Arcana Labs Integrates 25 AI Tools to Slash Film Production Timelines

Arcana Labs整合25款AI工具 大幅缩短电影制作周期 Arcana Labs通过统一的多引擎AI平台将长片制作周期从数年缩短至数天，同时仍需人类创作者掌控创意主导权。 与会嘉宾深入探讨了AI在电影制作中的应用现状。Breckin Meyer指出了当前存在的“恐怖谷”等局限，同时主张AI应作为辅助人类创作者的创意工具，而非取而代之。Guy Ronen解释了Arcana Labs如何将25种以上的各异AI工具整合至单一平台，将制作周期从数年压缩至数天。展望未来五年，演讲者预见行业将走向平民化，独立创作者能够制作完整的长篇电影，并向制片厂推介完全具象化的视觉概念，避免其构想在沟通转译中走样。讨论最后展示了一段Arcana短视频，呈现了AI生成氛围场景的概念验证。 Breckin Meyer指出传统长片开发需7到14年，而Arcana Labs的AI工作流将制作周期缩短至数天，但目前的AI内容仍面临恐怖谷效应等局限。 (15:53) Arcana Labs在单一平台中集成了25种以上的AI工具以处理专项任务，但嘉宾强调技术无法替代人类创作者，单纯依赖自动化只会产出千篇一律的乏味内容。 (17:24) 未来五年AI将推动电影制作民主化，创作者无需巨额预算即可拍摄完整长片，并可通过直观的视觉提案替代口头陈述来简化项目立项流程。 (19:17) 圆桌论坛以Arcana Labs的概念验证视频展示作结，直观呈现了其AI引擎在氛围序列与音效生成方面的创意能力。 (21:55)

While generative artificial intelligence can compress production cycles from years to days, industry panelists argue that avoiding formulaic content requires pairing unified multi-model tools with human storytelling. Compressing Production Timelines Amid Uncanny Valley Constraints (15:53) Actor Breckin Meyer observes that current AI-generated material still struggles with an uncanny valley effect, likening artificial scenes to video game cutscenes that viewers want to skip and noting that automated tools struggle with nuanced humor. Even so, moderator Roger Cheng and Meyer view hybrid workflows—combining traditional filmmaking with AI—as a practical path forward given traditional industry friction. Meyer points out that conventional feature film development typically spans 7 to 14 years, while his eight-year voice work on King of the Hill required a full year from script to screen because animation had to be sent overseas and returned. In contrast, Arcana Labs co-founder Guy Ronen states that Arcana's AI-assisted pipelines compress turnaround times down to days, allowing creators to ideate concepts in seconds and combine visual assets rapidly. Despite current aesthetic shortcomings, Meyer views these tools as an accessible creative outlet that he is glad younger creators can use to experiment directly on home computers. Consolidating Multi-Engine Workflows Under Human Creative Direction (17:24) Addressing technological fragmentation, Ronen explains that dozens of AI engines are evolving on a weekly basis, each tailored to narrow proficiencies such as rendering explosions and action sequences versus capturing precise skin tones and facial expressions. To mirror a filmmaker's natural workflow, Arcana Labs integrates more than 25 separate AI engines into a unified platform covering the progression from text or image prompts to generated video and audio. Both panelists emphasize that technological integration cannot substitute for human oversight. Meyer cautions that relying solely on automated generation will produce formulaic and boring action projects, asserting that creative control must remain with human directors to achieve compelling results like Echo Hunter. Ronen concurs, stating explicitly that filmmakers cannot skip human storytellers and that there are no shortcuts around genuine creative talent. Democratizing Independent Production and Streamlining Studio Pitches (19:17) Forecasting the entertainment sector over a five-year horizon, Ronen anticipates that AI software will democratize filmmaking by functioning as an end-to-end studio in software. In his view, this capability enables individuals who cannot afford cameras or multi-million-dollar budgets to direct full feature films independently, noting that industry conference attendees have expressed eagerness to realize lifelong production dreams. Addressing institutional Hollywood practices, Meyer expects AI to fundamentally streamline the studio pitching process. Rather than relying on junior executives to verbally relay concepts to decision-makers—a process where creative visions are often lost in translation—emerging creators will be able to supply fully realized visual decks and presentations that show studio executives exactly what the final production will look like. The panel concluded with an Arcana Labs proof-of-concept video demonstrating atmospheric sequences, background music, and minimal dialogue cues, illustrating the platform's generation capabilities while questions remain regarding how audiences and traditional studios will embrace AI-assisted productions at scale.

行业从业者指出，尽管生成式人工智能在幽默感和拟真度上仍受制于“恐怖谷效应”，但通过多工具融合与人类叙事主导的混合制作模式，AI正将影视制作与提案周期从数年大幅缩短至数天，并推动电影制作的平民化。 制作周期压缩与多模型生态整合 (15:53) 在传统影视制作流程中，项目开发周期通常极其漫长。演员兼创作者Breckin Meyer指出，一部标准故事片的开发与制作往往需要历时7到14年；以他参与配音长达八年的动画片《King of the Hill》为例，由于需要将素材送往海外加工并等待返回，从剧本到最终成片耗时整整一年。针对这一痛点，Arcana Labs的Guy Ronen表示，通过AI工作流，创作者可以在数秒内构想内容，并将整体制作周期大幅缩减至数天。 面对快速更迭且高度分散的AI技术，Guy Ronen解释称，目前市面上存在数十种每周都在迭代的独立引擎，且各具所长——例如部分模型擅长渲染动作场面与爆炸特效，而另一些模型则专精于皮肤质感与面部捕捉。Arcana Labs通过将超过25种不同的AI工具集成至单一平台，模拟电影创作者从文本提示、图像上传到最终视频和音频生成的完整工作流，从而解决了单一模型功能局限的问题。 技术局限与人类叙事的核心地位 (17:24) 尽管AI技术进步显著，Breckin Meyer强调当前技术仍存在明显瓶颈。他将目前的AI生成内容比作电子游戏中让人想按“A键”跳过的过场动画，观众能轻易识别出其非真实性，尤其是能够察觉作品并非出自真正懂得幽默的创作者之手。Meyer警告称，如果完全依赖自动化生成，最终只会产出遵循俗套公式的枯燥动作电影。他认为AI应被视为辅助现代创作的重要工具，但唯有结合人类主导的混合制作模式，才能孕育出具有艺术活力的作品。 针对技术与创意的关系，Guy Ronen高度认同Meyer的观点，并明确表示绝不能跳过人类故事讲述者。Ronen强调在影视创作中没有捷径可走，多模型集成的核心目的在于赋能创作者，而非取代编剧与导演的审美判断与创意方向。 五年展望：创作平民化与提案流程重构 (19:17) 展望未来五年的行业趋势，Guy Ronen预测AI技术将彻底推动影视制作的民主化，使任何人无论预算高低都能成为创作者。他认为软件驱动的工作室将打破资金壁垒，“下一位Steven Spielberg”或许就出自此前无力购买昂贵摄影器材、但如今能独立制作完整故事片的独立创作者群体中。 在商业运作层面，Breckin Meyer认为AI将深度重塑好莱坞传统的工作室提案流程。传统提案依赖初级管理人员对创意的理解并层层转述，容易导致核心构想失真；而借助Arcana等AI创作工具，即便是一位刚高中毕业的年轻创作者，也能直接拿出高度视觉化的完整提案图册与动态演示，向决策高管精确呈现最终成片效果，彻底解决创意在沟通过程中“被转译丢失”的行业顽疾。 本场讨论以Arcana Labs的概念验证视频展示作结，该视频演示了基于AI引擎生成的气氛渲染镜头、配乐与极简对白。综合讨论表明，虽然AI在概念可视化和制作降本上展现出颠覆性潜力，但其商业与艺术落地仍严格受制于拟真度短板，依赖人类创作者在叙事与审美层面的把控。

## Background 本内容由 Ai4 于 2025 年 8 月 24 日发布在 YouTube 平台上。节目围绕人工智能如何重塑好莱坞电影制作的艺术与商业模式展开，邀请了 TheWrap 执行主编 Roger Cheng、Arcana Labs 首席运营官 Guy Ronen 以及资深演员兼编剧 Breckin Meyer 共同探讨生成式 AI 在影视创作中的前沿应用与行业未来。 ## TL;DR Arcana Labs 介绍了其整合 25+ 专用生成式 AI 引擎的一体化创作工作流，并以首批获得演员工会（SAG-AFTRA）批准的短片《Echo Hunter》为例，展示了将制作周期与预算大幅压缩的颠覆性潜力。对谈强调坚守“艺术家驱动 AI”而非取代人类核心创作者，并通过严格的工会协议与模型销毁条款保障演职人员权益。嘉宾们还深入解析了 AI 混合工作流在突破传统定格动画与长片制作周期瓶颈上的实践价值，指出未来五年 AI 将彻底瓦解高昂的资本壁垒，实现影视创作与视觉提案的全民化普及。 ## Atmosphere & style 整场对谈兼具行业前瞻性与轻松诙谐的基调，Breckin Meyer 凭借 35 年工会老将的幽默感与通俗比喻，巧妙化解了业界对 AI 夺取饭碗的恐慌情绪。Guy Ronen 与 Roger Cheng 则从技术架构与产业观察的角度沉稳呼应，形成了传统影视工匠精神与前沿科技工具之间的理性对话。现场交流充实而接地气，既没有盲目的技术狂热，也消解了防备与抵触，传递出一种务实、包容且对影视未来充满底气的人机协作氛围。 ## Key takeaways - [00:04](https://www.youtube.com/watch?v=70J77BLX-vE&t=4s) Arcana Labs 致力于打造聚合 25 种以上生成式 AI 引擎的一体化平台，为电影制作人与创意工作者提供端到端的统一制作工作流。 - [00:04](https://www.youtube.com/watch?v=70J77BLX-vE&t=4s) 短片《Echo Hunter》是首批获得美国演员工会（SAG-AFTRA）批准认可的 AI 影视项目之一，消除了演职人员对 AI 取代真人的抵触与顾虑。 - [01:59](https://www.youtube.com/watch?v=70J77BLX-vE&t=119s) AI 极大压缩了制作周期与成本，同等体量项目的制作时间从数周缩短至 3~4 天，制作预算甚至低于传统好莱坞剧组一天的餐饮开销。 - [04:33](https://www.youtube.com/watch?v=70J77BLX-vE&t=273s) 行业应当坚持“艺术家驱动 AI”而非“AI 驱动艺术”，核心创作者永远不可替代，AI 的核心价值在于大幅拉低独立电影的资金与技术门槛。 - [07:54](https://www.youtube.com/watch?v=70J77BLX-vE&t=474s) 类比《星战》早期全 CG 角色 Jar Jar Binks，AI 制作仍依赖导演围读剧本和演员真实表演，是对表演形式的扩展而非剥夺。 - [09:40](https://www.youtube.com/watch?v=70J77BLX-vE&t=580s) 为防止侵权与低质内容滥用，正规制作严格遵守 SAG-AFTRA 协议，专门训练的演员数字模型在项目完结后必须被强制销毁。 - [11:53](https://www.youtube.com/watch?v=70J77BLX-vE&t=713s) 预计在 6 到 18 个月内 AI 将跨越“恐怖谷”效应，未来制作模式更倾向于实拍与 AI 按比例融合的混合工作流（Hybrid Model）。 - [14:13](https://www.youtube.com/watch?v=70J77BLX-vE&t=853s) 掌握高颗粒度、具备导演思维的“提示词工程”成为创作者的新核心技能，能将传统动画长达一年的生产周期缩短至数天。 - [17:24](https://www.youtube.com/watch?v=70J77BLX-vE&t=1044s) 由于没有单一大模型能胜任所有细分场景，聚合多专项模型的集成化系统比单一专有大模型更能满足专业影视制作需求。 - [20:27](https://www.youtube.com/watch?v=70J77BLX-vE&t=1227s) AI 驱动的视觉化方案彻底改变了传统影视提案的“信息失真”痛点，让年轻独立创作者能以极低成本实现“所见即所得”的视觉呈现。 ## Detailed walkthrough ### [00:04](https://www.youtube.com/watch?v=70J77BLX-vE&t=4s) 认识 Arcana Labs 与 Breckin Meyer 的 AI 初体验 TheWrap 执行主编 Roger Cheng 作为主持人开场，欢迎观众并引出主题。Arcana Labs 首席运营官 Guy Ronen 介绍了公司的核心定位：Arcana Labs 致力于打造一个全方位的 AI 平台，将市面上所有生成式 AI 引擎整合为一套面向创作者（包括电影制作人、营销人员等各类创意工作者）的统一工作流，旨在极大简化和加速创作流程。 随后，主持人询问资深演员兼制片人 Breckin Meyer 是如何涉足 AI 领域并与 Arcana 建立联系的。Breckin 幽默地分享道，经纪人最初给他打电话推介这个 AI 项目时，他的第一反应是困惑与怀疑——“等等，我不是 AI 啊？”然而，当他得知这是首个获得美国演员工会（SAG-AFTRA）批准认可的 AI 项目时，作为拥有 35 年工会会龄的老成员，他的兴趣瞬间被点燃。在与 Arcana 的联合创始人 Yoni 以及短片导演 Cavin the Kid 见面沟通后，对方以通俗易懂的方式（如同“AI 傻瓜书”一般）向他解释了运作原理。Breckin 结合自己参与制作长达 20 年的定格动画《机器鸡》（Robot Chicken）的技术演进经验，迅速意识到生成式 AI 在优化制作管线上的巨大潜力，消除了“AI 参与就意味着不需要真人演员”的误解。 ```mermaid flowchart TD accTitle: Arcana Labs 核心模式 subgraph S1["技术整合层"] A["多源生成式 AI 引擎"] --> B["Arcana Labs 统一平台"] end subgraph S2["创意赋能层"] B --> C["创作者 / 电影制作人"] B --> D["SAG-AFTRA 工会专业演员"] end subgraph S3["最终交付"] C --> E["电影短片与创意内容<br/>(如 Echo Hunter)"] D --> E end ``` - **Arcana Labs 定位**：汇聚多种底层 GenAI 引擎的一体化创作工作流工具。 - **工会合规**：短片《Echo Hunter》作为首批获得 SAG-AFTRA 批准的 AI 电影项目之一，打消了传统演职人员的抵触情绪。 - **嘉宾背景**：Breckin Meyer 兼具 35 年工会演员资历与 20 年《机器鸡》动画制作人经验。 --- ### [01:59](https://www.youtube.com/watch?v=70J77BLX-vE&t=119s) 展现未来制作管线：《Echo Hunter》幕后与成本颠覆 现场播放了短片《Echo Hunter》的幕后花絮片段。短片导演兼 Phantom X 联合创始人 Cavin the Kid 在片段中介绍，该片是一部完全基于 Arcana Labs 前沿生成式 AI 工具打造的影片，融合了下一代技术与全明星 SAG 演员阵容，力图将未来的电影制作方式带入现实。 在随后的讨论中，Guy Ronen 深入拆解了该项目的制作周期与惊人成本。在拍摄《Echo Hunter》时（距对谈约四个半月前），整个制作周期耗时 3.5 周；而随着技术工具的飞速迭代，同等体量的项目在当下只需 3 到 4 天即可完成。谈及成本，Guy 直言虽然不透露具体数字，但整体预算“甚至低于传统剧组单日餐饮的开销”。尽管采用了 AI 技术，团队仍坚持采用传统电影制作规范——雇佣 Breckin 等 SAG 工会演员、签订正式工会合同、配置专业调色、传统后期剪辑，并聘请作曲家专门谱写配乐。这种做法旨在主动融入好莱坞工业体系，向行业证明 AI 不是要消灭影视业，而是可以通过传统工艺与 AI 的结合实现产业平滑过渡。 - **制作周期对比**：从 4.5 个月前的 3.5 周大幅压缩至当下的 3~4 天。 - **极致预算**：全片制作成本低于常规好莱坞剧组一天的剧组盒饭/餐饮预算（daily catering）。 - **专业工业化流程**：保留传统录音、调色、配乐及工会标准合同，探索人机协作规范。 --- ### [04:33](https://www.youtube.com/watch?v=70J77BLX-vE&t=273s) 艺术家驱动的 AI：降低门槛与打破大制片厂垄断 Roger Cheng 指出，如今好莱坞大制片厂几乎被超级英雄电影和超级 IP 续集垄断，独立创作者举步维艰，AI 的出现为中小规模创作者带来了破局机遇。Breckin Meyer 深表认同，并回顾了影视制作门槛的历史变迁：十年前粉丝问如何做《机器鸡》，只能建议买专业摄像机；后来智能手机和 YouTube 普及，彻底解放了短视频创作；而 AI 正是这一赋能进程的最新一步。 Breckin 坦言自己完全不擅长画画，但 Arcana 这样的工具让他能将脑海中的梦幻角色和创意视觉化。“这是 AI 在协助艺术家，而不是取代艺术家”，这种技术让没有超级英雄大片预算、无法敲开制片厂大门的独立创作者，也能做出像《机器鸡》或《穿鞋子的贝壳马塞尔》（Marcel the Shell with Shoes On）那样富有灵气和创意的作品。主持人将其精辟概括为：“艺术家驱动的 AI（artist-driven AI），而非 AI 驱动的艺术（AI-driven art）”。 Guy Ronen 进一步阐述，Arcana 是由电影人创造、服务于电影人和故事讲述者的。公司的核心使命是“降低创作门槛、出售创作自由”，只要创作者能想象出来，就能用工具呈现。制片人、导演、编剧、剪辑等核心岗位（Above-the-line）永远不可替代，AI 所做的是以极低的成本赋予他们更高的制作价值（Production Value）。 - **创作赋能**：帮助不具备手绘技能的创作者直接将剧本想法具象化为视觉镜头。 - **行业定位**：坚守“艺术家驱动 AI”原则，工具旨在解放表达而非替代核心创作者。 - **核心价值**：大幅拉低独立电影与小众题材进入市场的资本与技术门槛。 --- ### [07:54](https://www.youtube.com/watch?v=70J77BLX-vE&t=474s) 消除行业恐慌：从《星战》Jar Jar Binks 到真正的表演协作 针对好莱坞业界对 AI 导致失业的普遍恐慌与抵触心理，Breckin Meyer 分享了身边年轻一代的真实反应。他 21 岁正在戏剧学院深造的女儿得知他接拍 AI 电影时，第一反应也是质问：“这里面有你的表演吗？有导演吗？”Breckin 告诉她，实际工作方式与传统排练毫无二致——他依然与导演坐在一起围读剧本、推敲台词，完全由真实的情感和导演指导驱动。 Breckin 引用了当年乔治·卢卡斯在《星球大战》前传中采用全 CG 角色“恰恰·宾克斯（Jar Jar Binks）”的历史先例：当时全行业也都在抗议卢卡斯会毁掉演员饭碗，但事实上 Jar Jar 的背后依然是动作捕捉与配音演员 Ahmed Best 的真实表演。在 AI 短片制作中，现场并非“一个演员身上贴满标记点、独自面对一个光球”，而是真实的导演、编剧、制片人在共同进行艺术创造。AI 只是扩展叙事和视觉优势的新媒介，当行业以尊重演职人员、规范签约的正确方式拥抱它时，它将成为推动影视工业进化的全新生产力。 - **认知误区澄清**：AI 影视制作并非算法黑盒独角戏，仍依赖导演现场调度、围读剧本及演员实时表演。 - **历史镜像案例**：类比《星球大战》早期动捕角色 Jar Jar Binks 与演员 Ahmed Best，论证技术变革是对演员表现手法的拓展而非剥夺。 ### [09:40](https://www.youtube.com/watch?v=70J77BLX-vE&t=580s) 抵制低质内容与安全防护：演员权益保护与数字模型销毁 TheWrap 的编辑 Roger Cheng 抛出关于当下网络上充斥着粗制滥造的“AI 垃圾内容”（AI slop）的现象，询问如何平衡公众因低质内容产生的负面偏见与高水准、合规制作的 AI 作品。Arcana Labs 首席运营官 Guy Ronen 强调，首要任务是建立严格的安全防护机制（Safeguards）。在面向消费者的 D2C 平台上，严禁滥用、深度伪造（Deepfake）、盗用公众人物肖像或侵犯版权；而在专业制作端，则是在保障安全边界的同时，为创作者提供易用的工具，确保创意的风格与故事自由度不被削弱。 针对编剧和演员对自身肖像与作品被喂给 AI 模型进行训练的深切担忧，演员兼编剧 Breckin Meyer 分享了他在拍摄 AI 参与制作的影片《Echo Hunter》时的实际法律保障。根据演员工会（SAG-AFTRA）的合同规定，制作方为他构建的 AI 形象模型在电影制作完成后必须被销毁，且严格限制其商业用途——未经许可绝不能拿去制作狗粮或猫粮广告。Breckin Meyer 幽默地调侃道，虽然理论上任何人都可能通过输入提示词生成一段“Breckin Meyer 和格温妮丝·帕特洛（Gwyneth Paltrow）一起吃意大利面”的诡异视频（他笑称“这绝对是一部烂片，大家千万别去看”），但在合法正规的生产体系中，这种侵权行为是不被允许的，而且未经精细打磨的生成内容往往一眼就能看出陷入了假模假样的“恐怖谷”（Uncanny Valley）。 - **平台安全底线**：严厉杜绝未经授权的公众人物肖像复制、Deepfake 以及侵犯知识产权的行为。 - **演员工会条款**：严格遵循 SAG-AFTRA 协议，专用训练的演员数字孪生模型在项目结束后强制销毁，严防跨项目的商业滥用。 --- ### [11:53](https://www.youtube.com/watch?v=70J77BLX-vE&t=713s) 突破“恐怖谷”效应：AI 生成与传统 CGI 的技术差异 Roger Cheng 谈到他在观看《Echo Hunter》时，仍然能辨别出某些带有明显 AI 生成痕迹的片段，并询问技术何时才能彻底跨越“恐怖谷”实现全拟真。Guy Ronen 解释道，《Echo Hunter》是 Arcana Labs 在四五个月前打造的一个概念验证（Proof of Concept）项目，旨在尽早证明 AI 全流程制作的可行性。AI 的迭代速度极其惊人，仅仅数月之后，团队现在就能在节省 90% 时间的前提下产出质量高得多的画面。 对于实现完全真实感的具体时间表，Guy Ronen 认为无需等待 5 到 10 年，预计只需 **6 到 18 个月**，动作捕捉（Motion Capture）以及对 AI 角色的精准控制就能达到成熟水准。在对比 AI 生成与传统计算机图形（CGI）时，Guy Ronen 指出： - **传统 CGI 流程**：往往依赖绿幕、物理跟踪标记点或大型道具（例如《权力的游戏》中演员抚摸的巨型实体龙形头部道具），后期再将数字资产与物理动作贴合。 - **AI 生成流程**：所有视觉要素都是端到端虚拟嵌入并实时生成的，因此目前要将实拍素材与 AI 生成元素无缝融合仍具备相当挑战。 Arcana 并不推崇极端、非黑即白的 100% 纯 AI 制作，而是倡导符合创作者意愿的混合工作流（例如 70/30、50/50 或 60/40 的实拍与 AI 混合比例）。 ```mermaid flowchart TD accTitle: CGI vs AI Production Workflow Comparison subgraph CGI["传统 CGI 制作模式"] A["实拍场地与物理道具<br>(例如绿色标记点/龙首实体道具)"] --> B["演员进行物理交互"] B --> C["后期数字嵌入与合成<br>(高时间/资金成本)"] end subgraph AI["AI 智能生成模式"] D["提示词与多模态输入"] --> E["全虚拟实时生成与嵌入"] E --> F["混合工作流输出<br>(实拍与 AI 依比例融合)"] end ``` --- ### [14:13](https://www.youtube.com/watch?v=70J77BLX-vE&t=853s) 提示词的艺术与定格动画变革：从一年周期到数天极速交付 谈到 Breckin Meyer 长期参与创作的经典定格动画《机器鸡》（Robot Chicken），主持人询问他是否会考虑在后续项目中引入 Arcana 的技术。Breckin Meyer 兴奋地透露他们已经与 Arcana 团队进行了接洽。通过与精通 AI 商业广告制作的导演 Caven 合作，Breckin Meyer 发现向 AI 下达指令的“提示词工程”（Prompting）与导演给演员讲戏有着异曲同工之妙——两者都需要深入了解性格与精准表达。例如，只说一句“给我一个汉堡”可能得到平庸的结果，但如果详细描述为“要一个沃尔夫冈·帕克（Wolfgang Puck）名厨风格、上面带有煎蛋的汉堡”，产出的细节就会完全不同。 Breckin Meyer 坦言，传统定格动画极其繁琐笨重，“点一下拍一帧、挪一下腿再拍一帧”，导致《机器鸡》从写剧本到最终成片需要整整一年时间（他此前配音了八年的经典动画《山丘之王》（King of the Hill）同样因为需要外包动画绘制而耗时一年）。对于一部极其依赖流行文化梗和时效性包袱的讽刺喜剧来说，长达 9 到 12 个月的滞后周期是非常致命的挑战。AI 工具能够让创作者迅速跟进当下热点，极大提高制作效率。 Guy Ronen 补充说明，通常一部传统电影长片的制作周期长达 **7 到 14 年**，而在 Arcana 的 AI 生态工具赋能下，概念构思与视觉混合可以缩短到**几天乃至数秒**内完成，创作者能够瞬间尝试不同的分镜与视觉风格，自由决定哪些镜头进行实拍、哪些场景使用 AI 快速生成。 - **创作周期对比**： - 传统长片：7–14 年 - 《机器鸡》/《山丘之王》传统动画：从剧本到屏幕耗时约 1 年 - Arcana AI 混合工作流：从设想到多版本视效测试仅需数秒到数天 - **提示词艺术（Prompt Craft）**：掌握高颗粒度、具有导演思维的提示词成为现代创作者的核心新技能。 --- ### [17:24](https://www.youtube.com/watch?v=70J77BLX-vE&t=1044s) 聚合 25+ 专用引擎的电影级全流程生态 Roger Cheng 探讨了技术集成策略，询问为什么 Arcana Labs 选择聚合多种外部模型，而不是押注单一的专有大模型。Guy Ronen 解释道，AI 底层模型的演进呈周级别爆发，没有任何单一大模型能擅长所有细分领域——有的模型在爆炸、动作场景的物理动态上表现卓越，而有的模型则在人类皮肤质感与面部表情捕捉上更具优势。 Arcana Labs 的核心壁垒在于打造专为影视制作人设计的集成化操作系统，平台聚合了超过 25 种前沿 AI 工具与模型，将提示词交互、图片生成、实景视频嵌入直到最终音效制作（Soundtrack/Audio）整合在统一的流水线中，创作者无需在不同软件工具间繁琐切换，即可在同一个套件中实现从构思到母带输出的端到端制作。 ```mermaid flowchart LR accTitle: Arcana Labs Integrated AI Ecosystem subgraph Engines["多元专项 AI 模型层 (25+ 引擎)"] E1["动作与物理爆炸模型"] E2["面部表情与肤质渲染模型"] E3["音频与音效生成模型"] end subgraph ArcanaPlatform["Arcana Labs 统一制作平台"] P1["文本/提示词输入"] P2["场景/图片嵌入"] P3["混合视频生成"] P4["声音后期合成"] end Engines --> ArcanaPlatform ArcanaPlatform --> Output["电影级成品输出"] ``` ### [18:45](https://www.youtube.com/watch?v=70J77BLX-vE&t=1125s) 人机混合创作与不可替代的故事讲述者 在探讨技术与艺术的关系时，嘉宾们强调了创作者参与的核心价值。如果只是单纯对 AI 下达“帮我做这个”的指令，产出的内容往往会显得偷懒、粗糙且毫无新意；正如一部优秀的传统电影需要顶尖的编剧、演员和制片人协同打造一样，AI 时代同样需要创作者全程深度介入。所谓的混合创作模式（Hybrid Model），要求创作者始终保持审美把控与叙事主导权，而不是将所有工作甩给算法。项目《Echo Hunter》的成功正是由于团队在每一个环节都倾注了创作者的构思与执念。 无论技术如何迭代，人类作为故事讲述者的角色永远不可替代，创作没有任何捷径可走。技术的意义在于放大创作者的表达力，而非抹杀叙事的核心灵魂。 ```mermaid flowchart TD accTitle: AI创作路径对比 subgraph S1["偷懒的纯AI模式"] A["简单粗暴的AI指令"] --> B["缺乏灵魂的粗糙产出"] end subgraph S2["深度结合的人机混合模式"] C["故事创作者理念"] --> D["AI生成与实时协作"] D --> E["持续的人工审美介入与微调"] E --> F["高水准的影视叙事作品"] end ``` ### [19:17](https://www.youtube.com/watch?v=70J77BLX-vE&t=1157s) 五年展望：影视创作门槛的彻底瓦解 谈及未来发展，主持人 Roger Cheng 指出，制作《Echo Hunter》最初耗时数月，而现在同样的流程只需短短几天，技术的进化速度令人惊叹。面对未来五年的行业走向，Arcana Labs 首席运营官 Guy Ronen 认为，AI 将催生大量全新的职业与岗位，并真正实现影视制作的全民化。未来的“下一个史蒂芬·斯皮尔伯格（Steven Spielberg）”可能就坐在此刻的观众席中，过去他们或许因缺乏资金与设备而无法起步，但如今 AI 赋予了每个人所需的一切创作能力。 在近期的多次行业会议中，不少影视从业者眼中闪烁着光芒与 Guy Ronen 交流，坦言自己曾因连一台摄像机都买不起而搁置梦想，如今却能凭借个人力量完成一部长片电影。未来的创作工具将更加贴合个性化需求，让人们能够以前所未有的速度制作高度定制化的精准内容。 - **创作平民化**：彻底消除摄影器材、昂贵算力与大型摄制组带来的资金壁垒。 - **效率飞跃**：影视制作周期从过去的数月、数年压缩至以天为单位。 - **定制化内容**：创作者能够根据细分受众与具体场景快速产出专属视听作品。 ### [20:27](https://www.youtube.com/watch?v=70J77BLX-vE&t=1227s) 从文字提案到精准视觉呈现与圆满收尾 谈到五年后的愿景，Breckin Meyer 先是幽默地打趣道，五年后自己可能会去出演“新版蝙蝠侠（Batman 14 或 Batman Pi）”，随后他严肃地指出，AI 技术在五年的时间跨度内将产生难以估量的颠覆性影响。对于一个刚高中毕业、怀揣天才构想却囊中羞涩的 17 岁年轻人而言，AI 平台让他们无需筹集 2000 万美元或苦恼于制作复杂的提案 Deck，就能拥有把想象力变为现实的直接通道。 传统的好莱坞提案流程往往充满损耗：创作者向基层主管阐述创意，只能祈祷对方能准确理解并向上级领导转述，整个过程极易发生“信息失真（Lost in Translation）”。而借助 Arcana 这样的 AI 引擎，创作者可以直接拿出一套高度精细、直观立体的视觉概念方案，向资方明确展示“你最终将看到什么”。最后，主持人向两位嘉宾致谢，并播放了由 Arcana 引擎制作的全新概念演示视频，直观展现了 AI 驱动未来影视视效的强大潜能。 - **解决传达损耗**：从依靠口头表述的“盲盒式提案”转变为“所见即所得”的高精度视觉展示。 - **赋能年轻创作者**：摆脱数千万美元的巨额预算束缚，让优秀创意直接获得落地机会。 - **闭幕演示**：通过 Arcana 实验室的全新概念短片，展示了引擎在画质细节与氛围渲染上的强劲实力。 ## Notable quotes > [00:04](https://www.youtube.com/watch?v=70J77BLX-vE&t=4s) 等等，我不是 AI 啊？ — Breckin Meyer > [01:59](https://www.youtube.com/watch?v=70J77BLX-vE&t=119s) 甚至低于传统剧组单日餐饮的开销 — Guy Ronen > [04:33](https://www.youtube.com/watch?v=70J77BLX-vE&t=273s) 这是 AI 在协助艺术家，而不是取代艺术家 — Breckin Meyer > [04:33](https://www.youtube.com/watch?v=70J77BLX-vE&t=273s) 艺术家驱动的 AI（artist-driven AI），而非 AI 驱动的艺术（AI-driven art） — Roger Cheng > [04:33](https://www.youtube.com/watch?v=70J77BLX-vE&t=273s) 降低创作门槛、出售创作自由 — Guy Ronen > [07:54](https://www.youtube.com/watch?v=70J77BLX-vE&t=474s) 这里面有你的表演吗？有导演吗？ — Breckin Meyer 的女儿 > [09:40](https://www.youtube.com/watch?v=70J77BLX-vE&t=580s) Breckin Meyer 和格温妮丝·帕特洛（Gwyneth Paltrow）一起吃意大利面 — Breckin Meyer > [09:40](https://www.youtube.com/watch?v=70J77BLX-vE&t=580s) 这绝对是一部烂片，大家千万别去看 — Breckin Meyer > [14:13](https://www.youtube.com/watch?v=70J77BLX-vE&t=853s) 要一个沃尔夫冈·帕克（Wolfgang Puck）名厨风格、上面带有煎蛋的汉堡 — Breckin Meyer > [19:17](https://www.youtube.com/watch?v=70J77BLX-vE&t=1157s) 下一个史蒂芬·斯皮尔伯格（Steven Spielberg） — Guy Ronen > [20:27](https://www.youtube.com/watch?v=70J77BLX-vE&t=1227s) 你最终将看到什么 — Breckin Meyer ## Mentioned Breckin Meyer, Guy Ronen, Roger Cheng, Arcana Labs, TheWrap, Echo Hunter, SAG-AFTRA, Yoni, Cavin the Kid, Phantom X, 机器鸡 (Robot Chicken), 穿鞋子的贝壳马塞尔 (Marcel the Shell with Shoes On), 星球大战 (Star Wars), 乔治·卢卡斯 (George Lucas), Jar Jar Binks, Ahmed Best, Gwyneth Paltrow, 权力的游戏 (Game of Thrones), 山丘之王 (King of the Hill), Wolfgang Puck, Steven Spielberg ## Video description Breckin Meyer, Actor/Writer, Echo Hunter Guy Ronen, COO, Arcana Labs Roger Cheng, Managing Editor, TheWrap Abstracts How will artificial intelligence redefine the magic and business of filmmaking? AI is rapidly transforming industries, and Hollywood stands on the brink of a revolution, poised to reshape creative processes, production pipelines, and audience engagement. Attendees will gain insights into Arcana Labs' vision for AI's role in entertainment and understand the emerging opportunities and challenges for the industry. Join us to explore the groundbreaking future where AI and Hollywood converge. [Source](https://www.youtube.com/watch?v=70J77BLX-vE)

[0:04] Breckin Meyer: 等等，我不是 AI 啊？ [1:59] Guy Ronen: 甚至低于传统剧组单日餐饮的开销 [4:33] Breckin Meyer: 这是 AI 在协助艺术家，而不是取代艺术家 [4:33] Roger Cheng: 艺术家驱动的 AI（artist-driven AI），而非 AI 驱动的艺术（AI-driven art） [4:33] Guy Ronen: 降低创作门槛、出售创作自由 [7:54] Breckin Meyer 的女儿: 这里面有你的表演吗？有导演吗？ [9:40] Breckin Meyer: Breckin Meyer 和格温妮丝·帕特洛（Gwyneth Paltrow）一起吃意大利面 [9:40] Breckin Meyer: 这绝对是一部烂片，大家千万别去看 [14:13] Breckin Meyer: 要一个沃尔夫冈·帕克（Wolfgang Puck）名厨风格、上面带有煎蛋的汉堡 [19:17] Guy Ronen: 下一个史蒂芬·斯皮尔伯格（Steven Spielberg） [20:27] Breckin Meyer: 你最终将看到什么

[15:53] Breckin Meyer: you're like all right it's not real [15:53] Breckin Meyer: wasn't written by somebody who knows humor very well [15:53] Guy Ronen: days [17:24] Guy Ronen: You can't skip the storytellers. You can't skip that. There's no shortcuts here. [19:17] Guy Ronen: the next Steven Spielberg [19:17] Breckin Meyer: the new Batman [19:17] Breckin Meyer: Batman 14 [19:17] Breckin Meyer: exactly what you'll be getting [19:17] Breckin Meyer: lost in translation

[0:04] It is it is early. I'm glad. I'm [0:06] thankful and grateful for all of you to [0:07] show up. This is going to be a great [0:08] chat. Uh Guy Breen, just to just to kick [0:11] things off, let's talk a little bit [0:12] about Arcana. We're going to get into [0:14] Hollywood. We're going to get into [0:14] entertainment, but let's let's talk [0:16] about Arcana Labs and what you guys are [0:18] doing. [0:18] >> Yeah, thanks for having us, guys. Arcana [0:22] Labs. Um, we're an AI platform [0:25] mitigating all the engines that you can [0:27] think of on Gen AI into one interesting [0:30] workflow for creators. It could be [0:32] filmmakers, it could be marketers, any [0:34] type of creators. And that's what [0:36] basically we're trying to streamline in [0:38] Arcana. [0:39] >> All right, Breen, just curious how you [0:41] got involved in this. You're you're a [0:43] longtime actor, producer. [0:44] >> Yeah. [0:45] >> Uh, how'd you get into AI and hooked up [0:47] with with Arcana? [0:48] >> I uh, you know, I got a call kind of as [0:50] a normal gig. I got a call from my [0:53] manager and they said, "Well, it's this [0:55] AI project." And I was like, "Well, that [0:57] doesn't I don't I'm not AI." And [1:01] >> are you sure about that? [1:01] >> I was I wasn't sure and I'm still not [1:03] sure now that I've seen the boxing and [1:05] the dog, but um I it was the it was [1:08] presented to me as the first SAG after [1:11] approved one, which immediately as a [1:13] union guy of 35 years now, uh [1:17] immediately maybe got me interested. And [1:19] I met Caven and Yoni who runs Arcana [1:22] with Guy and they presented this to me [1:25] and it sounded really really interesting [1:27] and it sounded really neat. And as a [1:28] tech fan and as an animation fan with [1:31] Robot Chicken that we've done for 20 [1:33] years now and has seen how we can [1:35] streamline our production uh just [1:37] through the last couple the last 10 [1:38] years we've done on our show. It was [1:40] really interesting to me and Cavin was [1:41] such a creative director of it that it [1:44] immediately kind of put everything, you [1:46] know, like most people when I heard AI, [1:47] I was like, "Wait, so that means I'm not [1:48] in it?" Like, "No, not at all. It's [1:50] actually quite different." So, once it [1:52] was presented to me and kind of spoken [1:54] to me as like as I needed to be told AI [1:56] for idiots, it was it helped a lot. [1:59] >> Well, just to better illuminate what [2:01] this actually means, let's show a clip [2:02] of Echo Hunter. This is a behind thes [2:04] scenes clip that uh really kind of dives [2:06] into what we're talking about. Echo [2:07] Hunter is a fully gen AI film built with [2:11] the cuttingedge tools of Arcana Labs. [2:14] I'm Cavin the kid, director of Echo [2:16] Hunter and co-founder of Phantom X. We [2:19] teamed up with Arana Labs to fuse [2:21] nextgen tech with an amazing cast of SAG [2:24] actors pushing the future of film making [2:26] into now. [2:29] >> It's been too long. [2:32] [Music] [2:35] >> I think it's frozen. [2:38] That's a corpse ever. [2:41] Consider the purpose of its actions. [2:48] That's a corpse ever. [2:51] Consider the purpose of its actions. [2:57] [Music] [3:09] Well, you could you could check out the [3:10] full video, the full 20-minute videos on [3:12] YouTube and on your site as well. Um, [3:14] it's fascinating stuff. And I think [3:15] what's what's interesting to me, um, [3:18] it's really the the time and sort of the [3:20] budget that was spent. If you talk a [3:22] little bit about that and just sort of [3:24] what that what went into this, how the [3:26] process differ from sort of the [3:27] traditional Hollywood model. [3:28] >> Yeah, Echo Hunter is a good good [3:30] example. It took us uh three and a half [3:33] weeks to produce that. By the way, today [3:36] we're talking about like four and a half [3:37] months ago. Today we can do it in [3:39] probably three to four days. Um budget [3:42] wise can get into numbers but probably [3:45] less than a typical budget for a daily [3:48] catering in a typical production. Um so [3:51] those are the numbers. I think for us we [3:54] wanted to emphasize the capabilities and [3:57] the tools. So we choose to do a lot of [4:00] things traditionally so we can actually [4:02] ease oursel into the industry. One of [4:04] the biggest problem is getting into the [4:05] industry and there's a lot of people [4:07] that really think the AI going to [4:08] eliminate the the entertainment industry [4:10] and that's totally not right. So we used [4:13] real actors like Ricken um and we paid [4:16] SAG contracts and we did a traditional [4:20] post with music and coloring and [4:23] everything and a composer to deal with [4:26] the soundtrack. So we literally try to [4:29] go with the traditional steps and [4:31] embedding AI in it. [4:33] >> Yeah. And I want to talk about that [4:35] because you know the budget being so [4:37] small that the time from you you know [4:39] this really opens up the opportunity to [4:40] create more content. And I want to throw [4:41] over to [4:42] >> Breen. I mean you've you've lived in [4:44] Hollywood for decades at this point. You [4:46] understand how hard it is to make a [4:48] project especially right now where [4:49] studios really just care about superhero [4:51] movies and these big franchise pieces. [4:53] What does AI what does AI enable or [4:56] allow for when it comes to these kind of [4:58] smaller projects or projects for [5:00] creators who maybe don't have those [5:02] opportunities? [5:02] >> Right. I mean that that's the one thing [5:04] to you know with with robot chicken [5:06] which is a stop motion animated show if [5:08] anyone's seen but with robot now we we [5:11] get asked all the time at panels how do [5:13] we get started how can I do my own robot [5:15] chicken 10 years ago it was like well [5:17] you got to get a video camera I guess [5:18] how we did and eventually the cell phone [5:20] came out and YouTube came out we were [5:22] able to say you can literally do it all [5:24] right here we weren't able to and you [5:25] can it streamlined it and the same with [5:27] AI is it streamlined there's no way we [5:30] could do what we do on with robot uh [5:32] with the antiquated ways we used to on [5:35] budget and what this does, what Arcana [5:38] does is it gives an opportunity. I don't [5:40] draw. I'm very bad with drawing. I'm not [5:41] that very good. I'm not good at it. And [5:43] what Arcana does is it is able to assist [5:46] me in getting my ideas and my kind of [5:49] dreams, not to be hokey, but to get my [5:51] dream of what I want this character to [5:53] look like and I can't draw to a reality. [5:56] And it's, you know, we you keep saying [5:57] with guy, it was artists assist, you [6:01] know, AI is assisting artists instead of [6:03] replacing. And it helps someone like [6:06] with budgets, it helps you streamline [6:08] it. And with people who couldn't [6:09] necessarily get into a studio to get a [6:11] pitch going or to get a if they don't [6:13] have a superhero movie, but they do have [6:15] something really neat, whether it's what [6:17] is it, you know, whether it's robot or [6:19] Marcel the Shell type things you can do [6:22] and this thing can really assist in [6:24] streamline it. So that was impressive to [6:26] me as a producer and as a creator y to [6:28] be able to do that [6:29] >> artistdriven AI and not AIdriven art. [6:32] >> Yeah. Well and that's that's a good [6:33] point and and I've talked to a number of [6:35] companies who are sort of following this [6:37] sort of ethical route of creating [6:39] content using AI. That said, uh AI I [6:42] would argue is still a bit of a dirty [6:44] word in Hollywood. I think most folks [6:46] here have embraced the technology. we, [6:48] you know, we kind of understand it, but [6:50] there's there's still a lot of fear and [6:51] trepidation when when you see a video [6:53] like Echo Hunter, your immediate thought [6:56] is, "Oh, am I to your point, is my job [6:59] gone? Is my job going to be gone in a [7:00] few years?" I'm curious [7:02] >> how that how that works or how you sort [7:04] of envision artists being empowered to [7:07] use these tools and like what that how [7:09] how we avoid the idea of job [7:11] displacement or replacement with AI. I [7:13] think in the essence are kind of built [7:15] by filmmakers to serve filmmakers and [7:17] storytellers. Okay? It's not here to [7:19] replace anyone. If if we're talking [7:21] about the type of a business we're in a [7:23] business of lowering the barrier and in [7:25] a business of selling freedom to create [7:28] whatever you want to. If you can imagine [7:30] that, we discuss that. If you can [7:31] imagine that, you can create that. Okay. [7:34] And talking on above the line uh uh [7:37] positions, we still need them. We need [7:39] storytellers, very talented. We need [7:42] producers. We need directors, we need [7:44] editors, we need all of those. We're [7:46] just easing the way into production and [7:49] getting a better production value [7:52] >> on a lower cost of course. [7:54] >> And Bracken, just uh in terms of uh [7:57] reaching out or when you talk you have [7:59] conversations with other actors and [8:00] producers, there there are I think [8:02] >> there's a decent segment of Hollywood [8:04] that probably dismisses AI out of hand [8:05] or is sort of burying their heads in the [8:07] sand a bit. I'm curious what you would [8:09] tell folks on that side in terms of how [8:11] or why they should be embracing it. [8:13] >> Yeah. I mean, having done it now, having [8:14] done Echo Hunter with uh Cabinet with [8:17] Arcana, you know, my I have a [8:20] 21-year-old who's at uh drama school [8:22] right now wanting to be an actor. And I [8:24] told her I was doing this and she had [8:26] the same reaction initially, which is [8:28] like, "Wait, so are you not in it?" And [8:30] I was like, "No, I'm in it." And she's [8:31] like, "But there's no director." And I'm [8:32] like, "No, there's a director. I sat [8:34] across from just like you would a normal [8:37] rehearsal." in production and read the [8:38] lines and blah blah blah and all that. [8:40] So it is showing it is kind of [8:43] explaining to people like no you're [8:44] assisting as I said you're assisting the [8:46] artist. It's like when George Lucas [8:48] started with Jar Jar and got all this [8:50] you know everyone was so angry that he [8:52] you know well they didn't like Jar Jar [8:54] but you know he was um but he was going [8:56] to put people out of work. It's like no [8:58] it's a different way. It's a different [8:59] way of doing it. Jar is still Ahmed [9:01] best. He's the actor who did it and if [9:03] you want to you know you can praise or [9:05] blame whatever you want to do with Amed. [9:06] He's a lovely guy. But to me, it is [9:10] another advantage that you can do [9:12] creatively. Another advantage you can do [9:15] creatively and story-wise as well with [9:17] the story and with the direction. And [9:19] but again, this [9:20] >> you know what you call AI short film. [9:23] >> I was the dire. I was meeting with the [9:25] director who was the writer who was the [9:26] producer. It was never me just with dots [9:28] alone in a room with an orb, you know, [9:30] which I think a lot of people think is [9:32] what it's going to be. [9:34] And you know the way the process here [9:36] you you know you've hired actors, [9:37] composers, you've I would argue you've [9:39] done this the right way. [9:40] >> Uh right now I mean YouTube is a good [9:43] example. There's I think a lot of [9:44] examples of AI done the wrong way. AI [9:46] slop is out there. I'm curious how you [9:48] sort of balance that the broader [9:51] perception of AI based off of AI slop [9:53] with some of the more slickly produced, [9:55] you know, ethically done AI uh work. [9:58] >> Well, there's a lot of aspect for that [10:00] starting from safety and security. Okay. [10:01] Okay. On a on the D2C side of the [10:03] platform, we're not allowing people to [10:05] abuse, to deep fake, to use u any public [10:08] domain figure and and kind of a [10:11] duplicate something or steal the [10:13] likeliness of anyone. Um that being [10:15] said, we're trying to give access to a [10:17] lot of tools to make things easier on [10:20] the production side, but not losing the [10:22] creativity, not losing the freedom of [10:24] like, you know, um style and [10:27] understanding how do you want to tell [10:28] your story. Um but hey that's that's AI [10:31] we're dealing with safety and security [10:34] issues and how to create a safeguards [10:37] right safeguards on a daily basis. [10:39] >> Yeah and re for you u you know [10:42] screenwriters actors I think are worried [10:43] about you know use of their images or [10:45] use of their content [10:46] >> being fed into these models for for [10:48] training. I'm curious what you know your [10:50] thoughts are on that and sort of what [10:52] the how the industry whether or not the [10:54] industry has done enough to sort of [10:55] protect uh [10:56] >> yeah I mean I'm not you know I would I [10:58] would definitely ask Jeffrey more than [11:00] me but as far as that goes but with I [11:02] can only speak on my experience with [11:04] Echo Hunter one of the things was again [11:06] with this SAG and Astra um contracts was [11:09] that technically my model that they made [11:12] was is destroyed after we do the movie [11:15] and part of that is even in promotion [11:17] and there were certain aspects of it [11:18] where I was like you if it's not from [11:20] Echo Hunter, you can't take my face and [11:22] put me on a dog food ad or whatever or [11:25] or a cat food ad. But um so I mean my [11:29] that was my specific experience with it [11:32] was okay well they're going to destroy [11:33] the model because you do have that [11:34] thought of all right there are people [11:36] who can put into AI give me breer you [11:39] know eating spaghetti with you know [11:41] Gwennneth Paltro. I don't know if that's [11:43] a great movie by the way don't go see [11:44] that movie but it's possible. Um, but [11:48] it's not going to be me. It's that [11:49] Uncanny Valley thing where you'd look at [11:51] it and be like, "That's not really, you [11:52] know, me." [11:53] >> That That's a good segue because I mean, [11:55] I watched all of the Echo Hunter and [11:56] there, you know, you there are parts [11:58] where you can sort of tell it is AI. I'm [12:00] I'm curious [12:01] >> at what point this uh the the content is [12:05] so good, the AI is so good that that we [12:07] would be past the uncanny valley aspect [12:09] and like things are actually fully [12:11] realistic. [12:12] >> Yeah, I think I think uh for us, Echo [12:14] Hunter was a proof of concept. We wanted [12:16] to be there as early as we can and to [12:18] prove something and and to show that we [12:20] can do that. And again, it's moving [12:22] rapidly. We're talking about almost four [12:25] and a half months ago. Today, we can cut [12:27] it in like 90% of the time um and do [12:30] actually a better results. And that's, [12:31] you know, the interesting thing in in AI [12:33] that it's getting better on a weekly [12:35] basis. What we're trying to do in Arcana [12:37] is assess, do the right diligence and [12:39] embed the right platforms into the [12:41] ecosystem so creators can get the most [12:44] updated software and capabilities to [12:46] start working and create. That being [12:48] said, I I think um going back to your [12:51] question is like we try to control as [12:54] much as we can. Motion capturing is [12:57] still not 100% um we're not talking [12:59] about five or 10 years. We're talking [13:01] about like 6 to 18 months to get there. [13:04] Controlling AI characters is still not [13:07] 100%. Okay. If you want to blend between [13:11] real uh footage and an AI the same [13:13] frame, it's still challenging. [13:16] >> Uh but we're getting there. We're [13:18] definitely getting there. [13:19] >> And just this is more from myification [13:21] if you want to dumb things down a little [13:23] bit. The difference, what is the [13:24] difference between sort of using AI to [13:26] generate something and and CGI? Mhm. [13:29] >> So, put aside the cost, the time, the [13:31] resources that you need. Basically, CGI, [13:34] you embedding a character in the frame [13:36] sometimes with like green dots or any [13:38] any other things and and you touching [13:41] between the characters and embedding if [13:43] you want Game of Thrones, you seeing the [13:45] head of the dragon, they're actually [13:47] using a a big huge head of dragons and [13:49] then they can touch something. In AI, [13:51] everything should be embedded virtually [13:54] in the frame. And this is one of the [13:55] challenges, by the way. Um, and we're [13:57] not encouraging anyone to do 100% AI. [14:00] >> Yeah. [14:00] >> The right blend could be anywhere that [14:02] works for the creators. It could be 70, [14:04] 30, 50, 64, whatever works for you. But, [14:08] but it's definitely getting there and [14:10] it's filling the gaps very fast. [14:13] >> And Bre, you talked about, you know, [14:14] your work on Robot Chicken. Curious, uh, [14:16] if your experience on Echo Hunter Yeah. [14:19] >> has inspired you to maybe take a look at [14:21] Arconer for for Robot Chicken. [14:22] >> Absolutely. We we've actually already [14:24] met with Arcana talking about a new [14:25] project that we're a follow-up to robot [14:28] and one of the things I noticed was [14:30] working with Cavin the director who has [14:32] worked with AI a lot and done ads and [14:35] commercials and I was totally new to it. [14:38] Um one of the things that was [14:39] interesting was the way Caven has to [14:41] talk to me, the way a director has to be [14:43] able to talk to an actor is similar to [14:45] how you also it's something you have to [14:47] learn. You have to learn personalities [14:49] and how you deal with certain actors and [14:50] whatnot. And this it's similar to [14:52] learning prompts now which is learning [14:55] the specific things. There is an art to [14:57] it which is you could say to it give me [15:00] a hamburger or you could say give me a [15:02] Wolf Gang Puck style hamburger with an [15:04] egg on top and blah blah blah blah blah [15:06] blah and things like that and it's a [15:08] learning curve and it's just the same [15:09] with CGI when it first came out and [15:11] products like that. And again robot is [15:13] still I mean we are the most antiquated [15:15] show in the world cuz we are still stop [15:17] motion. We are still click move the leg, [15:19] click, move the leg. And it's tedious, [15:22] you know. It's all [15:23] >> It's why people love the show though. [15:25] That's the thing. [15:25] >> It's why we and we we love the show. The [15:27] one thing is it takes us a year from [15:29] script to screen. [15:30] >> Wow. [15:31] >> And when you're doing a show that's pop [15:33] culture with references, you want to [15:35] make sure they're evergreen or current. [15:37] And having nine months is tough. Um, so [15:40] it definitely helps us streamline and [15:41] stay relevant. Um, but I think that's [15:44] another tool that an artist has to learn [15:46] now because it's a new tech are these [15:49] prompts and there is an art to it and [15:51] there is a specificity to it that I [15:53] haven't mastered yet but I've dabbled [15:54] and it's pretty exciting. But I also [15:57] think there's you know I don't know if [15:58] you guys play video games but the [15:59] cutscenes you know cut scenes in video [16:01] games you're like all right it's not [16:03] real doesn't sometimes I want to hit A [16:05] and get through it. And I think there's [16:07] still that with AI where that uncanny [16:09] valley thing where you're like I can [16:11] still tell it's not real. I can still [16:13] tell it wasn't written by somebody who [16:16] knows humor very well. [16:17] >> And so I think there's still a long way [16:19] to go with it. But I do think it's it [16:22] >> it's important to adopt it and learn it. [16:24] I mean I think it's another tool for [16:25] creativity that I didn't have as a kid [16:28] and I wish I did. You know I'm glad that [16:30] my 14-year-old can mess with it a little [16:32] bit on my computer. Well, I mean to [16:35] guy's point, I think that the this [16:36] hybrid model of sort of incorporating [16:38] traditional techniques with AI makes a [16:40] lot of sense. I'm just curious from a [16:41] from a time frame perspective because [16:43] that kind of blew me away that you said [16:44] it takes about a year for this stuff to [16:45] get produced. [16:46] >> How with Arcana's war technology with AI [16:49] in general, like how much can you [16:50] actually shorten that time frame? [16:52] >> A year is fast. By the way, a typical [16:54] feature is like seven to 14 years in [16:56] >> Oh, yeah. I mean, and it's, you know, I [16:58] did King of the Hill for eight years, [17:00] and King of the Hill was a year from [17:02] script to screen because it's got to get [17:03] sent away to animate and come back. [17:05] >> So now, I mean, you could tell better [17:07] than I was Arcana. I assume it'd be [17:09] shorter [17:10] >> days. I'm talking about days. And [17:11] obviously, we need to to discuss what [17:13] type of genre, what what's the length, [17:15] but but you can envision what you want [17:17] to in seconds, and then pick and choose, [17:20] mix and match. I like that. I don't like [17:21] that. I want to shoot this one outside [17:23] or I want to use AI. [17:24] >> Interesting. And uh going back to sort [17:27] of the the workflow like you said at the [17:28] beginning you sort of integrate a bunch [17:30] of different models technology. I'm just [17:32] curious what how that what's the [17:34] advantage of that versus say you know [17:36] signing on to just a single comp single [17:37] AI company kind of doing your work for [17:39] you. So there's tons of engines, models, [17:42] different type of tools with AI today, [17:44] right? And each one, first of all, [17:46] they're moving so fast and developing so [17:48] fast. So it's changing on a weekly [17:49] basis, but each one is doing some thing [17:52] greater than other models. [17:53] >> This one is great in action scenes and [17:55] and and you know, explosions. This one [17:57] with skin tone and and face capturing. [18:01] We at Arcana try to bridge everything, [18:04] embedding or pouring 25 or actually even [18:06] more tools into one ecosystem. The flow [18:09] is the flow of of filmmakers of creators [18:12] exactly how to start exactly as Breaken [18:14] mentioned starting from a prompt or any [18:16] image or just take an image on the scene [18:18] and embedded a video on that. So any [18:21] type and flexibility is existing in the [18:23] platform and we're making it easier for [18:25] creators to just sit down in one suite [18:27] of tools and create everything from [18:29] scratch up to the end up to sound. [18:31] >> So everything [18:32] >> Wow. No. Um, and I I think I'm sorry. I [18:35] think there is a part of [18:37] >> creativity that yeah, you can do a bad, [18:41] you know, a very kind of boring action [18:44] movie that's just by the books, blah, [18:45] blah, blah. Or you can get a great [18:48] writer to create a really great script [18:49] and great actors and great producers. [18:51] And I think similar here, which is you [18:53] could just say, "Hey, make me this, but [18:56] it's going to look like that. It's going [18:57] to look lazy." Yeah. [18:58] >> And it's not going to be anything [18:59] exciting. And then you have the other [19:01] version which is the hybrid which is no [19:02] you put creativity involved in it. You [19:04] don't just send it off on its own. You [19:07] stay involved and you stay creative with [19:09] it. You get something exciting like I [19:11] think Echo Hunter definitely had that. [19:13] >> You can't skip the storytellers. You [19:14] can't skip that. There's no shortcuts [19:16] here. [19:17] >> Just sort of looking ahead. I mean you [19:19] sort of mentioned with Echoh Hunter it [19:21] took a couple months now it takes a [19:22] couple days. The rate of advancement is [19:24] tremendous here. [19:26] >> Looking out about five years which is [19:28] which is hard to say. I know 5 years is [19:30] a long time from now. Like what what are [19:32] what do you sort of see as the [19:33] capabilities the the opportunities when [19:35] it comes to AI and entertainment? [19:36] >> I think the opportunities will [19:38] definitely going to open a lot of new [19:40] positions and roles to people. Anyone [19:42] can start creating. I think uh the next [19:44] Steven Spielberg is could be sitting [19:47] here and and doesn't have the means to [19:49] start creating and right now he has [19:51] everything that he needs. you know uh [19:52] we've been in a few conferences it's [19:54] interesting and a few um executive [19:56] approach with like you know shining eyes [19:58] and said you know that was my dream to [20:00] create something and I couldn't afford [20:01] even to buy a camera right [20:02] >> now I can do actually a feature yeah [20:05] >> by myself [20:06] >> and that's that's exactly what what [20:07] that's that's exactly the business that [20:09] we are into right now and I think moving [20:11] forward it's going to be easier more [20:13] adjusted to the needs people can create [20:16] more things faster and dedicated the [20:20] specific content to their needs And [20:22] that's telome or like you know that [20:25] that's a capability that we're going to [20:26] see a lot. [20:27] >> Okay, what do you think? Five years from [20:30] now, it's a long time. I get it. [20:31] >> Five years from now, I'll be the new [20:33] Batman [20:36] heard here first. [20:37] >> Exactly. [20:37] >> Batman 14. [20:38] >> Yeah, it'll be Batman Pie. And uh No, I [20:41] I think the same thing. I I almost can't [20:44] fathom how because of how fast it moves [20:46] as far as five years. Um, but I think [20:50] it'll I think it'll be a platform that [20:52] again, like the the 17-year-old kid just [20:56] out of high school who is really [20:57] brilliant and wants to create will have [20:59] an avenue to create that doesn't require [21:01] him having $20 million or a pitch deck [21:06] that he can't create and things like [21:08] that. I mean, that's what it really [21:09] helps streamline is you can now show [21:12] someone exactly what you're thinking [21:13] instead of just telling you an idea that [21:15] now I have to hope when you when you [21:17] pitch something, the idea is I'm going [21:19] to pitch to this executive and my only [21:21] hope is that they retain enough of my [21:23] idea to translate to their boss who I [21:26] don't get to go in to see and that [21:29] they'll then go, "Okay, that's a good [21:30] enough idea." as opposed to being able [21:33] to present them a deck that I've created [21:35] with the help of Arcana and say this is [21:37] exactly what you'll be getting. This is [21:39] the exact story. So I don't have to [21:41] worry about it being lost in translation [21:43] which is a good movie but um I don't [21:45] have to worry about that and hopefully [21:46] in 5 years it's even kind of better and [21:48] even stronger and giving more [21:50] opportunities to creative people. Yep. [21:52] >> More assisting them. [21:54] >> Five years. It's a lot in AI. [21:55] >> Yeah, that's a long time in AI. Well, [21:57] thank you both for your time. Really [21:59] appreciate this chat. We wanted to wrap [22:01] this up with another short video from [22:02] Arcana to show a sort of a proof of [22:04] concept of some of the other things you [22:06] can do with this uh with this powerful [22:07] engine. So, cue the last video. [22:22] [Music] [22:34] Are you looking? [22:37] [Music] [22:48] Heat. Heat. [22:53] [Music]