Elon Musk calls the Grok Agent model a 'significant unlocking of capabilities', with attention paid to AI video generation capabilities According to BlockBeats, on May 19th, Musk forwarded a user tweet and commented that "Grok agent mode is a major ability unlock". The statement cites a user's presentation, in which the user stated that they have completed a video production of more than 4 minutes using Grok Imagine's Agent mode, and claimed that this mode significantly improves the efficiency of the video generation process, allowing users to directly rely on AI understanding to complete script, image, and background generation without the need for additional dubbing processing. The user also stated that the image reference function is now available and believes that it performs well in terms of consistency and realism. They plan to release it as a video later. Musk's repost can be understood as a strengthened endorsement of xAI's multimodal "Agent based creation tool", indicating that it is pushing Grok to extend from a chat model to an executable content production system. [Original link]