Babbitt | Metaverse Daily Must-Read: WWDC2023 is approaching, Apple MR details continue to emerge, the media said that domestic OEMs are required to start work before July, and the first year’s shipment is expected to be less than 500,000 units

Abstract: According to the "Kechuangban Daily" report, expectations for Apple's MR devices have been rising in the A-share market recently. Although it is still unknown whether the product will be unveiled as scheduled, the entire industry is undoubtedly looking forward to it. According to a report by The Information on May 31, Apple's mixed reality device may be the most complex hardware product the company has ever made, and it uses an unusual design-curved surface design. The design is said to be the main reason the device retails for as much as $3,000. The Information said that although mass production of the device has not yet begun, Apple has asked Luxshare Precision to start construction by July. This suggests that the device will officially go on sale in the fall or winter, with fewer than 500,000 units expected to ship in the first year.

Image source: Generated by Unbounded AI tool

Hot info:

Renault Korea plans to create a metaverse car experience showroom

According to Sina VR reports, recently, Renault Korea announced that it plans to establish and operate a car experience metaverse space called "Renault Korea Center" this fall. The metaverse space will be operated by K-verse, which has already settled in the metaverse platform The Sandbox, providing services exclusively for Korean brands.

According to Renault Korea, the Renault Korea Center is The Sandbox's first complete car project and is expected to provide a metaverse space where visitors can enjoy various experiences and games, such as designing their own vehicles. In addition, visitors can experience Renault cars realized in voxel (volume and pixel) format, learning the story behind each car.

XiaoIce’s first batch of AI clones of GPT Internet celebrity stars went online ahead of schedule, including “Hanzo Forest” and others

According to news from IT House on June 1, two weeks ago, Xiaoice announced the launch of the "Human Clone Project" in the Chinese and Japanese markets. Create AI clones based on your personality, skills, voice, and appearance. According to the latest official news, at present, the first batch of Internet celebrity clones have been launched ahead of schedule, including the previously controversial "Hanzo Forest". In this regard, Xiaoice believes that under the premise of strictly abiding by the red line of the law, everyone can have the opportunity to start a second life with AI cloning. Xiaobing said that at present, the total number of Internet celebrity fans who have signed up for clones in the background has exceeded 500 million. However, for safety reasons, the first batch of clones is still limited to 300 people. After testing and verification, it is expected to be gradually released to a larger scale before the end of the year. Xiaoice plans to unlock new clones every Thursday.

In addition, in order to avoid hidden dangers caused by improper use of clones, Xiaoice Company does not open technical interfaces, but restricts clones to communicate with users only in specific X Eva Apps. The technologies used at the bottom of X Eva App include Xiaoice XLLM language model, sound and neural network rendering technology, etc., so that clones have both personality and the ability to complete work.

British Prime Minister: I hope to make the UK a global AI center and will lead the development of "safe and reliable" rules

British Prime Minister Rishi Sunak wants the UK to become a global hub for artificial intelligence, promising to play a "leading role" in setting "safe and secure" rules to boost the industry, the Financial Times reported on May 31. Sunak said he would discuss AI regulatory "guardrails" with U.S. President Joe Biden in Washington next week, an issue that must be addressed internationally. Last week, Sunak met with the CEOs of leading artificial intelligence companies OpenAI, Google DeepMind and Anthropic, saying they will work together to ensure society benefits from artificial intelligence technology, and discussed the risks posed by AI, involving disinformation, national security and existential threats. British officials believe the UK could promote a regulatory model that would not be as "draconian" as the approach taken by the European Union but would be stricter than any US framework.

a16z-backed AI startup Character.AI releases mobile app, tops 1.7 million installs in first week

On May 23, a16z-backed AI startup Character.AI launched a mobile app that has attracted more than 1.7 million new installs in less than a week, TechCrunch reported. The mobile version of the AI chatbot platform is said to be rolling out to iOS and Android users globally, with strong performance on Google Play in particular. The app surpassed 700,000 Android installs in its first 48 hours, putting it ahead of top entertainment apps like Netflix, Disney+, and Prime Video. Before the launch of the app, Character.AI’s web app received more than 200 million monthly visits, and users spent an average of 29 minutes per visit, a figure the company says is 300% higher than ChatGPT, Character.AI said. As previously reported, AI start-up Character.AI received US$150 million in financing led by a16z.

Alibaba Cloud Intelligence releases audio and video AI Tongyi Hearing, which will be officially tested on June 1

According to a report by the "Kechuangban Daily" on June 1, Alibaba Cloud Intelligence released Tongyi Tingwu today, and it will be officially tested on June 1. Zhou Jingren, CTO of Alibaba Cloud Intelligence, said that Tongyi Tingwu is an AI assistant based on the large model of Tongyi Qianwen and the audio and video AI model. Regular and so on.

The first China Digital Art Exhibition is open to the whole country, covering AIGC and other works that are deeply integrated with digital technology

Babbitt News, on May 31, the China Artists Association issued the "Notice of Call for Papers for the First China Digital Art Exhibition". The notice pointed out that the China Artists Association, Zhejiang Provincial Federation of Literary and Art Circles, and China Academy of Art will jointly host the "First China Digital Art Exhibition" at the end of December 2023. This exhibition aims to sort out the creative achievements of contemporary Chinese digital art, highlight the social value of digital art, broaden the field of digital art, and stimulate the new energy of art development that drives social innovation through artistic innovation.

The exhibition is now open to nationwide collection of works. The categories of works include but are not limited to CG paintings, digital images, interactive art, mixed reality (AR, VR, MR), generative art/artificial intelligence art/AIGC and other works that are deeply integrated with digital technology.

Fudan University MOSS system leader: There will be plans to launch MOSS models with greater computing power and stronger reasoning capabilities

According to a Shanghai Securities News report on June 1, Qiu Xipeng, the person in charge of the MOSS system at Fudan University, said "the performance of MOSS after running for more than a month" said that the MOSS system currently ranks high in downloads in the open source community, and the following users on the GitHub platform It has exceeded 10,000 people.

Qiu Xipeng said, "Up to now, the MOSS background has received applications from hundreds of companies, most of which are small and medium-sized enterprises, covering traditional industries, intelligent information services, finance, medical care, education, etc. At present, the main functions of the MOSS model are centralized It can help companies apply large models to specific fields at a lower cost." In terms of feedback from the use of the MOSS model, Qiu Xipeng said bluntly that the current chat function of the model has achieved a good interactive experience, but it still needs further improvement in dealing with complex problems. Complete. Qiu Xipeng revealed that Fudan will plan to launch a MOSS model with greater computing power and stronger reasoning capabilities.

HKUST Xunfei: Large-scale model user data and license information cannot be disclosed now, and the current computing power can meet large-scale model training

According to a report by China Business News on June 1, iFlytek stated on the situation of the large model that the user data of the large model of Spark and whether it has obtained a license are core information and cannot be disclosed. When asked whether the chip and computing power can support the current large-scale model application, the staff said that the company now has relevant cooperation with domestic mainstream chip manufacturers, and the current computing power of Xunfei can meet the training of large models.

Investment and Financing:

Google invests in artificial intelligence startup Runway

Google's investment in New York-based artificial intelligence startup Runway, which includes the new funding, values the company at $1.5 billion, according to two people familiar with the matter, The Information reported on June 1. Runway lets users use artificial intelligence to generate videos from text descriptions. The investment, part of a funding round of about $100 million, underscores fierce competition among cloud providers for access to companies with cutting-edge artificial intelligence services that could become major cloud customers or acquisition targets in the future. Amazon Web Services (Amazon Web Services) had touted Runway as a significant customer for artificial intelligence startups, but Runway is now expected to lease cloud servers from Google, a person familiar with the deal said.

Informed sources: AI start-up MiniMax is close to completing over US$250 million in financing, and Tencent affiliated entities participated in the investment

According to Reuters news on June 1, people familiar with the matter said that MiniMax, a Chinese AI startup focusing on general-purpose large-scale models, is close to completing a round of financing of more than 250 million U.S. dollars, and its valuation is expected to reach nearly 1.2 billion U.S. dollars. The company is developing an AI solution similar to ChatGPT. The latest funding round for MiniMax has attracted new investors, including entities linked to Tencent, according to two people familiar with the matter. It is reported that MiniMax attracted Mihayou to participate in the previous financing. Other early investors include Yunqi Capital and Mingshi Capital.

Metaverse recommended reading:

"Apple MR details have emerged or become its "most complex hardware in history", and the industry and commerce have been required to start work before July"

According to The Information on May 31, Apple's mixed reality device may be the most complex hardware product the company has ever made, and it uses an unusual design - curved design, which has proven to be An unprecedented challenge in history. The design is said to be the main reason the device retails for as much as $3,000. It is worth noting that, according to The Information, although the mass production of the equipment has not yet started, Apple has asked Luxshare Precision to start construction by July. This suggests that the device will officially go on sale in the fall or winter, with fewer than 500,000 units expected to ship in the first year. As of the close at noon today, Luxshare Precision rose 4.03%, the largest single-day increase since April 13.

"Solving the biggest obstacle to AI landing, OpenAI has found a solution? "

OpenAI seems to have found a solution to the "serious nonsense" of generative artificial intelligence. On May 31, OpenAI announced on its official website that it has trained a model that can help eliminate "illusions" and other common problems common to generative AI. OpenAI says that reward models can be trained to detect hallucinations, and reward models are further divided into outcome-supervised (provide feedback based on the final outcome) or process-supervised (provide feedback for each step in the chain of thought) model. That is, process supervision rewards each correct step of reasoning, while outcome supervision simply rewards correct answers.

"Game Manufacturer, all in AI"

Since the launch of ChatGPT, what kind of imagination AIGC will bring to the game industry has been a hot topic in the market. Although the final consensus has not yet been reached, one thing is for sure. The fastest speed is applied in every link of game industrialization development. The combination of AI and games is regarded as one of the scenarios where commercialization will land faster, and it has also become an important direction for the future development of more game manufacturers.