LAS VEGAS — Amazon Web Services showed a new service that uses machine learning to generate code suggestions for software developers, stepping up its efforts in the growing area of AI-powered programming.
The new service, Amazon CodeWhisperer, offers code recommendations based on contextual information, such as past code a developer has written; and natural language prompts, such as “generate a subroutine to upload files to S3.
Swami Sivasubramanian, AWS vice president of database, analytics and machine learning, announced CodeWhisperer Thursday morning at the company’s re:MARS conference in Las Vegas.
It’s part of a trend in software development. The announcement comes a day after Microsoft-owned GitHub made its Copilot AI pair programming service generally available for $10/user per month.
CodeWhisperer automatically determines which cloud services a developer commonly uses, and best practices for coding in specific languages, Sivasubramanian said during the unveiling this morning. The machine learning model has been trained on Amazon code and open source code, he added.
Sivasubramanian also outlined CodeWhisperer’s security precautions and protections against bias.
Amazon says CodeWhisperer supports code written in Python, Java, and JavaScript, using VS Code, IntelliJ IDEA, PyCharm, WebStorm, and AWS Cloud9, with support for AWS Lambda Console coming soon.
