Figure, a U.S. robotics company, has achieved a breakthrough in humanoid robot technology, enabling its Figure 02 robot to walk with a more natural gait. This advancement, utilizing reinforcement learning, enhances robots’ mobility in real-world environments, paving the way for practical applications in industries like manufacturing and warehousing.
Aparna Thakur’s research introduces a transformative framework that combines AI with quantum-resistant cryptography to enhance financial transaction security. This innovative approach addresses vulnerabilities posed by quantum computing, improves payment reconciliation efficiency, and strengthens fraud detection, paving the way for a secure and scalable future in global finance.
A new AI agent, Manus, developed by Chinese startup Monica, showcases its ability to handle complex tasks autonomously, from travel planning to financial analysis. Unlike traditional chatbots, Manus executes actions directly through web interfaces, raising the bar for AI capabilities and sparking comparisons to Western innovations like OpenAI’s Operator.
AI pioneers Andrew Barto and Richard Sutton have been awarded the 2025 Turing Award for their groundbreaking contributions to reinforcement learning. Their research has profoundly influenced AI, bridging connections between computer science and neuroscience, and shaping the future of intelligent systems and decision-making across various disciplines.
Explore the transformative landscape of artificial intelligence with a deep dive into four leading models: DeepSeek AI, ChatGPT, Gemini, and Perplexity AI. Each platform offers unique strengths, catering to diverse needs from coding to real-time information retrieval, reshaping how we work and innovate in today’s digital world.
At the 2025 Spring Festival Gala, Unitree Robotics showcased 16 humanoid robots, “Fuxi,” performing a captivating Yangko folk dance. This event highlighted China’s advancements in robotics, with Unitree’s innovative AI-driven technology enabling lifelike movements, aiming to establish robots as reliable household assistants and redefine the future of humanoid robotics.
OpenAI’s SWE-Lancer benchmark evaluates AI models against over 1,400 real-world software engineering tasks, revealing that while AI has made significant strides, it still lags behind human engineers. The best model earned just $400,000 of a potential $1 million, highlighting the challenges AI faces in practical coding scenarios.
