Usama is a news reporter at XDA who has a decade of experience in covering technology news across a variety of topics, with the primary area of focus being anything related to Microsoft.
He also has a bachelor's degree in Computer Science and loves to play video games and read books in his free time.
AI applications and large language models (LLMs) are all the rage these days, with Microsoft and OpenAI mostly leading the charge. However, prior to the recent explosion in the popularity of this domain, Microsoft had already given us a hint of what's next when it introduced GitHub Copilot back in 2021. The Redmond tech firm's lead in this area has become evident with the rollout of applications like ChatGPT and Bing Chat, which are both quite proficient — if not perfect — in writing code and fixing bugs in it. Although Google Bard can write code too, Google now seems to be following in the footsteps of Microsoft by launching AI-powered pair programming features in Colab, likely to compete against Microsoft.
For those unaware, Google Colab offers Jupyter notebooks hosted on the cloud so that developers can write and execute code in Python directly in the browser without downloading any additional software. While it is free to use, Google offers multiple payment tiers in case your code is compute-intensive and cannot run sufficiently well on the free tier. Colab is a fairly mature tool since it was launched back in 2017, and now, Google wants to supercharge it by offering AI-assisted programming capabilities similar to Microsoft's GitHub Copilot.
Google has announced that Colab will leverage the code models offered by Codey to facilitate code generation and completion capabilities, along with a specialized chatbot to enable developers to write high quality code faster. The company has emphasized that Codey's training dataset uses "permissively licensed" code that is tailored to Python and Colab coding environments. Codey is based on PaLM 2, which is the latest LLM powering the new version of Google Bard.
The new AI features in Colab will initially be available only to customers based in the United States, with Google emphasizing that the initial release will prioritize code generation based on natural language prompts. This should reduce the time needed to write boilerplate code, enabling developers to focus on higher-priority aspects of their software. Programmers will be able to press the "Generate" button at the top of their Colab notebook to write a prompt and generate code. Meanwhile, the integrated chatbot will serve as a conversational interface where you can ask specific questions about the code and other use cases such as "How do I import data from Google Sheets?"
Google has noted that the Codey-powered integration will be available free of charge, which is good news for the seven million customers, mostly comprising students, that Colab currently boasts. That said, paid customers in the U.S. will get access to the AI features first, followed by free-tier users in the same region. Once this initial rollout is complete, Google plans on expanding to more countries.
It is important to note that this is essentially the first competitor against Microsoft's GitHub Copilot in the big tech domain. Although it does have a major selling point in the form of free access compared to GitHub Copilot's $10/month base tier, it is still severely restricted in what it can do. Codey is limited to Colab code on the cloud, while Copilot integrates with multiple popular on-premises integrated development environments (IDEs), including Visual Studio, Visual Studio Code, Neovim, and more. Additionally, Codey's coding capabilities cover only Python, whereas Copilot caters to developers across multiple tech stacks, including Python, Java, C, Ruby, and Javascript, among others.
Regardless, the intention from Google is clear here. The company is intent on challenging Microsoft in the AI-assisted programming space, and the logical next step might include offering integrations with IDEs, including its own Android Studio.
