Babbitt | Metaverse's daily must-reads: Large models of Huawei and Xiaomi have surfaced, Honor is ready to try, and AI large models are about to set off a "new war" in the mobile world?

Abstract: According to Titanium Media reports, on August 4, Huawei released the HarmonyOS 4 system at the developer conference. Through the blessing of the Pangu large model, the smart assistant Xiaoyi will have the ability of AI large models, bringing a more intelligent experience. Just a week after Huawei released the HarmonyOS 4 system, Xiaomi's large-scale model also surfaced, and the large-scale pre-trained language model MiLM-6B developed by it appeared on the large-scale model evaluation list. In the down cycle of the entire consumer electronics, the integration of large models is not only an AI game among mobile phone manufacturers, but also an active change and "self-help" of the entire industry chain.

Image credit: Generated by Unbounded AI tools

Hot info:

U.S. investment firm Ripplewood plans to invest $1 billion in OpenAI

According to a report by Webmaster's Home on August 14, Ripplewood, a well-known American investment company, plans to invest US$1 billion in artificial intelligence software company OpenAI. The investment will leverage products developed by OpenAI, including chatbot software ChatGPT and image artificial intelligence software Dall-E. This investment could make Ripplewood a significant financial investor in the artificial intelligence industry, especially if OpenAI goes public in the future, and Ripplewood's stake in ChatGPT could be very valuable.

In addition, Ripplewood plans to launch artificial intelligence-based investment tools that provide accurate market forecasts and optimize investment portfolios. Such a solution is expected to be available by the end of 2023, subject to successful participation in OpenAI. It is reported that Ripplewood Holdings LLC was established in 1995 by American investment banker Timothy C. Collins, and the company makes decisions on Ripplewood's investment. Collins currently holds a 15% stake, and he led the investment team to expand investments in Europe and Southeast Asia, with partners including Blackrock, Nathaniel Rothschild and Bank of America.

Microsoft Research: GPT-4 is sufficient for medical tasks

According to a report from the Webmaster's House on August 14, Microsoft Research believes that GPT-4 is sufficient for medical tasks, can speed up medical processes and improve efficiency. Large language models (LLMs) such as GPT-4 have huge potential in the medical field, according to the Microsoft Research team. These models can help speed up medical processes, such as improving the efficiency of cancer drug development by processing large-scale unstructured patient data.

Microsoft has also introduced language models such as BioGPT specifically for medical tasks, but has now made it clear that it will mainly rely on GPT-4 in the future. In addition, Microsoft is developing the "LLaVA-Med" model to more closely integrate medical data and research to accelerate medical care and research.

British deputy prime minister: AI may have a bigger impact on the UK than the industrial revolution

According to a report from the Webmaster's House on August 14, British Deputy Prime Minister Oliver Dowden stated that artificial intelligence may have a greater impact on the UK than the Industrial Revolution. He believes that artificial intelligence will increase productivity and speed up government decision-making. Dowden said that artificial intelligence is developing rapidly, and it can draw information from large data sets to help people make decisions.

Dowden emphasized that the development of artificial intelligence will lead to a major restructuring of the economy, but the government will ensure that artificial intelligence does not cause harm to humans. He likened the spread of artificial intelligence to the invention of the car, and believed that artificial intelligence can complete boring tasks and allow humans to focus on more interesting things. He also emphasized the need to recognize the risks in the development of artificial intelligence and take corresponding measures to solve them.

Ming-Chi Kuo: The release of NVIDIA L40S is good for AI edge computing

According to the "Kechuang Board Daily" report on August 14, the well-known analyst Ming-Chi Kuo posted on the X platform (formerly Twitter) that Nvidia's release of L40S may shift investors' focus on AI investment to AI edge computing. Beneficiaries include Dell, ASUS, etc.

It is understood that Nvidia released the L40S GPU in early August. This computing acceleration graphics card can be used for graphics rendering, artificial intelligence large-scale model training and reasoning, 3D design, and video processing. The AI computing performance is higher than that of the A100 GPU. Since this accelerator card does not support NVLink, and does not choose to use HBM video memory, but uses GDDR6 ECC video memory, the L40S is more suitable for edge computing than A100 and H100.

Deloitte Japan will use AI to review corporate financial information

According to a report by the "Technology Board Daily" on August 14, Deloitte Tohmatsu Japan (Deloitte Tohmatsu) will officially start using artificial intelligence (AI) to review corporate financial information as early as September. By letting AI learn the financial reports of companies with past mistakes and misconduct, it can find out the risk of non-compliance of audited companies.

Investment and Financing:

Google-backed U.S. AI company Anthropic receives $100 million investment from South Korea's SK Telecom

According to a Reuters report on August 14, SK Telecom, South Korea's largest telecom operator, announced that it will invest US$100 million in Anthropic, an American AI company, to strengthen its telecom-driven AI business. SK Telecom said the two companies plan to jointly develop a large language model that can support multiple languages for global telecom companies and build an artificial intelligence platform.

Anthropic was founded in 2021 by former OpenAI researchers and is headquartered in San Francisco, USA. The company developed the chatbot Claude and has received investment from technology giants such as Google and Salesforce. It completed a $450 million Series C financing in May. As previously reported, Anthropic released the AI model Claude Instant 1.2.

Metaverse recommended reading:

"Domestic large-scale model has no company to vote for"

The large-scale venture capital market has come to a new stage of "no company to invest in". Entrepreneurs want money, but new rounds of large-scale financing are becoming more and more difficult; investors find it difficult to find suitable targets, and those who are promising can't invest, and those who can afford it can't. This phenomenon is both unexpected and reasonable. Ten thousand arrows have been launched for less than half a year, and the frequency of ordinary people's use of AI and large models has declined. When the enthusiasm has calmed down, even the growth rate of the leader ChatGPT's visits has begun to slow down. But the volume and congestion at this track is far from over.

"Human data is in a hurry, Microsoft OpenAI began to feed AI with AI, Altman said: all data in the future will become synthetic data"

According to the "Financial Times", many companies are feeding the results generated by large models, the so-called synthetic data (Synthetic data), to large models with smaller parameters, and found that the results are not bad. For the use of synthetic data, OpenAI CEO Sam Altman not only does not mind, but also said that "all data in the future will become synthetic data." Cohere, a big-model startup valued at $2 billion, is also using synthetic data. Aidan Gomez, CEO of the company and one of the authors of the classic large model Transformer paper, even believes that synthetic data may accelerate the path to a "superintelligent" AI system.

"Annual loss of 500 million knives, OpenAI will go bankrupt in 2024? Altaman reveals that there is a shortage of GPUs, and top talents are leaving the job

Nine months after its birth, ChatGPT has spent nearly 200 million US dollars! According to foreign media Analytics, OpenAI is likely to go bankrupt by 2024. OpenAI burns about $700,000 a day just to keep ChatGPT going. This fee does not include other AI products such as GPT-4 and DALL-E 2.

"Huawei and Xiaomi rushed to land on the beach, and the AI large model set off a "new war" in the mobile phone arena"

In the first half of this year, large-scale models were released one after another, and a "hundred-model battle" was staged among enterprises. In the second half of the year, the large model has somewhat "ebbed", replaced by the landing of the application side, allowing more users to have a deep understanding of the large model, and at the same time changing the development direction of terminal hardware. Among them, the mobile phone is one of the terminal devices with the strongest perception ability, and the blessing of the large model on the terminal side and the cloud side is promoting the innovation of its acceleration experience.