- AI KATANA
- Posts
- OpenAI launches Operator AI agent research preview
OpenAI launches Operator AI agent research preview
Also: China’s cheap, open AI model DeepSeek thrills scientists

Hello! The ever-evolving world of AI is making waves across industries, with OpenAI unveiling a powerful browser-based agent that can tackle complex tasks independently, and China introducing a revolutionary open-source AI model challenging global leaders with its affordability and transparency. On the energy front, discussions around powering AI’s massive data centers are intensifying, with nuclear energy emerging as a potential solution to meet growing demands. Meanwhile, Samsung’s latest flagship smartphone showcases how AI-driven innovation continues to enhance user experience, and political shifts in the U.S. spotlight the delicate balance between AI innovation and governance.
Sliced just for you:
🤖 OpenAI launches Operator AI agent research preview
🇨🇳 China’s cheap, open AI model DeepSeek thrills scientists
⚡ Is nuclear energy the answer to AI data centers’ power consumption?
📱 Samsung Galaxy S25 Ultra hands-on: AI-driven features lead the way
🇺🇸 Trump orders AI action plan and more work erasing Biden’s AI efforts
OpenAI has introduced “Operator,” a browser-based AI agent capable of independently performing tasks online, such as filling out forms, ordering groceries, and creating memes. Built on the Computer-Using Agent (CUA) model, it combines GPT-4o’s visual recognition with reinforcement learning to interact with graphical interfaces. Designed for efficiency and accessibility, Operator allows users to manage workflows with personalized instructions and task parallelization, all while maintaining control over sensitive actions like payments or logins. It is available as a research preview for Pro users in the U.S., with plans for broader integration into ChatGPT. Safety is prioritized through features like user confirmations, privacy management, and defenses against malicious inputs. Operator is poised to enhance productivity for individuals and businesses, while also exploring public sector applications. OpenAI plans to expand its functionality and access based on real-world feedback, with the eventual goal of releasing CUA as an API for developers.
China’s DeepSeek-R1, a newly developed AI model, is making waves as a cost-effective and open alternative to leading reasoning AI systems like OpenAI’s o1. Capable of performing on par with o1 in tasks related to chemistry, mathematics, and coding, DeepSeek-R1 stands out for its affordability and openness, as researchers can freely examine and build on it under an MIT license. This model demonstrates that efficiency can rival raw computational power, a significant achievement given U.S. export restrictions on advanced AI hardware. With running costs approximately thirty times lower than competitors and the availability of smaller distilled versions for limited computing resources, DeepSeek is poised to democratize AI applications globally. Its emergence underscores a narrowing technological gap between global AI leaders and highlights the potential for collaborative innovation over competitive arms races in AI development.
AI data centers are projected to increase energy demand by more than 160% by 2030, driving exploration of sustainable power solutions. Nuclear energy, with its low carbon emissions and reliability, is emerging as a key option for meeting these demands, particularly for 24/7 operations. However, challenges like high costs, permitting hurdles, and limited uranium availability hinder rapid adoption. Renewables such as wind and solar are also critical, capable of covering up to 80% of demand when paired with storage, but their intermittency necessitates complementary sources like nuclear or natural gas. Companies are adopting hybrid strategies that combine renewables, battery storage, and traditional power to ensure consistent energy. While efficiency improvements in data centers have slowed since 2020, continued advancements and investments in nuclear and green energy are expected to significantly reduce emissions and support the burgeoning energy needs of AI infrastructure.
The Samsung Galaxy S25 Ultra distinguishes itself in the competitive smartphone market with significant AI-driven innovations and refined hardware. Building on the Galaxy S24 Ultra, the device introduces advanced AI features such as enhanced contextual photography and improved gallery search, aimed at simplifying everyday tasks through intuitive automation. Powered by the exclusive Snapdragon 8 Elite processor, it balances performance and energy efficiency while maintaining a 5,000 mAh battery. The phone’s design, featuring a thinner and lighter build with a squarer-edge aesthetic, accommodates an enlarged display for a more immersive experience. Launched alongside the S25 and S25+ models at Samsung's Unpacked 2025 event, the Galaxy S25 Ultra underscores Samsung’s commitment to iterative improvement, leveraging AI to enhance user experience and cement its flagship status.
U.S. President Donald Trump has issued an executive order to establish an AI Action Plan aimed at solidifying America’s global leadership in AI, emphasizing economic competitiveness, national security, and technological advancement. The plan has a 180-day timeline for development and seeks to enhance AI innovation while dismantling regulations implemented under former President Joe Biden. Among Biden’s policies revoked was a 2023 executive order requiring AI developers to disclose safety assessments for high-risk systems. The move signals a shift away from prior safety-focused AI governance, prioritizing growth and dominance in the AI sector. Critics argue the approach may sideline essential safeguards, while proponents highlight its potential to accelerate AI-driven advancements.
🛠️ AI tools updates
Researchers at Yale School of Medicine’s Cardiovascular Data Science Lab have developed an AI tool capable of identifying individuals at high risk of developing heart failure through electrocardiogram (ECG) images. This innovation represents a major advancement in early detection, offering a noninvasive, widely available method to stratify risk before symptoms appear. Validated across diverse populations in the U.S., U.K., and Brazil, the model could reduce hospitalizations and premature deaths by enabling timely intervention. By integrating this AI tool into routine ECG screenings, clinicians can significantly improve cardiovascular outcomes, especially in resource-limited settings. The study highlights the global scalability of AI-driven health technologies and was published in the European Heart Journal.
💵 Venture Capital updates
In 2024, Andreessen Horowitz emerged as the most active AI investor, completing 42 post-seed deals, including major funding rounds such as Databricks’ $10 billion raise and xAI’s $6 billion funding. Other top investors included Lightspeed Venture Partners with 37 deals and Y Combinator with 36, followed by Alumni Ventures, Sequoia Capital, and General Catalyst. Thrive Capital led the highest-value rounds, totaling $17.3 billion, including investments in Databricks and OpenAI. The year saw significant increases in AI-related investments across the board, with VC firms focusing on major bets in a booming sector. Despite high activity, the need for massive exits underscores the risks inherent in these substantial investments.
🫡 Meme of the day

⭐️ Generative AI image of the day

Before you go, check out Pope to World Economic Forum: AI must promote and never violate human dignity.
