The world of artificial intelligence (AI) is constantly evolving at an incredible pace, bringing us new tools, improvements to existing systems, and fascinating discussions about AI’s impact on our lives. The past seven days have brought numerous significant updates, from Google’s innovations to speculations about new models and controversies within the industry.
Google’s Dominance in Innovation
Google has been particularly active this week, introducing several key updates and new tools:
- Opal – AI Mini Apps at Your Fingertips: Google Labs introduced #Opal, a tool that allows you to describe, create, and share AI mini-apps. It is currently available in the US during its public beta phase and enables the creation of small, AI-powered workflows. You can use pre-built applications like a blog post writer, a book recommendation tool, or a photo transformer into “claimation” characters, and then remix them as desired. You can also start from scratch and describe what you want the app to do, and the AI will build the workflow itself, which can then be manually adjusted. For example, you can ask for an app that researches AI news from the past 24 hours and creates a page with bullet points and sources.
- Web Guide – AI-Organized Search Results: Google is also testing changes in the appearance of search results with a new feature called Web Guide, an experimental AI-organized search results page. This feature groups search results into categories, such as “comprehensive guides for solo travelers” for queries about traveling to Japan. Web Guide is rolling out gradually and needs to be enabled within Google Labs.
- Photos to Videos and Remixing: Google Photos is rolling out an update that allows you to turn your own photos into videos using the “I’m feeling lucky” function. This feature is starting to roll out in the US on Android and iOS devices. Concurrently, remixing features are being introduced directly into Google Photos, where you can use AI to restyle photos in different styles, such as cartoon or anime. The same photo-to-video feature is also being added to the YouTube app, allowing users to convert images into videos directly within the app and post them to YouTube Shorts.
AI Video and Image Tools
- Leonardo V3 Fast – Faster and Cheaper Video: In the world of AI video, Leonardo introduced V3 Fast, which is three times faster than the original V3. It still generates 8-second videos with sound, but it’s faster and cheaper to use, costing 2,000 credits compared to 2,500 for V3, with a minimal difference in quality.
- Higsfield Steel – Recreating Images from the Web: Higsfield AI introduced Higsfield Steel, a feature that allows you to recreate any image from the web via a Chrome extension. The installation process requires manual steps and enabling developer mode, suggesting caution when using.
- Pika Social AI Video – Social Media with AI-Generated Content: Pika Labs is introducing a new app, Pika Social AI video, which is currently invite-only. It allows users to take selfies, and Pika’s AI then animates these images or places them in different scenes, creating a social media feed of AI-generated versions of friends and family.
Privacy and Developments in AI
- Lumo – Proton Mail’s Confidential Chatbot: Proton, known for its Proton Email service, introduced its own chatbot called #Lumo, which is a confidential chatbot. Lumo does not collect data, does not record conversations, and conversations are never used for model training, ensuring complete privacy. It appears to be free to use, but also offers paid plans with more options.
- Warp – AI-Powered Development Environment: #Warp is a new agentic development environment that serves as an AI-powered assistant for development. It allows users to describe what they want to build in natural language, and the AI then performs tasks such as spinning up environments, writing and refactoring code, debugging issues, and even deploying, all from a single input bar.
Industry News and Controversies
- Meta Acquires AI Talent: Meta continues its tactic of acquiring talented individuals from other companies. This week, Meta brought over three Google AI researchers who worked on a model that solved problems in the International Mathematical Olympiad at a gold medal level.
- Amazon Buys B AI Wearable: Amazon has acquired B AI wearable, a wristband that records conversations by transcribing them, allowing you to ask questions about recorded sessions later. The wristband does not save audio recordings and is anonymous.
OpenAI Updates
- GPT-5 – Release Date Rumors: It is rumored that #GPT5 is coming out in August. Although Sam Altman did not provide an exact date, only mentioning “soon” and “in the summer”. The next OpenAI Dev Day is on October 6th in San Francisco, and this year attendance costs $650.
- ChatGPT Agent Available to All: The ChatGPT agent is now fully rolled out to Plus, Pro, and Team users. It allows users, for example, to search X accounts for interesting AI news and tutorials and organize information into spreadsheets.
- Gold Medals at Math Olympiad: OpenAI and Google DeepMind have achieved significant successes, claiming their models achieved gold medal level performance in the International Mathematical Olympiad. This demonstrates an improvement in “general artificial intelligence” in “narrow skills,” bringing us closer to AGI (Artificial General Intelligence) or ASI (Artificial Superintelligence).
- DeepMind Anas – Contextualizing Ancient Inscriptions: DeepMind presented research called Anas, a model that contextualizes ancient inscriptions, helps historians interpret them, and can restore fragmented texts.
Microsoft Windows and Other Developments
- Windows 11 AI Features: Windows 11 received a number of new features, including Recall (full-computer search history), Click-to-do (additional AI features on right-click), improved Windows search, and Copilot Vision (an interactive tutorial where AI can guide the user step-by-step).
- Alibaba Qwen Model: Alibaba has a new model called Qwen 3235B A22B2507, which outperforms the previous Qwen 3235B A22B model in benchmarks.
This week once again demonstrated the extraordinary progress and diverse applications of artificial intelligence, from content creation tools and improved user experience, to ethical questions and strategic policy guidelines. We will continue to monitor how these innovations develop and impact our future.
More news from Ai World can be found on our Blog.
Stay updated on key AI innovations!