One week of AIGC丨China's first AIGC regulatory document was officially released, and American AIGC unicorns began to lay off staff

0e91ee6a51ae242d3e8c6277d4aabe04.gif

There is no such thing as a circle without rules. After the large model task force of the National Artificial Intelligence Standardization General Group officially launched the formulation of national standards for large model testing, China's first AIGC regulatory document was officially issued. Overseas, there are hot and cold. AIGC unicorns in the United States with a valuation of 1.5 billion US dollars have started layoffs. Nvidia has won the large-scale model arms race. The price of A800 chips is one day.

066b83d50605685327a09c3ecd3bdbcf.jpeg

Image source: Generated by Unbounded AI

01

financing

Indian IT service provider Wipro plans to invest $1 billion in AI over next three years

Indian IT service provider Wipro said it plans to invest $1 billion in AI over the next three years, Reuters reported on July 12. The investments will focus on the expansion of artificial intelligence, big data and analytics solutions, as well as new research and development and the development of new platforms, the company said in an exchange filing.

Nvidia invests $50 million in biotech Recursion to train AI models for drug discovery

According to Reuters news on July 12, Nvidia announced a $50 million investment in biotechnology company Recursion to accelerate the training of the company's artificial intelligence models for drug discovery. Shares of Recursion jumped about 167% in premarket trading on the news.

It is reported that Recursion will use its biological and chemical data sets of more than 23,000 terabytes to train artificial intelligence models on Nvidia's cloud platform. Nvidia can then license the AI ​​models to biotech companies through BioNeMo. BioNeMo is a generative artificial intelligence cloud service for drug discovery that Nvidia launched earlier this year.

AI voice startup Resemble AI completes $8 million in financing

According to a TechCrunch report on July 12, AI voice startup Resemble AI announced that it has received $8 million in Series A financing. This round of financing was led by Javelin Venture Partners, with Craft Ventures and Ubiquity Ventures following. Co-founder and CEO Zohaib Ahmed said the funding will be used to further develop Resemble's enterprise products and double the size of the team to more than 40 people by the end of the year.

South Korean tech giant Kakao launches AI image generator Karlo 2.0 and establishes 10 billion won AI fund

According to TechCrunch, on July 12, Kakao Brain, the artificial intelligence division of South Korean technology giant Kakao, announced the launch of the AI ​​image generator Karlo 2.0. Users only need to input English or Korean prompts to generate pictures in seconds. In addition, Kakao Brain established the Karlo 100X fund, which has a size of 10 billion won (approximately US$7.7 million) to support AI start-ups developing image generation technology.

Artificial intelligence-based large-scale model company reInvent AI completed tens of millions of dollars in the first round of financing, Lenovo Ventures exclusive investment

According to 36 Krypton’s report on July 13, recently, reInvent AI announced that it has completed tens of millions of dollars in the first round of financing, which was exclusively invested by Lenovo Venture Capital. reInvent AI focuses on basic models and general artificial intelligence (AGI), aiming to create industry-oriented professional large models. The founding team members include academicians of the European Academy of Sciences, core members of OpenAI GPT, and the former SAP global senior vice president and global AI leader, as well as the original author of the Transformer paper as a consultant. reInven tAI will operate simultaneously in Berlin, Germany and Shanghai, China, and its German company is called nyonic.

Sapphire Ventures plans to invest more than $1 billion in AI startups

Jai Das, president of venture capital firm Sapphire Ventures, announced at a conference on Tuesday that the company plans to invest more than $1 billion in artificial intelligence startups to accelerate the adoption of the technology in global enterprises, according to a report by the "Technology Board Daily" on July 13. "As AI is affecting how new software is built, it's really changing the game," Das said. AI startups raised more than $40 billion in the past six months, accounting for nearly a quarter of all startup funding, according to research firm PitchBook.

People familiar with the matter: Hugging Face is reviewing the terms of the D round of financing competition, or raising at least $200 million at a valuation of $4 billion

AI startup Hugging Face is reviewing competing term sheets for a Series D funding round that could raise at least $200 million at a valuation of $4 billion, with Ashton Kutcher's venture capital firm Sound Ventures currently leading the round, Forbes reported on July 14, people familiar with the matter said.

AI startup Nomic raises $17 million led by Coatue

According to SiliconAngle reports, on July 14, artificial intelligence startup Nomic AI announced the completion of a $17 million Series A round of financing, led by Coatue, with participation from Contrary Capital, Betaworks Ventures, SV Angel, Story Ventures and Factorial Capital, as well as many angel investors, valuing Nomic at about $100 million.

Artificial intelligence industrial application service provider "Yunshuo IoT" completed tens of millions of yuan in Series A financing

According to a report from the "Kechuang Board Daily" on July 17, recently, Yunshuo IoT completed the A round of financing. The leading investor was Qifu Capital, and the old shareholders followed suit. Yunshuo IoT is an artificial intelligence industrial application service provider. Starting from the AI ​​predictive monitoring and decision-making of production quality, it has implemented "industrial artificial intelligence" in this important field, helping industrial enterprises to improve quality, increase efficiency and reduce costs, and strive to change the current situation of mismatch between enterprise pain points and technical capabilities.

Data: Since the beginning of this year, AIGC has raised more than 100 billion yuan, of which 18 single financings of more than 100 million yuan

Babbitt News, "Kechuangban Daily" issued an article on July 16, saying that in the era of AI 2.0, the amount of global financing has experienced a huge increase. Within two years of the release of GPT-3, the investment of venture capital institutions in AI companies has increased by 4 times, and there will be 1.37 billion US dollars in financing in 2022 alone. Since the beginning of this year, AIGC has presented a capital feast, raising more than 100 billion yuan in half a year, of which 18 single financings exceeded 100 million yuan.

Tom Cat: It is planned to raise no more than 2.3 billion yuan for projects such as AI interactive terminal products

According to a report by the Financial Associated Press on July 14, Tom Cat announced that it plans to raise no more than 2.3 billion yuan in additional funds, which will be used for Tom Cat family IP series AI interactive terminal product projects, vertical field model development projects, high-density big data computing power infrastructure projects, and supplementary working capital.

Unisound went public in Hong Kong with the concept of AI, with a net loss of 375 million yuan in 2022

According to the Securities Daily, Yunzhisheng went public in Hong Kong, and the large model of mountains and seas has attracted market attention. This large language model will be launched in 2023, and it is reported that it has 60 billion parameters. Through a series of tool industry knowledge enhancement, command optimization and human feedback enhancement, Yunzhi claims that it can further upgrade the mountain and sea model into a large language model with comprehensive general knowledge and industry professional capabilities, adapt to various application scenarios through MaaS, and adopt it in smart medical and smart life products. According to more media materials, Unisound is the fourth largest AI solution provider in China and the fastest growing among enterprises with annual revenue exceeding RMB 500 million. The company leads AGI (General Artificial Intelligence) technology in China and has succeeded in the commercialization of large language models, and is regarded as one of the pioneers on the critical path to AGI. However, the financial data is a bit surprising. In 2022, Yunzhisheng will have a revenue of 601 million yuan and a net loss of 375 million yuan.

Force Energy, an AI operator of urban virtual power plants, completed tens of millions of yuan in financing, led by Saitian Development

According to 36 Krypton’s report on July 15, recently, Force Energy, an AI operator for urban virtual power plants, has successfully completed tens of millions of A-round financing, led by Saitian Development, Chuangwei Venture Capital Fund and Suzhou Chinese Consortium Venture Capital Fund as strategic investment institutions.

02

start a business

The "Eight Sons of Transformer" all leave Google, and the last resigners will devote themselves to AI entrepreneurship

According to the "Kechuangban Daily" report on July 12, eight Google researchers previously jointly published a paper explaining the Transformer architecture, and they were known as the "Eight Sons of Transformer". Now one of the authors, Llion Jones, will leave Google at the end of this month. With his resignation, all the authors of the paper have left Google. Jones said that leaving Google was not an easy decision, but given the momentum of AI development, it was time to try something different. According to reports, he himself plans to take a break before starting a company. It is reported that the Transformer architecture is the most important part of Google's famous algorithm BERT, and it is the "T" at the end of ChatGPT.

U.S. unicorn Jasper AI announces layoffs

Babbitt News, on July 12, Dave Rogenmoser, co-founder and CEO of the American AI unicorn company Jasper AI, announced on LinkedIn that he would reshape the team, centralize and integrate resources. Rogenmoser said that the laid-off employees are the first batch of people in the industry to develop, promote and support artificial intelligence products. If a team is recruiting, it will be recommended.

03

large model

Users complain that GPT-4 has become "lazy" and "stupid", industry insiders say it may be related to OpenAI's reset architecture

According to the "Kechuangban Daily" report on July 14, recently, some doubts have begun to appear on the Internet. Many users complain that OpenAI's GPT-4 has become less powerful. Some people say that this model has become "lazier" and "dumber" compared with its previous reasoning ability and other outputs. Some people in the industry believe that this may be related to OpenAI's major redesign of the system. They speculate that OpenAI uses a hybrid expert (MoE) model to make the output of the generative model higher quality, lower cost, and faster response.

GPT-4 model architecture leak: Contains 1.8 trillion parameters, uses a mixed expert model

According to a report by IT House on July 13, foreign media Semianalysis recently revealed the GPT-4 large model released by OpenAI in March this year, including the GPT-4 model architecture, training and inference infrastructure, parameter volume, training data set, token number, cost, Mixture of Experts and other specific parameters and information.

JD Yanxi large model is expected to be officially launched in August

According to the Beijing Business Daily, on July 13, JD.com officially released the JD Yanxi large model at the 2023 JD Global Technology Explorer Conference and JD Cloud Summit, and at the same time released the Yanxi AI development computing platform. Cao Peng, Chairman of JD Group Technical Committee and President of JD Cloud Business Unit, mentioned that the platform has started appointment registration and is expected to be officially launched in August.

Domestic large-scale model ChatGLM2-6 B allows free commercial use

According to the announcement of Zhipu AI & Tsinghua KEG on July 14, in order to better support the prosperity and development of domestic large-scale model open source ecology, Zhipu AI and Tsinghua KEG laboratory decided that the weights of ChatGLM-6 B and ChatGLM2-6 B are completely open to academic research from now on, and free commercial use is allowed after completing enterprise registration and obtaining authorization.

04

application

Focus Media: The vertical marketing model has been deployed and is being fine-tuned and iterated

According to Jiemian News on July 11, Focus Media stated on the interactive platform that the company's large-scale marketing model has been deployed and is being fine-tuned and iterated. Through in-depth mining of historical advertising data, unique high-quality data sets and reasonable human-labeled feedback data are carefully constructed, and continuous iteration and optimization can be carried out in actual scenarios to continuously improve its capabilities. Such a new AI-driven industry solution will greatly enrich its comprehensive capabilities in advertising creative generation, advertising strategy formulation, advertising delivery optimization, advertising data analysis, etc., and then empower more advertisers.

Meitu Xiuxiu's "AI image expansion" function is online, which can predict and generate larger frames based on image context

Babbitt News, Meitu Xiuxiu’s official Weibo posted on July 11, saying that the “AI image expansion” function is now online, and both Meitu Xiuxiu App and Wink App can immediately experience the AI ​​​​enlargement function.

It is said that this function is based on AI algorithm intelligent recognition, which can predict and generate missing parts according to the context and texture of the image, and can bring a larger frame and wider viewing angle to the original picture.

Stability AI launches Stable Doodle, a sketch-to-image tool

According to news from IT House on July 16, Stability AI, the startup behind the image generation model Stable Diffusion, recently launched a new service that can convert sketches into images. The service, called Stable Doodle, uses the latest Stable Diffusion model to analyze the outlines of sketches and generate "visually pleasing" artwork. Users can already use the service through the ClipDrop platform, which was acquired by Stability AI in March this year, and its predecessor was Init ML, an AI startup founded by former Google employees.

Indian e-commerce company Dukaan announced layoffs of 90% of customer service staff to embrace artificial intelligence

According to IT House, Indian e-commerce startup Dukaan recently announced that it will lay off 90% of its customer service staff and embrace artificial intelligence. Suumit Shah, CEO of Dukaan, said that the generative AI chatbot has greatly improved customer service work, reducing the first response time from 1 minute and 44 seconds to instant; the average resolution time was shortened from 2 hours and 13 minutes to 3 minutes and 12 seconds; customer service costs were reduced by 85%.

Beijing released the first batch of industrial AI large-scale model application cases

According to the news from Beijing Daily on July 15, a few days ago, at the first phase of the Beijing Artificial Intelligence Ceremony Leading and Promotion Event held by the Beijing Municipal Science and Technology Commission and the Zhongguancun Management Committee, Beijing released the first batch of 10 typical application cases of large-scale industry models jointly developed by the technology supplier and the scene demander. According to Liu Hang, the second-level inspector of the Beijing Municipal Science and Technology Commission and the Zhongguancun Management Committee, the first batch of 10 typical application cases of large-scale industry models announced this time mainly focus on key areas such as urban governance, medical health, scientific research, smart finance, smart life, and smart cities. They are jointly developed by 10 technology suppliers and scene demanders including Baidu, Zhipu Huazhang, and the Institute of Automation of the Chinese Academy of Sciences.

Unbounded AI has opened the copyright market for AI works, and can perform blockchain registration for works

Babbitt News, Unbounded AI has recently opened the copyright market for AI works, and will perform blockchain registration for works entering the AI ​​copyright market. In response to the previous "Kun Family Army" incident, Unbounded AI provided important evidence support for the blockchain copyright certificate issued by the creator hero for his rights protection. On June 27, Daxia uploaded the work to the Wujie AI digital copyright market and made a corresponding confirmation registration. Afterwards, the communication with the transporter was fruitless, so he used this as a basis to complain about his behavior of transporting the works. Lawyer Liu Honglin said that taking Unbounded AI as an example, the platform clearly stipulates in the user agreement that for works created using Unbounded AI, the copyright of the author belongs to the creator, that is, the user.

05

Supervision

Seven departments including the Cyberspace Administration of China jointly announced the "Interim Measures for the Management of Generative Artificial Intelligence Services"

Babbitt News, recently, the National Cyberspace Affairs Commission, the National Development and Reform Commission, the Ministry of Education, the Ministry of Science and Technology, the Ministry of Industry and Information Technology, the Ministry of Public Security, and the State Administration of Radio, Film and Television announced the "Interim Measures for the Management of Generative Artificial Intelligence Services" (hereinafter referred to as the "Measures"), which will come into force on August 15, 2023. The relevant person in charge of the State Internet Information Office stated that the introduction of the "Measures" aims to promote the healthy development and standardized application of generative artificial intelligence, safeguard national security and social public interests, and protect the legitimate rights and interests of citizens, legal persons and other organizations.

FTC launches sweeping investigation into OpenAI over ChatGPT leaks, fabricated answers

According to IT House citing the Washington Post, the US Federal Trade Commission (FTC) recently launched a comprehensive investigation into the San Francisco-based company OpenAI because the popular chatbot ChatGPT developed by the company may have violated consumer protection laws and put personal reputation and data at risk.

The FTC sent the company a 20-page demand for record about how it is addressing risks posed by its AI models, an investigation that poses a significant regulatory threat to OpenAI’s U.S. operations. The FTC emphasized that although the administration and Congress have not clearly stipulated comprehensive regulatory rules, existing consumer protection laws still apply to AI. Senate Majority Leader Charles E. Schumer predicted that new AI legislation would still be months away.

06

big company

Microsoft launches CoDi, an artificial intelligence model that can interact and generate multimodal content

According to a report from IT House on July 11, Microsoft recently launched the Composable Diffusion Model (CoDi), a unique AI model based on Composable Diffusion, which is designed to interact and generate multimodal content.

Microsoft designed CoDi with the goal of addressing the limitations of traditional unimodal AI models. Taking synchronized video and audio as an example, there may be inconsistencies and alignment issues when splicing independently generated information streams together. CoDi employs a unique composable generation strategy to align multimodalities during diffusion to generate interwoven modalities, and CoDi is able to process arbitrary input modalities and generate content of arbitrary modalities.

It is reported that CoDi was developed by the Microsoft Azure Cognitive Services Research Team in cooperation with the University of North Carolina at Chapel Hill, and is also part of Microsoft's project i-Code, which uses artificial intelligence to enhance human-computer interaction.

The price of Nvidia A800 chip is one day, if the quantity is small, there is no 110,000 yuan/chip and you can’t get it

According to The Paper, the price of the A800 chip sold by AI chip giant Nvidia in the Chinese market has risen sharply recently. Recently, an agent of Nvidia in China told The Paper that the current price of Nvidia A800 chips is one day a day. If you buy a lot, you can buy more than 100,000 yuan/chip.

According to a Wall Street Journal report at the end of June, the U.S. Department of Commerce may take action as soon as early July to prohibit chipmakers such as Nvidia from exporting chips to customers in China and other relevant countries without obtaining a license in advance.

07

View

106 large-scale models have appeared in the country, and the "capital of large-scale models" is still competing

According to incomplete statistics from the Economic Observer, in the new round of generative AI upsurge, 106 large-scale models have emerged across the country. Among them, the number of large-scale model companies and the number of R&D companies in Beijing occupy the "C position", as many as 54; Shanghai has accumulated advantages in AI computing power; cities such as Shenzhen, Guangzhou, Hangzhou, and Chengdu also have many large-scale model companies. At present, the large models in various places are still in internal testing and are not open to the public. The development of large-scale models is still at the beginning of chaos, and it remains to be observed who will win in the end and become the real "capital of large-scale models".

Bill Gates: AI is the most transformative innovation of our time and will have the same impact as the advent of the PC

According to the "Kechuangban Daily" report, on July 11 local time, Microsoft founder Bill Gates published an article pointing out that the future of artificial intelligence will not be as grim as some people imagine, nor will it be as optimistic as some people imagined. The risks do exist, but these risks can be controlled. He believes that the impact of AI will not be as great as the industrial revolution, but it will definitely have a major impact like the advent of personal computers. Many problems caused by AI can be solved by AI. Gates encourages everyone to pay as much attention as possible to the development of AI, because it is the most transformative innovation of our time.

Musk: AGI will definitely appear, xAI can surpass Google Microsoft OpenAI

On July 14, local time, Musk held a “Twitter Space” meeting with the newly formed xAI team. Musk said that AGI (General Artificial Intelligence) will definitely appear, and xAI will participate in it, which can provide more competitive products than Google and Microsoft OpenAI.

08

research report

Report: The global AI computing market will grow from $19.5 billion in 2022 to $34.66 billion in 2026

According to Jinshi's report on July 12, Inspur Information, Tsinghua University Global Industry Research Institute and IDC released the "2022-2023 Global Computing Power Index Evaluation Report" on July 12. The "Report" shows that AI computing represented by generative AI will show a skyrocketing trend in the future. According to IDC's forecast, the global AI computing market will grow from US$19.50 billion in 2022 to US$34.66 billion in 2026, of which the generative AI computing market will grow from US$820 million in 2022 to US$10.99 billion in 2026. Generative AI computing will grow from 4.2% to 31.7% of the overall AI computing market. In the field of AIGC, artificial intelligence algorithms and technologies are applied in various fields and industries, such as self-driving cars, medical diagnosis, financial forecasting, etc.

AI tools save workers nearly 400 hours a year, survey shows

According to news from IT House on July 10, workforce analysis and planning company Visier recently conducted a survey of 3,030 employees in more than 250 companies in the United Kingdom, the United States, Canada, and Germany. The results show that British employees who have adopted AI tools save an average of about 93 minutes a day. Based on this calculation, AI tools can help enterprise employees save 390 hours per year. Even so, more than half (56%) of respondents to the survey claimed that their employers have not embraced AI tools or encouraged employees to use them to increase productivity.

Additionally, more than half (53%) of workers are concerned that AI may replace the skills they currently possess, contributing to widespread fears of layoffs as governments race to put in place protective regulations in place. Roughly two-thirds (67%) say having AI skills is important for their future career.

Babbitt Park is open for cooperation!

8c6a97bed14cb8202ad5ae12802f7726.png

cc8983e73e8557472374b748c7109acb.jpeg

4f70bb0afe26bfa1f1e50f730f2dfad4.gif

Chinese Twitter: https://twitter.com/8BTC_OFFICIAL

English Twitter: https://twitter.com/btcinchina

Discord community: https://discord.gg/defidao

Telegram channel: https://t.me/Mute_8btc

Telegram community: https://t.me/news_8btc

99cfd47a9e1f14f13c42bd10e6251dbc.jpeg

Guess you like

Origin blog.csdn.net/weixin_44383880/article/details/131777768