World Labs发布Frameboy，支持实时探索AI生成世界
动察 Beating AI 报道称，World Labs发布Frameboy，用户可通过键盘在AI生成虚拟世界中实时控制镜头探索，最高15 FPS。公开版目前仅提供预设场景。
Loading…
bibibi.ccbibibi.ccLoading…
Connect scattered updates into a story you can understand.
动察 Beating AI 报道称，World Labs发布Frameboy，用户可通过键盘在AI生成虚拟世界中实时控制镜头探索，最高15 FPS。公开版目前仅提供预设场景。
北卡男子James Gates起诉OpenAI及Sam Altman，称ChatGPT助长妻子投毒计划。其妻曾被控谋杀未遂，但检方于2026年4月以证据不足撤案。
谷歌发布 Gemini Omni 1.2 Flash，最长可生成15秒视频，720p生成价格降至每秒0.07美元，并已接入部分产品。
PANews称，Google AI Studio在X平台宣布Gemini Omni 1.2 Flash已上线AI Studio和Gemini API，速度更快且成本降低30%。
据IT之家报道，高通宣布AI Hub向中国开发者开放，提供预优化模型、云端设备调优、示例应用及端侧推理框架等资源。
Sabi discloses RLNF research: combining EEG, eye movements, and existing ratings to fine-tune an image model through reinforcement learning to generate faces that better match individual aesthetic preferences. The experiment included only 5 participants, and EEG recognition results still require further validation.
Google DeepMind releases its new-generation world model, Genie 4, which can generate explorable virtual worlds and simulate sound, dialogue, and physical behavior; it is currently available only to a small number of invited testers.
Google DeepMind has announced the launch of its next-generation world model Genie 4 for real-time interactive simulations, now available to a small number of trusted testers.
AI Creator Sac said that they used Opus 5.5 and other tools to complete, within 16 days, a browser-based 3D The Little Prince story game, which is now available for free trial.
The report says that the Epic–Disney shooting game was delayed from its scheduled 2027 release, with no new date announced for now, and reportedly requires the team to fully adopt a March workflow.AI
According to reports, Pan Tianhong, founder of Filmkraft, demonstrated at the event a model that can generate a virtual environment from text, in which users can roam and make modifications. ByteDance has not confirmed its relationship with the project reported previously.
谷歌开放 SynthID Detector，公众可检测图片、视频和音频中的部分AI水印，并支持OpenAI、英伟达和Kakao等厂商。工具每日约提供10次检测，但无法识别未采用SynthID的AI内容。