Data intelligence industry innovation service media
——Focus on digital intelligence and change business
Entreprise's news
Li Bin responds to Weilai's price cut: It is impossible to satisfy everyone
On June 12, Li Bin, the founder, chairman, and CEO of Weilai, responded to the news that "the price of all NIO models has been reduced by 30,000 yuan." and suggestions. There are too many aspects to consider, and I am still deliberating on it at three o'clock today. Now is the most suitable time for release, and there must be some things that are not well thought out, and it is impossible to satisfy everyone.
NIO invests in wireless charging chip maker Yichong Technology
Recently, Sichuan Yichong Technology Co., Ltd. underwent industrial and commercial changes. Sichuan Innovation and Entrepreneurship Equity Investment Fund Partnership (Limited Partnership) withdrew from the ranks of shareholders, and Hefei Weilai Industrial Development Equity Investment Partnership (Limited Partnership) was added as a shareholder. Sichuan Yichong Technology Co., Ltd. was established in February 2016, the legal representative is Pan Siming, and the registered capital is about 178 million yuan. According to reports, Yichong Technology is a wireless charging chip manufacturer, focusing on magnetic resonance coupling wireless charging technology.
TSMC's advanced process capacity utilization rate has increased significantly
On June 12, according to Taiwan media reports, the industry reported that driven by Nvidia’s increase in AI chip production, the capacity utilization rate of TSMC’s advanced process has recently increased significantly, and the capacity utilization rate of 5nm has increased from more than 50% to about 70% to 80%; The capacity utilization rate has also gradually increased from the low level of 30% to 40% to around 50%.
Dongfangtong: It plans to realize AI recognition services such as intelligent voice reminders for fraud risks before the end of the year
On June 12, Dongfangtong stated on the interactive platform that the company’s TongGPT intelligent voice interaction system is mainly based on the to-be-launched TongGPT multi-modal intelligent interaction model to realize intelligent audio content recognition and other functions. Call reminders to achieve smooth communication with people, and plan to implement AI recognition services such as intelligent voice reminders for fraud risks before the end of the year.
Visual China launches AI-inspired drawing and screen expansion functions
On June 12, Visual China announced the launch of AI-inspired drawing (that is, text to generate images) and screen expansion (OutPainting) functions. For the former, users can generate four different styles of picture materials: photography, cartoon, 3D, and illustration based on text to match different usage scenarios. The latter is a creative tool that can help users expand more creativity. It only needs to extend the picture beyond its original boundary, and add the same style of visual elements, including shadows, reflections and textures. It can be expanded to any size, using Through language description, the original work creation can be brought to a new direction while maintaining the effect of the original pictures and texts.
Giant Network Shi Yuzhu: Plan to make a game with metaverse elements
On June 12, it was learned from Giant Network that Shi Yuzhu, chairman of Giant Network, recently participated in the 2023 Journey Carnival held in Nianhuawan Town, Binhu District, Wuxi City. In his speech, Shi Yuzhu shared the opportunity to return to the front line of R&D and participate in "Original Journey" in September last year, and bluntly said that the current public beta version is still a semi-finished product, and he hopes to polish the product in 12 months, and then push another one he is satisfied with. formal edition. He also revealed that in the future, after completing the research and development of the official version of "Original Journey", he plans to make a game with metaverse elements.
Apple files for Vision Pro trademark
The Tianyancha App shows that recently, Apple has applied for the registration of multiple "Vision Pro" and "Apple Vision" trademarks. The international classification involves website services, communication services, scientific instruments, etc., and the current trademark status is mostly under application. Not long ago, Apple officially released the Apple Vision Pro, its first mixed reality head-mounted display device, at the Worldwide Developers Conference.
Meta open source AI language model MusicGen, which can convert text and melody into complete music
Recently, Meta open sourced its AI language model MusicGen on Github, which is based on the Transformer model launched by Google in 2017. As the model name suggests, MusicGen is mainly used for music generation, which can convert text and existing melodies into complete compositions. In addition, MusicGen also supports the combined input of text and melody. For example, you can propose to generate "a light song" and at the same time ask to "combine it with Beethoven's "Ode to Joy".
The R&D team also tested the actual performance of MusicGen. The results show that compared with Google's MusicLM and other music models such as Riffusion, Mousai, and Noise2Music, MusicGen performs better on indicators such as the degree of matching between music and text prompts and the credibility of composition, and is generally slightly higher than Google. MusicLM level. Meta has licensed the model for commercial use and published a web application on Huggingface for demonstration purposes.
Online meeting software Zoom introduces new features based on generative AI, which can automatically reply to messages according to the user's tone
Recently, video conferencing software developer Zoom announced the introduction of new features based on generative AI technology in the software, and its AI models come from companies such as OpenAI and Anthropic. Zoom said that it will use generative AI technology to bring meeting highlights to users. This feature uses Zoom's large-scale language model to automatically organize the meeting points. The meeting host does not need to record the whole process, but only needs to send the meeting points to the participants through chat after the meeting.
In addition, Zoom also brings an automatic reply function based on generative AI. This function uses the chat writing function provided by OpenAI, which can automatically assist in drafting information that matches the user's tone and sentence length according to the context of the conversation, and provide users with reply suggestions.
Apple Vision Pro headset low configuration exposure: display, processor and sensor will shrink
According to news on June 12, following the release of the Vision Pro headset at the WWDC23 event last week, the latest news is that Apple is developing a lower-end version of the Vision Pro that is cheaper and more suitable for the mass market. Compromise on some fronts.
According to Bloomberg's Mark Gurman (Mark Gurman), this low-profile version of the Vision Pro may be called "Apple Vision" or "Apple Vision One." The device may differ from the high-end version in terms of processors, displays, and sensors to keep costs down. Through these measures, as well as optimizing the production process and economies of scale, the price of the low-end version may be reduced by "hundreds of dollars."
Policy/Report
The revenue of the top ten foundry manufacturers in Q1 decreased by nearly 20% quarter-on-quarter, and will continue to decline in the second quarter
Recently, according to TrendForce research, due to continued weak end demand and multiplied off-season effects, the revenue of the world’s top ten wafer foundry manufacturers in the first quarter fell by 18.6% quarterly, approximately US$27.3 billion. The biggest change in the ranking this time is that GlobalFoundries surpassed UMC to take the third place, and Gotower Semiconductor surpassed PSMC and World Advanced Technology, and reached the seventh place this quarter. TrendForce predicts that the output value of the top ten foundry manufacturers will continue to decline in the second quarter, and the quarterly decline will be smaller than that in the first quarter.
Ministry of Industry and Information Technology: Strengthen the top-level design of the open source system and build a higher level of open source infrastructure
News on June 12, the 2023 Open Atom Global Open Source Summit was held in Beijing on June 11. Zhang Yunming, Vice Minister of the Ministry of Industry and Information Technology attended and delivered a speech. Zhang Yunming emphasized that we should accelerate the construction of an open source system, release the potential of open source development, and promote high-quality development. Improve the systematic layout. Adhere to the national "one game of chess", face the world's cutting-edge science and technology, face the main economic battlefield, face the country's major needs, and face the people's life and health, strengthen the top-level design of the open source system, build a higher level of open source infrastructure, cultivate more dynamic open source projects, and create A better policy environment. Create an open source application demonstration benchmark. Give full play to the advantages of a large-scale market and a complete industrial system, support high-quality open source communities at home and abroad to accelerate incubation and application promotion, help improve key software source innovation and supply capabilities, and allow more star open source projects and open source companies to take root and blossom.
Passenger Federation: The number of new cars reached 8 in June, of which 6 are new energy models
On June 12, according to the Passenger Federation, the number of new cars reached 8 this month. 6 models are new energy models, namely: Buick E4, Xiaopeng G6, Ian Haobo Hyper GT, Denza N7, Deep Blue S7, Song PLUS DM-i / EV Champion Edition, and only 2 models are traditional fuel models—— The brand new Yingshi style, modern Musa.
Investment and Financing
JBD received another investment from Alibaba, BYD and other institutions, and the cumulative financing reached hundreds of millions of dollars
Recently, Shanghai Xianyao Display Technology Jade Bird Display (hereinafter referred to as "JBD"), a MicroLED microdisplay manufacturer, has received hundreds of millions of RMB in A3+ and A4 rounds of strategic financing.
Among them, the A3+ round of financing was jointly led by BYD, the intellectual property fund of Shanghai Kechuang Group, and GF Qianhe; the A4 round was led by Ali, Geely and Xingyuan Capital of the Xingji Times, and the old shareholders of the two rounds of financing continued to increase their stake . At present, JBD shareholders include Alibaba, Samsung, BYD, Geely and other industrial investors. This round of financing will be used for R&D investment in MicroLED core technology, mass production of products, introduction of high-end talents and market development. JBD focuses on the research and development of Micro LED technology below 0.3 inches.
Zhuoshi Zhitong received tens of millions of yuan in Pre-B round financing
On June 12, Beijing Zhuoshi Zhitong Technology Co., Ltd. (hereinafter referred to as "Zhuoshi Zhitong"), a multi-modal artificial intelligence-based integrated perception and vehicle-road collaboration service provider, announced the completion of tens of millions of yuan in Pre-B round financing. The round of financing was led by Haoshuo Zhier Fund, followed by Yaotu Capital. The funds will be mainly used for product market expansion, new technology research and development such as AI large models, and the construction of AI training platforms.
Zhuoshi Zhitong is a computer vision original algorithm, product and solution provider. Focusing on the artificial intelligence recognition and digital scenarios of vehicles and roads, focusing on the original research and development of vehicle face recognition, pedestrian recognition, traffic scene recognition, traffic video analysis and digital twin technology, as well as the application of artificial intelligence in the vertical industry of transportation and safety.
Collov, AIGC's automated design platform, completed its A-round financing of nearly 10 million US dollars
A few days ago, Collov, an automated design service provider for AIGC, completed its Series A financing of nearly US$10 million. This round led the investment by Fang Detao Capital, the old shareholder Amoeba Capital, followed by Jiangmen Venture Capital, followed by Boston deep technology investment fund Taihill Venture, and Xinghan Capital served as the exclusive financial advisor. The financing amount will be mainly used for technology research and development and market expansion.
Collov was established at the end of 2021 and is headquartered in Silicon Valley, USA. It is a design automation cloud platform starting from interior design, and is committed to making design automation an industrial-grade solution driven by large models and generative AI.
Financial SCRM manufacturer Mogu completed a new round of financing of tens of millions of yuan
Recently, Shenzhen Mushroom Fortune Technology Co., Ltd. (hereinafter referred to as: Mushroom) has completed a new round of financing of tens of millions of yuan, jointly invested by Dewu Capital and Fukun Venture Capital, and Beishen Capital served as the exclusive financial advisor. This round of financing will be mainly used for the research and development of large-scale model-driven AI digital assistant products, as well as the expansion of cross-financial market segments, further extending from the securities industry to institutional clients such as funds, futures, and banks.
Founded in 2014, Mushroom is an intelligent social marketing solution provider focusing on the financial vertical field. Its past investors include Tencent, Country Garden Venture Capital, KIP Capital, Ivy Capital and Zhongnan Venture Capital, etc.
other
It is reported that ChatGPT is expected to add more functions in the future: upload file analysis information, and remember user portraits
According to the news on June 12, OpenAI seems to be bringing file upload and custom user portrait functions to ChatGPT. Reddit user kocham_psy said that he found a new version of the interface in the source code of ChatGPT. kocham_psy claimed that he used a script to trick the website into thinking that his account had the highest privileges, so he could access some pre-made but not yet live projects. He also said that he discovered ChatGPT's newly launched chat sharing function a week ago.
At present, OpenAI has not officially announced the above functions, so the accuracy of the above functions has yet to be tested by time.
OpenAI CEO Altman visited Japan and expressed "very optimistic" about global AI coordination and cooperation
On June 12, according to Reuters, OpenAI CEO Sam Altman (Sam Altman), who is visiting Tokyo, Japan, said that this trip around the world made him "very optimistic" about the prospect of global artificial intelligence coordination and cooperation. .
Altman had just completed his visit to EU leaders and legislators, and last week he visited Israel, Jordan, Qatar, the United Arab Emirates, India, South Korea and other countries and regions. After visiting Japan, he will also go to Singapore , Indonesia and Australia.
News that Twitter refuses to pay Google cloud bills, trust and safety team affected
On June 11, Platformer reported that Twitter's refusal to pay an upcoming Google Cloud bill could paralyze Twitter's trust and safety team.
According to reports, before Musk acquired Twitter, Twitter signed a multi-year contract with Google, involving services in combating spam and protecting accounts. But since Musk took over, Twitter has slashed costs, laid off thousands of employees, and asked the company to cut infrastructure expenditures, such as cloud service fees, by US$1 billion (currently about 7.13 billion yuan). Twitter has been trying to renegotiate its contract with Google since at least March, according to The Information.
Twitter is not entirely dependent on Google Cloud, the company also uses Amazon's cloud services. But according to The Information, Amazon also warned Twitter to withhold its advertising revenue because it owed Amazon Web Services (Amazon Web Services) cloud computing fees.
American University team finds new way to test whether papers were authored by AI with 99% accuracy
News on June 12, scientists from the University of Kansas published a paper on June 8 detailing what they say is an algorithm that can detect academic writing involving ChatGPT with more than 99% accuracy . As AI chatbot content begins to flood the world, one of the biggest concerns is reliably spotting literal discrepancies between bots and real humans.
The peer-reviewed paper, published in Cell Reports Physical Science, describes a technique for sniffing out academic research articles written by artificial intelligence. According to the researchers, their algorithm detected 100% of the entire articles written by ChatGPT. At the paragraph level, its accuracy was lower, but still impressive: the algorithm found 92 percent of AI-generated passages.
Director Guo Fan: Artificial intelligence applications may affect the creation and development of "The Wandering Earth 3"
According to news on June 11, at the 25th Shanghai International Film Festival "Science Fiction Film Week", Guo Fan said when talking about "The Wandering Earth 3", more and more applications of new technologies have entered our world. in life.
Whether it is ChatGPT or Midjourney, these artificial intelligence applications may affect the process from film creation to production. Before shooting, I hope to spend more time to investigate the technology of artificial intelligence, and then explore how to apply it. "The Wandering Earth 3" mainly explores how humans face artificial intelligence.
Guo Fan said that the editorial team searched for relevant information when creating the second episode, and predicted that products like ChatGPT would appear around 2030. Unexpectedly, it was ahead of schedule, and it became part of reality as soon as it was filmed.
