Sora

Kuaishou's "Keling" video generation model is officially launched. Officials said that the Keling model was developed by the Kuaishou AI team. It uses a similar technical route to Sora, and claims to combine multiple self-developed technical innovations and benchmark Sora's effects. According to reports, the Keling model has the following advantages: It can generate large-scale reasonable movements and simulate the characteristics of the physical world. It has a strong ability to combine concepts and imagination. The generated video resolution is 1080p, the duration is 2 minutes (frame rate 30fps), and it supports free adjustment of aspect ratio. According to Kuaishou, the Keling model has been used in the Kuaiying App...

The Tribeca Film Festival in New York, founded by the famous actor Robert de Niro, is about to usher in a historic moment. Generative AI films will be on the stage of this 20-year-old film festival for the first time. On June 15, the Tribeca Film Festival will work with OpenAI to premiere "Sora Short Films", which will screen five short films created using artificial intelligence. This marks the first time that OpenAI's text-to-video tool Sora will appear at a film festival...

Recently, OpenAI's Vincent video model Sora will be on display at the upcoming Tribeca Film Festival, showcasing five short films produced by AI. These five films were produced by famous directors such as Bonnie Discepolo, Ellie Foumbi, Nikyatu Jusu, Reza Sixo Safai and Michaela Ternasky-Holland through Sora. This is also one of the few cases where a film produced entirely by AI has been presented at a film festival...

OpenAI launched the text-to-video Sora three months ago, sparking widespread discussion among netizens, the media, and insiders. At the 2024 I/O Developer Conference held today, Google also launched a benchmark product, Veo, which can generate "high-quality" videos with a length of more than 1 minute and a resolution of up to 1080P, with a variety of visual and film styles. According to Google's official press release, Veo has advanced understanding of natural language and can understand film terms such as "time-lapse photography" and "aerial scenery". Users can use text, images...

Recently, blogger @maxescu shared his experience of participating in the internal test of an AI music app. He spoke highly of this AI music app, believing that it will be a groundbreaking product, and even compared it to Sora, believing that it will shock people again. According to @maxescu's description, the music generated by this AI music app is almost the same as the songs we hear on Spotify. His evaluation is undoubtedly the highest evaluation of this AI music app, and it also aroused my great interest. @maxescu even said that this AI…

"AI resurrection short video production, easily popular on all platforms, public and private domains dual monetization operations. This type of video gains fans quickly. Basically, Xiaohongshu can gain more than 10,000 fans a month. If Douyin does well, it can gain 10,000 fans in a week. And people with this kind of demand are willing to spend money to make videos." "Using AI to make deceased relatives appear in front of you again, even if it is only for a short minute, it can make up for the regrets of many people, so there should be a lot of people who have this demand, and now there are few people who know how to make this type of video, and the competition is relatively small." "You can make this type of video yourself...

Market research firm Factorial Funds recently released a report that OpenAI needs 720,000 Nvidia H100 AI accelerator cards at peak times to deploy its text-to-video model Sora. If each Nvidia H100 AI accelerator card costs $30,000, 720,000 cards would cost $21.6 billion (currently about RMB 156.168 billion). And this is just the cost of the H100 accelerator card, which consumes a lot of electricity to run. Each H100 consumes 700W of power…

Recently, OpenAI invited some artists to try out its AI video creation tool Sora. By using Sora, the artists created a series of stunning experimental short films, showing the power and novelty of artificial intelligence. Unlike OpenAI's familiar chatbot ChatGPT and image generation platform DALL-E, Sora is not yet open to the public. On Monday, OpenAI published an article titled "Sora: First Impressions" on its blog, showing the visual...

OpenAI plans to hold a meeting with Hollywood studios, media executives and talent agencies in Los Angeles next week, according to Bloomberg. People familiar with the matter said the meeting plans to establish entertainment industry partnerships and encourage filmmakers to integrate its new artificial intelligence video generator into their workflows. OpenAI hopes to promote its AI video generation tool Sora to Hollywood, and the company's CEO Altman has attended several parties in Los Angeles during the Oscars weekend. The OpenAI Sora model can strictly…

Microsoft recently launched a video generation project called Mora, which aims to replicate Sora's capabilities to generate universal videos. Mora uses a multi-AI agent framework and integrates multiple cutting-edge visual AI agents to achieve the universal video generation capabilities demonstrated by Sora. Paper address: https://arxiv.org/html/2403.13248v1 Mora's main functions include: Convert text to video: Mora can generate videos related to the input text content. This means that users only need to provide...

After OpenAI launched the Sora model in February this year, it quickly became a headline in the AI circle. It can generate videos up to 1 minute long based on a user's sentence, and the effect is close to real shooting. OpenAI has not yet released Sora to the public, but developer @LinusEkenstam recently posted a series of tweets showing 9 new videos generated by Sora, including an elephant covered in leaves, a bee flying through flowers, and a rainbow-colored waterfall. Prompt: "An elephant made o…

OpenAI recently launched the hot text-to-video generation model Sora, but the company's Chief Technology Officer (CTO) Mira Murati was vague in an interview with the Wall Street Journal and could not clearly explain the source of Sora's training data. During the interview, when the reporter directly asked Murati about the source of Sora's training data, she only used vague official language to prevaricate: "We use publicly available data and licensed data." When the reporter asked whether the specific source included YouTube videos, Murati...

Meitu disclosed its 2023 performance report yesterday, achieving total revenue of 2.7 billion yuan for the whole year, a year-on-year increase of 29.3%. The adjusted net profit attributable to the parent company's equity holders was 370 million yuan, a year-on-year increase of 233.2%. According to Sina Technology, Wu Xinhong, founder, chairman and CEO of Meitu, said at the performance meeting that AI native workflows will be a huge opportunity in the future, and Meitu will be committed to further building AI native applications. At the same time, he also emphasized that Meitu will not make general large models, but Meitu will train and...

With the rise of AI video technology launched by domestic companies, the short drama market has ushered in new changes. Seven Volcanoes was the first in China to reproduce Sora's achievements and made major breakthroughs through innovative technologies such as the Etna model. The Etna model uses Diffusion Transformer to process video data, achieving ultra-high definition of 4K60 frames in 15 seconds, while also having the ability to understand space and time and deep semantics. Official website address: https://etna.7volcanoes.com/ Paper address: https://a…

Recently, the YouTube channel WVFRM Podcast invited the core team members of OpenAI's video model Sora to participate in an interview. The three members include Bill Peebles, Tim Brooks and Aditya Ramesh, all of whom are Sora project leaders. The three leaders talked about when Sora will be available to users. They said that Sora is still in the feedback acquisition stage, "It is not a product yet, and it will not be released to the public in the short term...

The Wall Street Journal published an interview with OpenAI’s chief technology officer, Mira Murati, on the 13th (today). She said Sora will be launched this year, “maybe a few months later.” Murati said that in addition to making Sora available to the public, OpenAI also plans to add audio generation capabilities, which may make the scene more realistic. In addition, the company plans to allow users to edit the content generated by Sora themselves, “because AI tools don’t always produce accurate results.” She said the company is trying to find ways to allow users to edit,…

Italy’s data protection agency (Garante) announced on Friday that it has launched an investigation into a service developed by Microsoft-backed OpenAI, called Sora, which can generate videos based on text prompts. The regulator asked OpenAI to clarify whether the way it informed users and non-users about the data used by its product Sora complies with EU regulations. OpenAI did not immediately respond to requests for comment. Italy’s data regulator is one of the most active in the EU and has been working to assess whether artificial intelligence platforms comply with EU regulations…

OpenAI recently released several videos generated by Sora, showing amazing 3D effects. These videos are rich in content, including diving to explore shipwrecks, stereo models depicting tranquil scenes from Japan's Edo period, etc. In addition, Sora also demonstrated some new features, such as seamless transitions. This means that interpolation can be gradually performed between two input videos, presenting a smoother transition effect. Some netizens began to discuss the source of the training data behind the videos generated by Sora. The mainstream view is that Sora may have used a 3D engine or Unreal Engine for training. Source Note: The picture is from…

OpenAI's Sora, a text-generated video tool released last week, has caused quite a stir, sparking discussions about how artificial intelligence will change the entertainment industry. However, Hollywood does not seem to be panicking. Sora can quickly generate relatively high-quality videos based on text, but industry insiders believe that such AI tools still have limitations and cannot completely replace the work of human creators. "A lot of people say that Hollywood is finished," said Reid Southen, a film concept artist and illustrator, "but I don't think so. Hollywood's production process is very complex...

OpenAI recently released a text-generated video model called Sora, which generates realistic videos based on user-provided prompts. However, according to user feedback, Sora takes more than an hour to render a 1-minute video. According to user feedback on the Reddit community, the researchers mainly showed pre-selected examples, and did not allow the public to access custom prompts. The longest demonstration video was only 17 seconds long. The screenshots are as follows: Others speculated on the potential reasons behind the long rendering time. One user mentioned OpenAI CEO…

Mikhail Parakhin, CEO of Microsoft's advertising and network services division, recently said in an interaction with netizens on Twitter that Sora, a video generator developed by OpenAI, will eventually be integrated with Copilot, but it will take some time. A user asked whether Sora would join Copilot, and Parakhin replied: "Eventually, but it will take time." Sora is a large language model chatbot developed by OpenAI. Users can enter any text prompt, and the AI model will generate up to...

Sora is able to generate realistic videos based on text prompts. The model can create videos up to one minute long, and the video resolution is up to 1080p. It performs well in handling reflections and shadows. It is not yet open to ordinary users because OpenAI is working with experts to conduct red team testing to assess the possible bias, risks and harms of the model. This article is divided into two parts. The first is Sora common sense, experience website, effect display, and the second is how to register Sora. 1. Common sense 1. What kind of company is OpenAI? Answer: Ope…

Sora's creative ability has shocked many people. The era of making your own movies has come. In the past, making a short video was enough to satisfy people, but now everyone can become a director. These days, everyone is waiting for when they can use it. I will share with them a method that may use Sora in advance. Register as an Open AI red team internal tester. Before the release of GPT-4 and DELL, there will be a group of people who participated in the red team test. They will have priority to use it, and it is free. Later, Sora, a model that consumes a lot of GPU, is destined to be unaffordable for ordinary people. Registration address: ht…

Today, there is news that before Sora detonated the Vincent video track, ByteDance in China also launched a disruptive video model - Boximator. Unlike models such as Gen-2 and Pink1.0, Boximator can accurately control the movements of people or objects in the generated video through text. According to Sina Technology, in response to this, ByteDance related personnel responded that Boximator is a technical method research project for controlling the movement of objects in the field of video generation, and it has not yet been implemented as a complete product, and it is far from the leading foreign video generation model...