AI application landing, faster than expected
In the vertical field of UI design, the most popular tool in the world is Figma. In last September, Figma was acquired by Adobe at a high price of $20 billion, which became a big news in the industry.
In order to keep the leading position in the industry, Adobe has offered a price that people cannot refuse.
Interestingly, Figma’s vision is to “eliminate the gap between imagination and reality”, which has been realized by AI technology to some extent.
Since the design was officially launched in September 2020, its user scale has exceeded 2.3 million. At present, the start-up company is still exploring more productivity tools. The launch of “real-time AI” has also brought more imagination to the industry.
Referring to the evolution speed of mainstream AIGC tools, perhaps in the near future, we will see that real-time AI will generate UI pages close to the level of senior designers, directly generate multiple series of UI design drafts, and even directly produce available final solutions, and finally have a complete design system.
It must be noted that AI based on large models cannot solve design problems in one step. However, designers can greatly save time and give full play to their ideas and ideas with the help of AI.
Think from this perspective, build a set of AI auxiliary tools, let designers use the “prompt” of natural language text or a small part of design as guidance to quickly generate, and then get started to improve, perhaps is the way of working in the future.
In short, “knowing what needs are” will be much more important than “knowing how to do it” in the future.
In the future, many workflows of each company will be automated by AI. From user research, requirements, prototypes, beautification, front-end to back-end, most of the processes may be simplified by AI.
Facing this trend, are you ready?
