Your AI Weekly: Agents That Act, AI's Geopolitical Game, and a Math Breakthrough

Hi everyone,

This past week in artificial intelligence was defined by a powerful shift from theoretical potential to practical application. The developments were driven by three main currents: the rapid maturation of agentic AI that can act on your behalf; an escalating platform war as tech giants race to lock users into their ecosystems; and the emergence of a clear geopolitical fault line between the US and China over the future of AI.

This week was a clear signal that the future of AI will be shaped as much by corporate strategy and national policy as by technical innovation.

Let's break down the biggest stories.

The Agentic Revolution Arrives

AI is evolving from a passive tool into an autonomous collaborator capable of executing complex tasks. This is the new frontier.

• OpenAI's ChatGPT Agent Rolls Out: After a week-long delay, OpenAI completed the full rollout of its highly anticipated ChatGPT Agent to all subscribers. This system is a unified powerhouse, combining web Browse, deep research, and conversational AI to act as your personal taskmaster. It can create slide decks, buy ingredients online, or even be scheduled to update a spreadsheet every week. This is a fundamental architectural shift, establishing the Agent as the new standard for proactive digital assistants.

  https://www.instagram.com/p/DMj_nYey4Kn/

• Google's Opal Lets You Build "Mini-Apps" Without Code: Google introduced Opal, an experimental tool that lets you build simple applications just by describing what you want in natural language. The AI translates your description into a visual workflow, which you can then share with a link. This dramatically lowers the barrier to creating software, empowering a new class of "citizen developers" whose core skill is effective communication with an AI.

The Platform Wars Heat Up

The battle to own the AI ecosystem is accelerating, with every major player launching new tools to capture user loyalty.

• Microsoft Fortifies Its Windows Moat: Microsoft's latest Windows 11 update is a strategic move to drive a new hardware upgrade cycle. The most powerful new AI features - like using natural language to change system settings or an AI "Relight" tool in Photos - are explicitly reserved for new "Copilot+ PCs" equipped with specialized AI chips. This transforms Windows into a tiered experience where the best features are a hardware purchase away.

• Google's Triple Play: Alongside Opal, Google launched Web Guide in Search Labs, which intelligently organizes search results into thematic categories instead of just a wall of links. This enhances user experience without completely replacing external publisher links. Additionally,

  Google Photos now has AI features to animate any static photo into a short video clip or "Remix" it in various artistic styles.

• The Creative AI Explosion: The content creation space is on fire.

  Leonardo.Ai has integrated Google's Veo 3 Fast model for high-fidelity video generation.

  Higgsfield AI introduced "Higgsfield Steal," a feature that lets you capture the complete aesthetic of any image on the web and apply it to your own AI character. And

  Pika Labs launched the first AI-only social video app, where the content is inherently generative and "remixable" by other users.

A Geopolitical Fault Line Emerges

The global AI race now has two clear, competing philosophies, led by the world's two biggest powers.

• The U.S. Unveils its "AI Action Plan": The White House released a 28-page plan to cement U.S. leadership in AI, structured around three pillars: accelerating innovation, building infrastructure, and leading in diplomacy. Key actions include aggressive deregulation to fast-track data center construction and a strategy to export a "secure, full-stack AI export package" to allied nations to counter China's influence.

• Alibaba's Open-Source Gambit: In a direct strategic response, Alibaba released Qwen3-235B-A22B-Thinking-2507, a flagship open-source reasoning model that tops several benchmarks dominated by proprietary Western models. By releasing such a powerful model with a permissive license for commercial use, Alibaba is leveraging open-source as a form of soft power, providing a potent, non-American alternative for countries and companies wary of dependency on U.S. tech.

Rapid Fire: Raids, Research, and Real-World AI

• AI Cracks the Toughest Math Test: In a landmark achievement, AI models from both Google DeepMind and OpenAI performed at a gold-medal level in the International Math Olympiad, solving problems that require deep creativity and reasoning. This signals a shift from AI merely retrieving knowledge to generating novel solutions.

• Meta's Talent Raid on Google: Meta successfully recruited three prominent AI researchers from the Google DeepMind team that built the Gemini model. This highlights the escalating "talent wars," where acquiring elite human capital is seen as a way to leapfrog years of development.

• Amazon Bets on Wearable AI: Amazon acquired Bee, a startup that created a wearable wrist device for "ambient intelligence" that continuously records and transcribes conversations to generate summaries and to-do lists. This is a clear move to evolve Alexa into a proactive, always-on agent.

• Privacy-First AI: Proton, known for its encrypted services, launched Lumo, a privacy-focused AI assistant featuring zero-access encryption and a policy of never using user data for model training.

• The "Baby Grok" Pivot: After multiple controversies over its "edgy" Grok chatbot generating harmful content, Elon Musk's xAI announced it will develop "Baby Grok," a kid-friendly version with robust content filtering and parental controls.

The pace of change is staggering, and staying ahead means understanding not just the tools, but the strategies behind them.

To help you master this new era, the 2025 edition of my book, "Prompt DOT AI: The Art of Writing Generative AI Prompts," is out now! It's a completely updated playbook for becoming an "AI Orchestrator," commanding a powerful suite of specialized AI agents to create anything you can imagine. This edition includes new chapters on AI video, 3D model generation, JSON prompting, and advanced creative workflows.

Stop just using AI - start directing it. Grab your copy to learn prompting and generating impactful content with AI.

Stay creative,

Da Sachin Sharma