In response to user demand, Google’s conversational AI tool, Bard, has acquired a new set of skills to aid software developers with programming tasks, such as generating and debugging code and explaining its logic.
According to a blog post by Google Research Product Lead Paige Bailey, coding has been one of the top requests from users. Google has announced that it is launching these software development capabilities in over 20 programming languages, including C++, Go, Java, JavaScript, Python, and TypeScript. With Bard’s new capabilities, users can export Python code to Google Colab and receive help with writing functions for Google Sheets.
Bard, which was launched earlier this year as an AI experiment, can now review users’ source code line by line and assist with debugging. Developers can ask Bard to fix their code, and it will provide assistance. Additionally, Bard can translate code between languages and explain code snippets, making it a valuable tool for those new to programming.
During TechCrunch’s testing on chatbot performance, Bard, which was designed to rival ChatGPT and other language models, did not perform as well as its AI counterparts. However, with its new set of skills, Bard may be able to at least match ChatGPT and Claude in terms of capabilities, at least theoretically.
Nevertheless, it remains to be seen how effective Bard is in creating, translating, and debugging code.
As mentioned by Bailey in the blog post, Bard is still in its early stages and may sometimes present inaccurate, misleading, or false information with confidence. Therefore, developers should be cautious while using Bard’s outputs, as it may provide incomplete or incorrect code that fails to generate the expected output.
Despite these challenges, Google believes that Bard’s new capabilities can assist developers in novel ways, such as writing code, creating test cases, or updating APIs. If Bard uses an existing open-source project as a reference, it will cite the source, according to Bailey’s blog post.
