Recently, researchers from ByteDance Research Institute and Tsinghua University jointly released a new study, pointing out that current AI video generation models, such as OpenAI's Sora, while capable of creating stunning visual effects, have significant flaws in understanding basic physical laws. This study has sparked widespread discussion on the capabilities of AI in simulating reality. The research team tested AI video generation models under three different scenarios: predictions under known patterns and predictions under unknown patterns.

Recently, Tim Brooks, the research head of OpenAI's Sora video generation system, announced on social media that he will be joining Google DeepMind to continue his work on video generation and the world simulator. Brooks expressed his excitement about collaborating with a talented team like DeepMind and thanked OpenAI for the wonderful two years he spent there, especially for the achievements made in the Sora project.

Recently, OpenAI has caused a stir in the video AI field as they are set to make significant upgrades to Sora, which was released in February of this year. According to reports from the Information Daily, the main goal of this upgrade is to significantly enhance Sora's performance, enabling it to generate longer and higher quality video content more quickly. Reflecting on the initial performance of Sora, there were indeed many pressing issues that needed addressing. The original version took over 10 minutes to generate a video, an efficiency level that clearly fails to meet practical application needs.

OpenAI is currently working on upgrading its video AI Sora, a project that first appeared in February this year. According to sources close to the company's management team cited by The Information, OpenAI is developing an upgraded version of its video AI Sora, aimed at producing longer and higher quality video clips more quickly than the original demonstration. Notably, Sora was not fully ready at its launch; reportedly, the initial video generation time exceeded ten minutes.

OpenAI's Sora Video AI model is currently in the research stage, with its original release date delayed. This is mainly due to ongoing communications with policymakers, cautious considerations regarding safety risks, and technical challenges leading to high costs. Sora faces legal risks concerning training data and it is yet to be confirmed whether YouTube data has been used. Additionally, OpenAI is actively collaborating with Hollywood, testing the model with several individuals, including Ashton Kutcher, to explore its potential applications in the film industry.

OpenAI has selected TSMC's 1.6nm chip technology as its primary supplier, marking a significant strategic upgrade in its AI hardware sector. Mass production is expected to begin in 2026, which will significantly enhance its technological performance. In collaboration with Broadcom and Marvell, OpenAI plans to develop ASICs dedicated to AI applications based on TSMC's 3nm process series in 2024 or 2025. Additionally, Apple plans to integrate OpenAI's AI solutions into the new iPhone and may.

According to reports from Titanium Media, the ByteDance team plans to unveil significant advancements in their latest artificial intelligence model technology on July 19th (this Friday) for the first time on a large scale. The content to be disclosed includes new AI models such as text-to-image and Sora-like video generation, with particular emphasis on innovative technologies in long videos and high-dynamic directions, which are expected to directly compete with OpenAI's Sora text-to-video mode

In the digital media era, video has become our primary means of self-expression and storytelling. However, creating high-quality videos typically requires professional skills and expensive equipment. Now, with Snap Video, you can automatically generate videos just by describing the scene you want in text.The current image generation models have already demonstrated excellent quality and diversity. Inspired by this, researchers have begun to apply these models to video generation. However, the hi

Recently, the AI video generation platform Runway announced that their latest product, Gen-3 Alpha, has been opened for testing to a select group of users. Compared to the previous generation, Gen-3 Alpha has achieved significant improvements in aspects such as lighting, quality, composition, text semantic restoration, physical simulation, and action consistency/continuity, making it a product that can rival OpenAI's Sora. If you are also interested in Gen-3 Alpha and would like to experience it

Recently, the Lu Chen Open-Sora team achieved breakthrough progress in 720p HD video quality and generation speed. They not only made headlines with their advancements in 720p HD video quality and generation time but also open-sourced this technology, causing a sensation in the entire community!

Sora's popular short film 'Balloon Man' has sparked controversy regarding human participation in post-production effects. Issues with video consistency: Sora cannot ensure consistency of subjects across different shots. Key to video post-production: Sora-generated materials require human cutting and speed adjustments. Internet users expressed dissatisfaction with OpenAI's marketing approach, claiming it to be misleading.

OpenAI's Sora generates 5 minutes of video per hour on the NVIDIA H100, allowing for 120 minutes of video creation daily to support the TikTok and YouTube creator community, requiring about 720,000 NVIDIA H100 GPUs. Sora will target visual artists and filmmakers after its launch, with plans to incorporate sound and editing flexibility. OpenAI is promoting Sora in Hollywood, allowing creators to generate multiple candidates before uploading videos.

Sora will open for public testing within the year. Detailed insights into Sora's technology, generation speed, and video sound effects. Red team testing is ongoing, inviting creators to cooperate in driving industry development. Future versions of Sora are expected to support video sound effects, enhancing user experience.

OpenAI has showcased its video generation AI system Sora to Hollywood studios and directors. Sora may be publicly released this year, sparking Hollywood's acceptance of AI technology in film production. The use of AI-generated imagery has prompted debates over job roles and authorship, which continue to evolve. The demonstration of Sora has garnered industry attention, and Hollywood's attitude towards AI technology in filmmaking remains to be seen.

Sora has several fatal flaws and cannot perfectly generate video content. Sora's training data comes from Shutterstock, with a data collaboration agreement with OpenAI. Sora has immense computational demands and will not be released to the market yet, aiming to reduce these requirements. Ultimately, Sora will generate videos with sound, with safety being a key focus area.

The Tiantu Wanjing team has launched a universal video and audio tool called Sora Opera, which allows users to lift the 'mute mode' of videos. Users can add music and sound effects to the generated videos. However, a membership is required to access the Sora Opera plan, which provides video dubbing and audio effect features. The cloud version will initially be available to partners. The founder showcased several demonstration videos of Sora Opera, which supports dubbing for videos up to 60 minutes long. Sora Opera also offers several unique features.

The Stable Diffusion 3 model introduces the DiT (Diffusion Transformer) architecture, which significantly improves quality. Stable Diffusion 3 outperforms other text-to-image generation systems. The parameter count of Stable Diffusion 3 ranges from 800M to 8B. The SD3 architecture is developed in collaboration with core members of Sora and an assistant professor from New York University.

["Sora is frequently making headlines", "LTX Studio released by Lightricks", "New technology products"]

Open-Sora has released a complete reproduction architecture scheme for Sora, which can reduce the reproduction cost by 46%. The model training input sequence length is extended to 819K patches, supporting dynamic resolution, various model structures, and multiple video compression methods. The performance improves by over 40% compared to the baseline scheme, while also reducing costs.

The emergence of the Sora model is driving a transformation in the text-to-video field. The development of the Sora model took nearly a year and involved multiple explorations in different directions. The Sora model utilizes the DiT technical framework, overcoming the limitations of diffusion models. Seven Volcanoes has established cooperative relationships with ByteDance, Xiaomi, and Kuaishou through the Lava platform, promoting both local and international content development. Supercom Communications has strategically invested in Seven Volcanoes, becoming its largest single shareholder with a 30% stake.