Chatgpt, a chatbot platform with AI OpenAI power supply, can now directly edit the code-if you are on macos, that is.
The latest version of the CHATGPT application for MacOS can take action to edit the code in supported programming tools, including Xcode, VS Code and Jetbrains. Users can optionally turn on the “automatic approval” mode so that chatgpt can edit without the need for additional clicks.
CHATGPT Plus, PRO and Team subscribers can operate the code editing function from Thursday, updating the MacOS application. Opeli claims that the code edition will be introduced to Enterprise, EDU and free users next week.
In the post on xAlexander Emricos, a member of the OpenAI product staff working on computer software, added that the CHATGPT application for Windows will receive a direct edition of the “soon” code.
Chatgpt for MacOS can now edit the code directly in ides. Available for Plus, Pro and Team users. pic.twitter.com/wpb2rmp0tj
– Developers OpenAI (@OPENIDEVS) March 6, 2025
The direct edition of the code is based on the “Work with Apps” ChatgPT function, which the company has launched in the beta version in November 2024. “Work with applications” allows the CHATGPT application for MacOS to read the code in a handful of coding environments focused on the developer, minimizing the need to copy and paste the code to chatgpt.
With the ability to directly edit the code, CHATGPT now competes more directly with popular AI coding tools, such as Cursor and Github Copilot. Apparently OpenAI has the ambition to introduce a dedicated product to support software engineering in the coming months.
AI encoding assistants become extremely popular, and The vast majority respondents in the latest GitHub survey, saying that they adopted AI tools in some form. Partner Y Combinator, Jared Friedman, claimed that a quarter of the W25 YC startup party has 95% of code databases generated by AI.
However, there is a number of risk of security, copyright and reliability related to AI supporting tools. Survey from a software seller bundle It was found that most programmers spend more time debugging the code generated by AI and security compared to inscriptions written by people. Google reportMeanwhile, it was found that artificial intelligence could speed up code reviews and benefits documentation, but at the expense of delivery stability.