Ant Group's CodeFuse launches the 'Picture to Generate Code” function, and more than 50% of programmers use AI to write code

On April 11, CodeFuse, Ant Group’s self-developed intelligent R&D platform, launched the new function of “Code Generation from Drawings”, which allows developers to generate code with one click using product design drawings, greatly improving the development efficiency of front-end pages. The relevant functions are currently in internal testing.

Like many Internet companies, Ant Group is fully promoting AI programming. More than 50% of engineers use CodeFuse to support daily research and development work, and 10% of the codes submitted by these engineers are generated by AI.

Gartner pointed out in the top ten strategic technology trends released in 2024: By 2028, 75% of enterprise software engineers will use AI programming assistants. CodeFuse is an exploration attempt under this trend. According to reports, CodeFuse’s functions cover the entire R&D link such as demand analysis, programming development, testing and construction, release and operation and maintenance, and data insights. For example, during the development and testing phase, CodeFuse can complete code, add comments, parse code, and generate orders. Testing, code optimization, etc. help developers write code faster and easier.

Currently, within Ant, more than 50% of programmers use CodeFuse in daily research and development every week. The overall adoption rate of code generated by CodeFuse is 30%, and the adoption rate in generating unit test scenarios can reach 50%.

(Picture caption: Ant intelligent R&D platform CodeFuse covers the entire AI R&D link)

The "graph generation code" function released this time mainly serves front-end engineers . In the Internet product development process, after the designer draws the design drawing, the front-end engineer needs to use code to implement the product design drawing. This work takes up a lot of development workload. Drawing-generated code can generate code with one click based on design drawings, which can greatly reduce the coding workload of the development team when developing web pages, small programs, and APPs. Taking a medium-sized web page as an example, if the final code is about 200 lines, it takes one person about 1 hour. After one-click generation, engineers only need to check and adjust, and the time is greatly reduced. This function is developed based on the multi-modal technical capabilities of the Ant Bailing large model.

The CodeFuse team pointed out that the popularity of AI can not only reduce the work pressure of developers, allowing them to devote more energy to more creative work, but also has greater significance in lowering the programming threshold and promoting software Innovation and progress in the development industry. CodeFuse's mission is to explore the next generation of AI R&D productivity tools and is committed to creating innovative solutions to make software developers' development process as smooth as silk. After generating code from natural language and generating code from graphs, CodeFuse will continue to launch new functions to help enterprises improve the efficiency of the entire R&D link.

The person in charge believes that the change in the AI ​​R&D paradigm no longer means that the role of "people" in R&D scenarios will disappear. Instead, it puts forward higher requirements for how AI and people can collaborate, especially when it comes to flexibility. and reliable operation and maintenance scenarios, which also require manual expert intervention to make the system run healthily. "AI is currently mainly focused on auxiliary programming (code copilot). There is still a long way to go to move from copilot to co-worker and realize the intelligence and automation of the entire R&D life cycle."

The above is the detailed content of Ant Group's CodeFuse launches the 'Picture to Generate Code” function, and more than 50% of programmers use AI to write code. For more information, please follow other related articles on the PHP Chinese website!