Ben Affleck (Hollywood star & Artists Equity CEO) talks about how he fine-tunes open video models by unfreezing weights and trained only the last cinematic layer so a film crew can hit real production standards. for context, Ben Affleck founded InterPositive in 2022, a 16-person AI shop for film post and Netflix bought it in March 2026 for $587 mn in cash. He needed that model because public video models were trained on his peers' films, and he did not think that was a real business. So InterPositive raised money, shot its own dataset for 8 months on a controlled stage, and used it only as late-stage training. Each new film then trains a private model on its own dailies, so the production keeps the footage and the learning. That is the product Netflix paid $587 million for. ---- From "Bloomberg Live" YouTube channel, (link in comment)
#视频
#视频
今日 4 条
Rohan Paul@rohanpaul_aiAI 评分6464
引用Rohan Paul@rohanpaul_ai
AYi@AYi_AInotes精选AI 评分6565
引用Rohan Paul@rohanpaul_aiBen Affleck (Hollywood star & Artists Equity CEO) talks about how he fine-tunes open video models by unfreezing weights and trained only the last cinematic layer so a film crew can hit real production standards. for context, Ben Affleck founded InterPositive in 2022, a 16-person AI shop for film post and Netflix bought it in March 2026 for $587 mn in cash. He needed that model because public video models were trained on his peers' films, and he did not think that was a real business. So InterPositive raised money, shot its own dataset for 8 months on a controlled stage, and used it only as late-stage training. Each new film then trains a private model on its own dailies, so the production keeps the footage and the learning. That is the product Netflix paid $587 million for. ---- From "Bloomberg Live" YouTube channel, (link in comment)
推荐理由:原文梳理了 Ben Affleck 用私有实拍数据微调开源视频模型的思路与产权闭环,读者可以借此对比公共模型与影视级生产的差距。
Rohan Paul@rohanpaul_aiAI 评分6464
引用Rohan Paul@rohanpaul_aiBen Affleck (Hollywood star & Artists Equity CEO) talks about how he fine-tunes open video models by unfreezing weights and trained only the last cinematic layer so a film crew can hit real production standards. for context, Ben Affleck founded InterPositive in 2022, a 16-person AI shop for film post and Netflix bought it in March 2026 for $587 mn in cash. He needed that model because public video models were trained on his peers' films, and he did not think that was a real business. So InterPositive raised money, shot its own dataset for 8 months on a controlled stage, and used it only as late-stage training. Each new film then trains a private model on its own dailies, so the production keeps the footage and the learning. That is the product Netflix paid $587 million for. ---- From "Bloomberg Live" YouTube channel, (link in comment)
Thomas Wolf@Thom_WolfAI 评分5353
引用Rohan Paul@rohanpaul_aiBen Affleck (Hollywood star & Artists Equity CEO) talks about how he fine-tunes open video models by unfreezing weights and trained only the last cinematic layer so a film crew can hit real production standards. for context, Ben Affleck founded InterPositive in 2022, a 16-person AI shop for film post and Netflix bought it in March 2026 for $587 mn in cash. He needed that model because public video models were trained on his peers' films, and he did not think that was a real business. So InterPositive raised money, shot its own dataset for 8 months on a controlled stage, and used it only as late-stage training. Each new film then trains a private model on its own dailies, so the production keeps the footage and the learning. That is the product Netflix paid $587 million for. ---- From "Bloomberg Live" YouTube channel, (link in comment)
Google Blog: AIAI 评分5151 独立导演 The Gifted 获 Future Vision XPRIZE 大奖,获 10 万美元奖金及 250 万美元制作资金
Google 与 XPRIZE、Range Media Partners 合办的 Future Vision XPRIZE 公布大奖,独立导演 Jeff Synthesized 凭借短片 The Gifted 从全球超 2500 部作品中胜出。
Josh Woodward@joshwoodwardAI 评分4242引用Google Flow@FlowbyGoogleMore than 25 million people are using Google Flow every month to dream up new ideas, create stories, and build cool things. Thank you. We’re continuing the 50 additional daily credits for all users. Dive in and keep creating.
Runway NewsAI 评分4040 Runway 详解实时视频生成的同步审核系统
Runway 为即将发布的实时视频生成模型设计了同步审核系统,采用 Zentropi 的 CoPE-B 模型,安全审查平均耗时不到 0.5 秒,可将违规内容暴露窗口压缩至半秒以内。CoPE-B 是基于 Gemma-4-26B-A4B-it 的 LoRA 适配器,该 MoE 模型总参数 25.2B、每次前向仅激活 3.8B,兼顾速度与准确率。实时模型输出将附带 C2PA 溯源信号。
Runway NewsAI 评分5353 Kinetix 团队加入 Runway,推进机器人世界模型研究
Runway 宣布巴黎 AI 研究实验室 Kinetix 团队加入 Runway。该团队在 3D 人体运动与物理约束视频生成方面的积累将用于 Runway 面向机器人学的世界模型研究,团队将加入 Runway 巴黎枢纽,公司同时在巴黎招聘研究与工程岗位。
Runway NewsAI 评分3939 Runway 谈企业视频生成下一阶段:模型授权与智能体执行
Runway 企业业务今年增长逾一倍,NRR 超过 300%,一家 Fortune 20 客户使用量增长 17 倍。其企业客户已覆盖 Amazon、Microsoft、Allstate、Adobe、Robinhood 等,欧洲占企业客户群 20% 以上,日本成为亚洲最大市场。
Runway NewsAI 评分4545 Populous 如何用 Runway 呈现全球最具标志性的场馆设计
全球设计公司 Populous 将 Runway 用于体育场馆概念设计阶段(Stages 0 到 2),生成能承载完整场馆碗形、航拍视角和数万座席规模的渲染图,而非用于生成实际设计方案。
Runway NewsAI 评分4343 Runway Characters 登上 SIGGRAPH 2026 实时直播舞台
Runway 的 Characters 入选 SIGGRAPH 2026 的 Real-Time Live! 环节,团队在数千名图形学专家面前现场演示了从单张图片实时生成可对话角色。
Runway NewsAI 评分5252 Runway 与 MIXI 达成企业级战略合作
Runway 宣布与日本科技公司 MIXI 达成企业级战略合作,MIXI 已在游戏、体育、生活方式和数字娱乐业务中部署 Runway。Monster Strike 繁中版视频制作从 21 天缩短到 3 天,品牌片从 20 天缩短到 7 天,角色动画视觉提案可缩短至 5 分钟。
Runway NewsAI 评分4949 Runway 在伦敦设立欧洲总部,专注通用世界模型研究
Runway 宣布在伦敦设立欧洲总部及新的通用世界模型研究枢纽,计划未来 18 个月向英国 AI 生态投资 1 亿美元,到 2028 年该数字将翻倍以上。过去十二个月 Runway 欧洲订阅销量增长 50%,超 20% 企业客户位于欧洲,BBC、Fremantle 和 WPP 均为其客户。新总部将招聘研究、产品、工程、销售等岗位,并招募一名 Head of Europe 统筹欧洲业务。
Runway NewsAI 评分3434 Runway 推出内部孵化器 Runway Labs,探索生成式视频新应用
Runway 宣布成立内部孵化器 Runway Labs,专注探索生成式视频与 General World Models 的下一代应用。该孵化器由联合创始人兼首席创新官 Alejandro Matamala Ortiz 领导,负责孵化处于 AI 视频前沿的实验性产品。Runway 同时启动招聘,招募设计、ML、全栈及 AI 工程师与创业者。