生成式AI 一、Runway发布全新Gen 4.5模型，57.1%的人分不清真假 1. Runway发布全新Gen 4.5图生视频模型，在镜头控制和故事叙事能力上实现显著提升，能在5秒内快速生成包含近 景、中景、远景的三个镜头； 2. 在1000人参与的测试中，仅有57%的人能分辨AI生成视频与真实视频，模型在人物面部一致性、光影逻辑和物理 规律表现上接近电影级水准； 3. 视频生成模型正进入新一轮升级期，真实度、声画同步、局部控制精细化和更长生成时长成为行业共同趋势。 https://mp.weixin.qq.com/s/_CryhmDwnF9C3O-6LM2-iQ 二、谷歌Gemini变身免费家教，接入SAT全真模考，讲错题 1. 谷歌联手The Princeton Review将全套SAT模拟题整合进Gemini，用户可免费进行全真模考，分数立等可取并获 得详细错题解析； 2. 测试涵盖阅读写作和数学两大模块，支持自定义倒计时和提示功能，Gemini会把解题思路拆解成详细步骤辅助理 解； 3. SAT只是第一步，谷歌计划将Gemini逐步扩展到更多标准化考试，同时通过垂直领域渗透策略让AI成为各行业的 ...

再看看这个，湿漉逼真的头发、肉眼可见的面部雀斑、超自然的景深，有点好莱坞大片内味儿了嗷： 还没完，咱再来看这个，机械义体与人脸的融合，以及构图处理都非常到位，妥妥滴赛博大片即视感！ 梦瑶 发自 凹非寺 量子位 | 公众号 QbitAI 不er，这个世界还有什么是真的？反正我是已经分不清了... 短短3秒，连续切了3个镜头，从人物脸上的皮肤纹理，到满天纷飞的大雪，细节真实到有点离谱！ 你就说逼真不逼真吧… 不卖关子，就是Runway刚刚发布的——全新 「 Gen 4.5」 模型。 这次更新主打的是 图生视频 ，在 镜头控制 和 故事叙事 上，明显往next level推了一步～ 这波效果一出来，网友当场坐不住了，直呼：感觉都能吊打好莱坞制作团队了好吧？太逼真！（doge） 甚至逼真到，在Runway做的一项1000人参与的调查中，结果只有约 一半 的人能分辨出该模型和真实视频的区别…… 问题来了，那这超超超逼真的——Gen 4.5模型效果到底咋样？咱一起来看！ 长故事表达能力 ：能承载更长时序的内容结构，视频的完整度和长度大大提升。 精准的镜头控制能力 ：镜头的景别、角度、运动轨迹以及切换节奏都更可控，输出效果 ...

Group 1 - Keling AI officially launched its new product "Keling O1," which integrates multi-modal inputs such as text, video, images, and subjects into a comprehensive engine, addressing consistency issues in AI video generation for applications in film, self-media, and e-commerce [2] - OpenAI is reportedly considering embedding advertisements in ChatGPT, with recent Android test versions containing code labeled as "featured ads," indicating a shift towards personalized advertising based on user interactions [2] - ByteDance's Seed team released GR-RL, achieving a significant improvement in the success rate of a shoe-lacing task from 45.7% to 83.3%, marking a notable advancement in reinforcement learning for fine manipulation tasks [2] Group 2 - AI startup Runway introduced its latest film generation model Gen 4.5, which outperformed Google and OpenAI in third-party evaluations, showcasing its ability to generate high-quality videos based on textual instructions [3]

Group 1 - Lenovo ISG in Shanghai reportedly laid off hundreds of employees, including pregnant women, with a brief communication meeting lasting only 15 minutes [3] - Nestlé is considering selling Blue Bottle Coffee, with an expected valuation below $700 million, as part of a broader strategy to streamline its business [11] - Instagram will require U.S. employees to return to the office five days a week starting February 2, 2024, following a previous policy requiring at least three days [11] Group 2 - OpenAI CEO announced a "red alert" status to improve ChatGPT and plans to delay advertising business projects [6] - Michael Burry disclosed that he is shorting Tesla stock, citing its "absurd" valuation [11] - Runway launched its latest video generation model Gen 4.5, which outperformed Google and OpenAI in third-party evaluations [20] Group 3 - Samsung launched its first tri-fold smartphone, Galaxy Z TriFold, priced at 3.59 million KRW, set to release in South Korea on December 12 [16] - The Chinese heavy truck market saw a nearly 50% year-on-year increase in sales in November, with total sales expected to exceed 1.1 million units for the year [20] - Microsoft is increasing its AI investment in Europe, focusing on establishing local facilities rather than U.S.-based operations [13]

Core Insights - Runway has launched Gen 4.5, a new video model that surpasses similar offerings from Google and OpenAI in independent benchmarks [1][2] - The model excels in generating high-definition videos from written prompts, demonstrating strong understanding of physics, human motion, camera movements, and cause and effect [1] - Runway's Gen 4.5 currently ranks first on the Video Arena leaderboard, outperforming Google's Veo 3 in second place and OpenAI's Sora 2 Pro in seventh place [2] Company Performance - Runway's CEO Cristóbal Valenzuela highlighted the achievement of competing against trillion-dollar companies with a relatively small team of 100 people, emphasizing focus and diligence as key factors for success [3]