Sora competitor, "Vidu", can create 16 seconds long HD video with 1080p resolution.

China's Shengshu Technology and Tsinghua University recently unveiled Vidu, a new text-to-video generator, at the 2024 Zhongguancun Forum in Beijing. Vidu is capable of producing 16-second video clips at 1080p resolution using a Universal Vision Transformer (U-ViT) architecture. This technology, developed in September 2022, enables the generation of realistic scenes with accurate physics, lighting, and detailed facial expressions. It also features multi-camera capabilities that allow dynamic transitions between various shot types within a single scene.

Vidu was demonstrated as a competitor to OpenAI's Sora, although it primarily generates shorter videos of 16 seconds compared to Sora's 60 seconds. Despite this, Vidu boasts an ability to create complex, surreal content and showcases impressive temporal consistency in its video output. However, when compared directly with Sora, Vidu's videos, while notable, do not yet match the visual fidelity and realism of OpenAI's generator.

The unveiling of Vidu marks significant progress in China's AI capabilities and suggests potential for further advancements. While it currently falls short of Sora in some aspects, its unique features and the ongoing development suggest that Vidu could see significant improvements in the future. This development underscores the vibrant and competitive nature of the global AI landscape, where continuous innovation leads to rapid advancements in technology.

Two Forbes 30U30 Founders Transforming Mental Wellness

Forbes 30 Under 30 winners founded Aura to solve the $100B problem - mental wellbeing. 

Aura has quickly grown to 8 million users & 100k+ paying subscribers, and attracted investments from legendary Silicon Valley VCs & executives from Spotify, Facebook, and Apple. 

The best part? You can invest now before the opportunity ends in a few days.

ChatGPT's Memory function is now available

OpenAI has finally begun rolling out its ChatGPT memory feature, which allows ChatGPT to carry over learnings between conversations. Users can track memories retained by ChatGPT in the settings → personalization tab.

ChatGPT's memory feature enables it to recall details from past conversations, enhancing its usefulness and personalization. This feature allows ChatGPT to retain specific preferences, such as formatting for meeting notes or personal interests like a user's business or family details. Users have control over what is remembered and can manage memories through settings, ensuring that only desired information is retained.

Memory is optional and can be disabled at any time. For those seeking privacy, there's a "temporary chat" option that doesn't use or remember any details shared during the session. Furthermore, memories are managed independently of chat histories, meaning deleting a chat won't erase its associated memories. Users can also instruct ChatGPT on how to use their data for model improvement or opt-out completely.

For enterprise and team users, memory enhances productivity by remembering work-related preferences and styles, which helps in drafting documents, coding, and preparing business reviews. Each GPT, such as those specialized in books or greeting cards, will also have its own memory, potentially requiring users to reiterate information shared with ChatGPT in these specific contexts. This feature ensures tailored assistance across different uses and platforms.

This cutting-edge robot is a Silicon Valley startup’s answer to labor shortages

Nvidia, a company that originally aimed to enhance video game graphics, has become a powerhouse in the AI industry with a market value of $2 trillion. Founded in 1993, Nvidia's remarkable growth is largely due to the demand for its GPUs, which are crucial for artificial intelligence technologies. CEO Jensen Huang, a key figure behind this success, believes AI will dramatically change the world. This belief was evident at Nvidia's annual developers conference, where Huang introduced Blackwell, the fastest GPU yet, claiming it's essential for future AI advancements.

During the conference, Nvidia showcased various applications of its technology. Pinar Seyhan Demirdag demonstrated how Nvidia's GPUs could generate virtual movie sets in real-time, significantly reducing costs for filmmakers. Similarly, Dr. Alex Snyder discussed how Nvidia's technology is instrumental in developing new protein-based drugs that could potentially revolutionize medical treatments. Additionally, Brett Adcock's startup, Figure, introduced an Nvidia-powered humanoid robot designed to address labor shortages, demonstrating the practical applications of Nvidia's AI technology.

Despite these advancements, there are concerns about the broader impacts of AI, such as job displacement and ethical considerations. However, Huang remains optimistic about AI's role in future progress and prosperity, envisioning a world where machines enhance human capabilities rather than replace them. This vision is shared by many in the industry who see AI as a tool for augmenting human potential rather than a threat.

Related: Major AI x Robotics developments this week.

Other stuff

• Apple Intensifies Talks With OpenAI for iPhone Generative AI Features

• Google’s new ‘Speaking Practice’ feature uses AI to help users improve their English skills

• Altman handpicked for Homeland Security's AI safety board

• Caught Between the US and China, a Powerful AI Upstart Chooses Sides

• In Race to Build A.I., Tech Plans a Big Plumbing Upgrade

• Podcast: WebSim, WorldSim, and The Summer of Simulative AI

• Creators of Sora-powered short explain AI-generated video’s strengths and limitations

• Humanoid robots are learning to fall well

• The seven lies of the AI expert who cited himself thousands of times on scientific papers

• Apple Rivals Retool to Challenge the iPhone and Vision Pro

• ChatGPT’s ‘hallucination’ problem hit with another privacy complaint in EU

Hope you enjoyed today's newsletter

⚡️ Join over 200,000 people using the Superpower ChatGPT extension on Chrome and Firefox.