OpenAI leak suggests new ChatGPT capabilities

OpenAI is gearing up to introduce several significant enhancements to ChatGPT's web interface, with internal testing already underway. These updates, expected to roll out over the coming weeks, promise to transform the platform from a simple chatbot into a more versatile and integrated tool for various tasks. The changes include a new feature codenamed 'Salute', smarter model selection for local queries, enhanced security for enterprise users, and improved inline editing for technical content.

Expanding ChatGPT's Capabilities with 'Salute'

One of the most notable additions is 'Salute', a feature designed to streamline task management directly within ChatGPT. This allows users to create specific tasks, attach relevant files, and then monitor their progress within the application. Essentially, OpenAI is pushing ChatGPT beyond conversational AI, moving it towards a lightweight project or workflow management system. Imagine using ChatGPT not just to draft an email, but to manage that email's attachments and track its completion status. This could be particularly useful for small businesses or individuals needing a more organised approach to their digital interactions.

Smarter Local Search and Enterprise Security

OpenAI appears to be refining ChatGPT's ability to handle location-specific queries, with references to an "is model preferred" flag. This suggests the platform will soon automatically select a specialised AI model optimised for local businesses, restaurants, and hotels. For users, this means more accurate and relevant results when searching for information about nearby services, potentially integrating better with mapping widgets and local directories. This move aligns with the broader trend of AI becoming contextually aware, as seen with developments like Gmail Gets Gemini AI Boost for email.

For enterprise clients, a new 'Secure Tunnel' feature for MCP (Model Context Protocol) servers is on the horizon. This aims to simplify secure integration by connecting internal MCP servers to OpenAI via a customer-hosted tunnel client using outbound-only HTTPS. This eliminates the need for inbound firewall changes, making it easier and safer for businesses to integrate ChatGPT into their existing infrastructure. Security remains a paramount concern for AI adoption, and features like this are crucial for fostering trust, especially given ongoing debates around data privacy and AI, as highlighted by organisations like the National Cyber Security Centre.

Enhanced Editing for Technical Content

A significant quality-of-life improvement for developers and technical users will be the introduction of inline editable code and math blocks. These function much like the existing rich-text formatting blocks, which users might recognise from tools like the email or writing draft areas in ChatGPT. When you highlight text, a mini editor toolbar pops up. Now, this functionality will extend to code snippets and mathematical expressions, making it far easier to edit and refine technical content directly within the chat interface. This could reduce the need to switch between applications and improve the overall flow for tasks involving programming or complex calculations. It's a step towards making ChatGPT a more comprehensive environment for various forms of content creation, echoing the utility seen in tools like Lost in AI Chat? Find What Matters With ChatGPT Pins for organising conversations.

These updates collectively indicate OpenAI's strategy to evolve ChatGPT into a more robust and versatile platform, catering to a wider range of user needs, from personal productivity to enterprise solutions.

What do you think of these upcoming changes? Will 'Salute' genuinely change how you use ChatGPT for tasks? Share your predictions in the comments below.