AI-powered coding assistant aims to help, not replace developers

GitHub Inc.’s new automated computer coding tool will help companies manage a chronic shortage of software developers, but it won’t replace them, according to the company’s CEO.

Thomas Dohmke, CEO of Microsoft Corp.

GitHub, owned by GitHub, explains that the artificial intelligence-based coding tool, launched this week as Copilot, is designed to make it easier to write long lines of computer code from scratch.

Software code is the math-like language that powers computer applications, which underpin many critical business functions, from sales and marketing to human resources, accounting and finance.

“The process of writing code remained mostly manual,” Dohmke said Wednesday at Collision 2022, a startup and investment conference held in Toronto.

Instead, he said, the AI ​​coding tool acts as a predictive text feature on smartphones. As developers work, the tool suggests options for the lines of code needed to perform given tasks, within a larger software program. Developers, who can also ask for suggestions, then choose which option to use.

The tool was created in partnership with OpenAI, an AI software maker best known for its GPT-3 natural language model that can mimic handwriting.

The same way natural language systems are trained on words, Copilot’s AI model is trained on billions of lines of code in languages ​​ranging from Cobalt to JavaScript, Dohmke said. Based on this huge code warehouse, it suggests the next line as developers type in programming apps like Visual Studio Code or Neovim.

The idea is to speed up the time it takes to build new apps by reducing the amount of code developers have to write manually.

He said the tool will also make the process of building custom apps more accessible to workers who have a rudimentary understanding of code, as the tool can walk them through the basics of app development.

GitHub, a San Francisco-based online coding platform that lets developers collaborate and share pieces of code, said that over the past year, more than 1.2 million developers have signed up. registered to use a preliminary version of the tool. The company said tests indicate the tool can cut app development times in half, with an average of 40% of code in users’ apps automatically generated. GitHub charges users $10 per month or $100 per year.

Mr Dohmke called software developers “the backbone of an organization”.

Last month, U.S. employers posted about 620,000 ads for open tech jobs, up about 50% from the same month in 2021, according to IT trade group CompTIA. Software developers were by far the most sought-after tech workers, accounting for more than 200,000 job openings in May alone, the group said.

Mr Dohmke said automated coding is meant to supercharge developers, accelerating what they are able to achieve and stretching their imagination. “AI will not replace developers,” he said. “He can’t predict the ideas in your head.”

Rob Zuber, chief technology officer at San Francisco software maker CircleCI – whose own coding platform competes with GitHub – said it’s risky for companies to rely too heavily on technology to do the job. work of a human developer, especially when building software-based operations at scale.

“Software development is complex, and you can’t replicate the fundamental practices and experience that a human brings,” Zuber said. CircleCI currently employs approximately 200 in-house developers.

A basic principle of coding that developers learn is not to repeat the same lines of code, and letting software generate copies of code goes against this fundamental principle, he said. “If you have long lines of baseline code, you should reorganize your code to eliminate them, not auto-generate them,” Zuber said.

As with most enterprise automation software, such as robotic process automation, GitHub’s Copilot is more likely to be used as an assistant, not a replacement for workers, said Peter van der Putten, director from the AI ​​Lab of software company Pegasystems. Inc.

“Automated text completion simply saves time, just like typing text on your phone or editing a document,” said van der Putten, who is also a professor. adjunct in AI at Leiden University. “It can provide a trigger to get started, even if the code is a little off,” he said.

Write to Angus Loten at angus.loten@wsj.com

Copyright ©2022 Dow Jones & Company, Inc. All rights reserved. 87990cbe856818d5eddac44c7b1cdeb8

Comments are closed.