ChatGPT Adds Library Tab, Google Overhauls AI Studio

ChatGPT Gets a Library for Your Documents

The way we work with artificial intelligence is changing. Many of us are now drafting more documents, from Word files to Excel spreadsheets, using AI tools. Recognizing this, ChatGPT has introduced a new feature: a Library tab. This isn’t just about making files easier to find. It signals a shift in how AI models handle information, especially managing context.

The Library tab collects files you’ve recently used with ChatGPT. This includes screenshots and documents like personal finance trackers. While it doesn’t show older files, it makes accessing and referencing them much simpler. You can now start a new chat directly from a file in your library. Imagine adding important context files, like a personal DNA file detailing your identity and goals, to this library. This makes it easy for ChatGPT to refer to them whenever you need.

The interface for viewing files has also improved. You can now see formatted markdown files and even chat directly within them. This move away from just text-based interactions to handling various document types like Excel sheets and infographics shows where AI is heading. The advice is clear: think about which documents can provide context for your AI conversations and tasks.

Google AI Studio Merges with Agentic App Builder

Google has significantly updated its AI Studio, merging it with its agentic app builder, ‘Antigravity.’ Previously, Google AI Studio was a developer-focused platform for Gemini, but many non-technical users also found it helpful for creating apps. Antigravity, on the other hand, was known for its autonomous capabilities.

The new, combined platform is designed to be more user-friendly. It now includes impressive demos of technically complex features, such as multiplayer games using 3D particles. A key addition is the integration of databases and authentication, allowing for the creation of shared tools and games for multiple users. This is a significant step, as building such applications was previously very difficult with AI tools.

Google has also integrated Firebase, a service that provides databases and user account management for apps. The platform now offers a simple way to build and even publish live online applications. Users can explore example applications, remix them, and then publish their own creations. While hosting and using these apps will incur costs, the ease of creating multiplayer experiences is a major advancement.

Midjourney V8 Shows Mixed Results

Midjourney, once the leading AI image generator, has released its V8 version in alpha. However, early user feedback and tests suggest it may not be a significant step forward. Many users are finding that prompt adherence is weak, with the AI sometimes generating incorrect subjects, like a DC superhero when asked for a Marvel one.

While the composition of images might be more cinematic, the overall impression is that V8 is more of a variation than a substantial improvement. This mirrors a trend seen in many AI image generators, which have reached a high level of quality and are now exploring different artistic styles rather than just better accuracy.

Anthropic Study Reveals Hopes and Fears for AI

A study by Anthropic, based on interviews with users, offers insights into what people hope for and worry about regarding AI. A recurring theme is that AI amplifies existing human traits – it makes us more of what we already are.

Users expressed strong hopes for personal transformation and life management through AI, with time freedom also being a significant desire. On the professional side, AI is seen as a powerful productivity tool and a cognitive partner for brainstorming. However, the biggest concern cited by users is AI’s unreliability. The need for a human in the loop remains crucial because AI tools can sometimes produce unexpected or incorrect results, making them difficult to depend on consistently.

Quick AI Updates: Gamma, OpenAI, Claude, Gemini, and More

Gamma Imagine: The presentation creator Gamma is expanding into graphic design with its new ‘Imagine’ feature, aiming to become a more comprehensive creative tool.

OpenAI Mini Models: OpenAI has released smaller models like GPT 5.4 Mini and Nano. These are primarily for developers building AI agents, offering quicker and cheaper performance for specific tasks.

Figma AI Agents: Figma has introduced AI agents, entering a competitive space that includes recent releases like Stitch.

Claude Code Auto Mode: Claude Code has rolled out an ‘auto mode.’ This feature is a safer version of a previous command that allowed the AI to act autonomously. Auto mode includes a classifier to check for potentially risky actions, asking for permission when needed, but otherwise proceeding to complete tasks efficiently.

Gemini Personal Intelligence: Google is now bringing its ‘personal intelligence’ features to free Gemini accounts. This aims to provide a more personalized experience by automatically gathering and using user context.

FAI MCP Server: For those interested in integrating external applications with AI tools like ChatGPT, the FAI MCP server offers an easier way to connect to various image and video generation models, simplifying AI-powered content creation.

Source: New ChatGPT Library Explained & More AI News You Can Use (YouTube)