Core Viewpoint - Midjourney has entered the video generation space, showcasing impressive capabilities in creating realistic animations and scenes, sparking significant interest and discussion among users [1][5][6]. Group 1: Video Generation Capabilities - The video generation model demonstrates smooth transitions in actions and environments, with realistic details such as reflections [2][3]. - Users have noted the high level of realism, with some stating that the videos are indistinguishable from real-life footage [9]. - Despite the impressive visual quality, the model currently lacks audio functionality, which has led to questions about its timeliness in entering the market [28][31]. Group 2: Image Generation Model Updates - Midjourney's image model, V7, is continuously being updated, with significant improvements in texture detail and rendering speed [10][41]. - The introduction of features like "draft mode" allows users to generate images through voice commands, enhancing user interaction and reducing generation costs by half [44][48]. - The V7 model has seen a 40% increase in image generation speed, with rendering times significantly reduced [51][52]. Group 3: User Engagement and Feedback - Midjourney has actively encouraged user participation in image scoring to refine the V7 model, indicating a commitment to user-driven development [38]. - The company has expressed a desire for user feedback on pricing to ensure accessibility for a wider audience [35]. Group 4: Competitive Landscape - The entry of Midjourney into video generation raises questions about its competitive position, especially compared to existing models like Veo 3, which already offer audio capabilities [28][31]. - Midjourney's focus on animation style may differentiate it from competitors that prioritize realistic video generation [34].

豆包前段时间新的图片模型的实力大家应该也看到了。 强大的提示词理解加上字体和营销图片生成能力直接让人人都能生成自己需要的营销图片或者进行字体设计。 就在前天 豆包又更新了超能创意 1.0 模式 ，我被灰度到了试了一下，给我整麻了。 图片的生成效率和修改效率大幅提升，让本来就很低的设计门槛又低了一大截。 我们可以先看个例子再介绍 我输入的提示词为： 参考下面的提示词帮我生成十个其他知名品牌的胶囊 16:9 图片，先基于品牌和主营业务更改提示词 中的内容然后在生成。 示例提示词为：一个高高的、外观逼真且充满活力的胶囊体水平漂浮着。它的左半边是标志性的星 巴克绿色，标有"Starbucks – Uplifting the Everyday"字样以及经典的美人鱼（Siren）标志。右半 边是透明的，里面填充着漂浮的烘焙咖啡豆、细腻的奶泡漩涡、手绘咖啡杯图标以及代表社区连接 的抽象暖色调线条，需要有背景色。 来看看他给我的结果，我根本没提要哪些品牌，也没提这些品牌的主营业务和典型产品。 他直接从LLM 模型拿到了这些知识然后还按照要求改了提示词 ，太离谱了，而且 这十张图片的生成速度比 4 o 一张都要快很多 。 我测 ...