OpenAI has introduced an upgrade to its macOS ChatGPT app, integrating the AI tool with coding environments such as VS Code, Xcode, Terminal, and iTerm2. The new feature, named “Work with Apps,” enables ChatGPT to read active code within these applications, offering assistance without requiring users to copy and paste code snippets manually.
“Work with Apps,” aimed at making coding more seamless, is accessible to Plus and Teams users, with plans to expand to Enterprise and Education accounts shortly.
A Step Forward for Contextual Assistance
With the “Work with Apps” feature, users must activate integration settings and control which applications ChatGPT can interact with. This step reflects OpenAI’s strategic push to create context-aware tools that enhance user productivity.
Unlike existing AI tools such as GitHub Copilot, which can write and edit code within integrated development environments (IDE), ChatGPT’s current function is limited to reading and analyzing code. The macOS accessibility API and VoiceOver screen reader underpin this new capability, allowing the tool to extract and process visible text.
Broader Implications for AI Development Tools
OpenAI’s efforts to make ChatGPT more interactive align with the increasing demand for AI models that integrate seamlessly with everyday developer tools. While the current iteration stops short of full code editing, OpenAI has hinted at expanding capabilities in future updates.
This positions ChatGPT within the competitive field of AI coding assistants, where Microsoft and Google have recently launched advanced AI tools with unique integrations.
|
Related: |
Microsoft and Google’s Competitive Edge
At GitHub Universe 2024, Microsoft rolled out GitHub Copilot for Azure in public preview. This feature allows developers to handle cloud deployments and troubleshoot directly within GitHub or Visual Studio Code using commands.
In addition, new AI-driven templates support rapid project prototyping, complete with infrastructure as code (IaC) and CI/CD pipelines, reducing the time needed for project setup and aligning with industry security standards.
Meanwhile, Google’s October release of Gemini Code Assist Enterprise added another layer to the evolving landscape of AI coding tools. The service, integrated into popular IDEs such as JetBrains and VS Code, goes beyond code suggestions by adapting to the developer’s specific project context.
With capabilities spanning SQL and Python assistance in BigQuery and connections to Google Cloud services, Gemini offers more than standard AI completion tools.
Alibaba’s Open-Source Approach
Adding to the global competitive mix, Alibaba launched its Qwen2.5-Coder series earlier this month. These open-source models, which range from 0.5 billion to 32 billion parameters, have drawn over 250,000 downloads within the first 48 hours. Supporting more than 40 languages, the Qwen2.5-Coder models cater to varied development needs, from code generation to debugging.
Notably, Alibaba’s release under the Apache 2.0 license makes these tools freely adaptable, appealing to budget-conscious startups and independent developers seeking robust solutions without costly subscriptions.
Potential for Expanded Features in ChatGPT
The release of “Work with Apps” lays the groundwork for OpenAI’s broader vision of creating agent-like AI systems that go beyond prompt-based interactions. While the current version focuses on reading and contextual responses, future iterations may extend to direct coding assistance. OpenAI’s gradual feature expansion keeps it in step with Microsoft, Google, and Alibaba, each pushing the boundaries of what AI-assisted coding can achieve.
