Rafomac News, Tech & Trends That MatterGoogle has just upgraded its Omni video generation model with the 1.1 release, bringing more control and higher resolutions. Meanwhile, Anthropic’s Claude Co-work AI now features its own built-in browser, shaking up the way we interact with intelligent agents.
Google’s Omni 1.1 Promises Smarter Video Creation
Google has quietly launched Omni 1.1 Flash, a significant update to their video generation AI introduced just a few months ago. Unlike the original, this version lets users extend scenes by specifying start and end frames, use videos as visual references, and upscale output to a crisp 4K resolution. There’s also a faster 360p mode designed for quick concept testing that’s lighter on credits.
Using Google Flow — a dedicated dashboard for generating images and videos — users can now organize projects and produce multiple images or videos from a single prompt to compare outcomes.
One hands-on example saw a medieval fantasy scene brought to life, with travelers stepping through an ancient portal that gradually lights up and dims. While the model delivered impressive consistency between frames, it sometimes subtly added unrequested characters or swapped actions, like random figures activating the portal or switching places mid-scene. Despite these quirks, Omni 1.1 delivers a remarkable step forward in video AI.
The model isn’t perfect — stubborn prompts will occasionally override clear instructions — but the ability to tightly control frame sequences is a game changer for creators.
Anthropic’s Claude Co-work Introduces Its Own Browser
Anthropic’s Claude Co-work, another serious AI contender, has just gained a built-in browser much like the one in ChatGPT Work. This addition enables Claude to perform web tasks with integrated browsing on the right side of the screen, offering users the option to take control anytime. This feature is gradually rolling out and promises to make Claude more practical and versatile for real-world web interactions.
Alongside this is a subtle makeover in the Claude desktop app’s interface. Tabs have been resized, project selectors tweaked and approval settings refined. It’s a nudge towards streamlined usability, even if the changes are barely noticeable.
Claude’s Unified Memory Across Platforms
One of Claude’s most interesting new features is a unified memory system spanning the chat and co-work applications. This approach lets users switch between platforms smoothly, with conversations and contexts remembered automatically. Sensitive topics like religion or health are excluded by default but can be enabled manually.
Categories organize memories—ranging from user interests to personal details—allowing easy review, editing, or deletion at any time. The AI builds this knowledge continuously during chats rather than waiting for manual input.
OpenAI’s ChatGPT Adds Sticker Packs and Transparent Images
On the consumer side, OpenAI has introduced a playful new feature in the ChatGPT app: creating custom sticker packs from your images, readily compatible with iMessage and WhatsApp. Users can select styles, remix photos, and generate adorable, meme-style stickers, complete with transparent backgrounds—a feature also now available for general images. This transparency is ideal for website design or overlay graphics, enhancing creative freedom.
Challenges with ChatGPT’s Cloud Browser; Desktop Version Gets Upgrade
While ChatGPT’s new cloud-based browser can navigate sites and log in securely without exposing user passwords, it struggles with human verification steps on many websites. In contrast, the desktop app’s on-device browser offers a smoother, more reliable experience, supporting full site access without the quirks of cloud-based restrictions.
Task Triggers and Temporary Chats Make Productivity Smoother
ChatGPT Work on web has improved triggered tasks, allowing users to automate workflows based on specific event triggers, like incoming emails or Slack messages, without constant polling. Regular users can now create up to three scheduled tasks for free, enhancing productivity.
Temporary chats have also evolved. Users can choose between personalized sessions that remember user settings and plugins, or unpersonalized ones that treat each session as new. Chats started temporarily can be saved anytime, turning them into regular, history-retained conversations.
Hugging Face Debuts Open-Source Micro Duck Robot
In a blend of fun and tech, the AI community Hugging Face launched the Micro Duck — a small, open-source robot priced at $399. It walks, learns from you, grabs objects, and even skates, all while being highly customizable for creators eager to tinker. It bridges robotics and AI in an accessible form, making advanced tech playful and approachable.
Google’s Gemini Live Upgrade Brings True Agent Intelligence to Phones
Google’s Gemini Live, recently beefed up with agentic functionalities, puts powerful AI assistants right on your phone. It integrates smoothly with Gmail, Drive, Google Photos, and Workspace, handling activities like reading emails, managing tasks, and processing images. It can even interact with videos — for instance, highlighting how to jumpstart a car in real-time instructions.
The Gemini Notebook now supports chatting with ebooks purchased via Google Play, enriching reading with interactive queries, infographics, and deep dives. Gemini’s new speech-to-text model promises improved real-time transcription that self-corrects as you pause, powering new creative and productivity tools.
Emerging Tools: Perplexity’s Portable Computer and SpaceX’s Grockbot
Outside the big players, Perplexity introduced a “portable computer” AI designed to run primarily on local models with a fallback to cloud when needed. This caters to users wanting privacy and efficiency but requires an expensive Nvidia DGX Spark setup, keeping it out of casual reach for now.
SpaceX AI cut prices on Grockbot, an autonomous AI agent, from $200 to $20 monthly, making advanced robotic assistance far more accessible. These developments suggest a growing diversity in how AI tools adapt to users’ needs and budgets.
Overall, the Claude Co-work browser stands out as a major practical step forward, but don’t overlook the simple joy of ChatGPT stickers or the impressive utility of Gemini Live. These advances combine to make AI less arcane and more human-friendly every week.
If you’re eager to see this wave of AI tech in action, exploring Google Flow’s new video control or creating stickers in ChatGPT might be the best place to begin.
