- Ainsider
- Posts
- Ainsider Weekly AI Newsletter vol. 51
Ainsider Weekly AI Newsletter vol. 51
Massive updates from Google Gemini | A lot of new AI updates: Perplexity, Cohere, Luma and others upgrades | New AI Tools in Ainsider Directory

Table of Contents
Massive Updates from Google Gemini
1. Deep Reasoning for Everyone in Gemini 2.0
Google has expanded Gemini 2.0's capabilities for free users by introducing deep reasoning features, previously available only to advanced users and developers.
Now, everyone can access the model’s enhanced analytical abilities, such as solving complex problems, data analysis, and logic-driven decision-making based on broad world knowledge. These improvements, powered by the 2.0 Flash Thinking Experimental technology, combine model speed with advanced reasoning, making Gemini a powerful tool for both professionals and casual users.
2. Native Image Editing in Gemini 2.0 Flash
Another breakthrough is the introduction of native image generation and editing in Gemini 2.0 Flash, available in Google AI Studio. Users can now create and modify images using multi-step natural language dialogues, eliminating the need for external models like Imagen. This feature enables iterative design—allowing users to generate initial visualizations and refine them with precision, opening new possibilities for content creators, designers, and artists.
3. Personalization Based on Google Search History
Gemini 2.0 introduces a personalization feature that allows AI to access your Google Search history to provide more relevant answers. For instance, if you ask for restaurant recommendations, Gemini can consider your previous food-related searches. This feature is experimental and powered by the Gemini 2.0 Flash Thinking model. You can disconnect Gemini from your search history at any time, ensuring full control over your data. It is currently available for Gemini and Gemini Advanced users on the web, with a mobile rollout coming soon.
4. Gems – Custom AI Assistants Now Available to Everyone
Gems are custom AI assistants that let users create personalized versions of Gemini for specific tasks, such as a fitness coach or a cooking assistant. Previously available only to Gemini Advanced subscribers, the latest Gemini 2.0 update has made them free for all users. This expanded access enhances the platform’s flexibility and usability for a broader audience.
5. Colab Integration – No-Code Data Analysis
Gemini in Colab notebooks now allows planning, analyzing, and visualizing data without writing code.
Users can simply describe their goals in the Gemini panel (e.g., “Visualize trends” or “Create a predictive model”), and the system generates a fully functional Colab notebook, eliminating the need to write and execute code manually.
Earlier this month, Google also announced Gemini integrations with apps like Google Photos, Calendar, and Tasks, aiming to enhance personalization and usability for end users.
The new features in Gemini 2.0—deep reasoning, image editing, and no-code data analysis—demonstrate how Google is continuously evolving its models to meet the needs of both individual creators and enterprises.
Perplexity AI Windows App Now Available
Perplexity AI has launched an official desktop app for Windows, providing users with access to key platform features, including:
Voice Search (Voice Dictation): Allows users to ask questions and generate responses using voice commands, enhancing convenience and efficiency.
Keyboard Shortcuts: Streamlines navigation and interaction with the app, which is particularly useful for power users.
Access to the Latest Models: The app integrates Perplexity’s newest AI models, ensuring cutting-edge search, content generation, and data analysis capabilities.
The app is available as a Progressive Web App (PWA) or a native Electron-based application, allowing users to install it on their Windows desktop with flexibility tailored to different user needs.
DeepWork - AI Agents platform by Convergence AI
DeepWork, launched by Convergence AI, represents a leap forward in AI-driven automation, enabling users to specify outcomes and let the system coordinate multiple AI agents to handle complex, multi-step workflows autonomously.
The technology builds on Convergence AI's earlier "AI Organizations" concept, introduced with Template Hub, signaling a broader industry trend toward integrating advanced AI systems into organizational structures.
It builds on their Proxy platform, used by over 100,000 users, and is available for pro users at a $20/month tier, with a free basic version.
DeepWork allows users to specify desired outcomes, and the system then plans and executes the necessary steps, handling tasks such as market research, brand analysis, and travel planning. It features advanced planning intelligence, seamless web interaction, and a natural language interface, making it user-friendly.