Figure AI's humanoid robot just raised $1B

9to5Google

NotebookLM's Video Overviews now support multiple visual styles via Google's Nano Banana model and offer Brief or Explainer formats, enhancing how users visualize and understand their source materials.

Philipp Schmid Blog

Deep Agents represent a fundamental architectural shift from reactive LLM loops to proactive systems with planning, delegation, and external memory, enabling AI to handle complex multi-day tasks instead of simple transactional queries.

TIME

Figure AI's Figure 03 humanoid robot represents a major step toward general-purpose domestic robots, leveraging advances in AI and massive data collection to train machines for real-world unpredictability, though critical safety and practical challenges remain unsolved before widespread home deployment.

Import AI

Advanced AI systems are growing in capability and displaying concerning signs of autonomy, making both technological optimism and appropriate caution necessary; society must shift from elite conversations to broad public engagement to develop effective governance before major disruptions occur.

Ming-UniVision unifies vision understanding and generation in a single continuous latent space, achieving 3.5× faster training convergence and eliminating inefficient pixel round-trips in multi-turn vision tasks. This breakthrough enables more efficient and intuitive multimodal AI systems.