Abstract: ByteDance recently established a new AI department called Flow, which mainly focuses on the AI application layer. Thanks to the high demand for AI GPUs, Nvidia made $14.5 billion in revenue from data center hardware. Research firm Omdia stated that Nvidia sold nearly 500,000 A100 and H100 GPUs in Q3, and the delivery time for servers based on the H100 has been extended to 36 to 52 weeks.
Image Source: Generated by Wujie AI
Hot News:
ByteDance Establishes New Department "Flow" to Focus on AI Application Layer
According to 36Kr's report on November 27, ByteDance recently established a new AI department called Flow, with the technical lead being Hong Dingkun, the Vice President of Technology at ByteDance. An insider revealed that the head of this new department is Zhu Wenjia, the head of ByteDance's large model team.
It is reported that Flow mainly focuses on the AI application layer. Within ByteDance, Flow recently released recruitment posts and has been conducting social recruitment for some time. In the post, it stated that it is an AI innovation business team under ByteDance, and "currently, it has launched two products, Dou Bao and Cici, respectively, both domestically and overseas, and is incubating multiple AI-related innovative products." As of the time of publication, ByteDance has not responded.
Omdia: Nvidia Sold Nearly 500,000 A100 and H100 AI GPUs in Q3, Meta and Microsoft Are the Largest Buyers
According to IT Home's report on November 28, in the third quarter of this fiscal year, Nvidia made $14.5 billion in revenue from data center hardware, thanks to the high demand for AI GPUs. Research firm Omdia stated that Nvidia sold nearly 500,000 A100 and H100 GPUs in Q3, and the delivery time for servers based on the H100 has been extended to 36 to 52 weeks. Omdia believes that Meta and Microsoft are the largest buyers of Nvidia's H100 GPU. They have purchased as many as 150,000 H100 GPUs, far exceeding the number of H100 processors purchased by Google, Amazon, Oracle, and Tencent (50,000).
Hugging Face Co-founder Releases 2024 Predictions: Open Source LLM Will Reach the Level of the Best Closed Source LLM
AI New Intelligence reported on November 27 that Clement Delangue, co-founder and CEO of the AI open-source community Hugging Face, made 6 predictions for the industry in 2024: 1. A leading AI company will go bankrupt or be acquired at an extremely low price; 2. Open source LLM will reach the level of the best closed source LLM; 3. AI will make significant breakthroughs in video, time series, biology, and chemistry; 4. The public will pay more attention to the economic and environmental costs of AI; 5. Most content of a popular media will be generated by AI; 6. The 10 million AI developers on Hugging Face will not lead to increased unemployment.
Google Introduces AI Core Update on Pixel 8 Pro, Providing Support for the Latest AI Models
According to the webmaster's report on November 28, Google has launched a new application update called AI Core today. According to the description on the Google Play Store, AI Core will provide support for the latest AI models and a range of other AI-driven features for the Android system and applications on the Google Pixel 8 Pro phone.
China Telecom Establishes Digital Government Technology Company with 500 Million RMB, Including AI Software Development Business
According to Caixin's report on November 28, Tianyancha showed that China Telecom Digital Government Technology Co., Ltd. was recently established, with Cai Xudong as the legal representative and a registered capital of 500 million RMB. Its business scope includes software development, big data services, AI basic software development, cloud computing equipment technology services, network technology services, electronic product sales, and internet access services. Shareholder information shows that the company is wholly owned by China Telecom.
Microsoft's Copilot "Brain" to Upgrade to GPT-4 Turbo and Increase Character Limit
According to IT Home's report on November 28, Mikhail Parakhin, head of Microsoft's advertising and network services, recently tweeted on the X (Twitter) platform, stating that the team is actively improving Copilot (Bing Chat) and will upgrade to OpenAI's GPT-4 Turbo after solving certain issues. Parakhin also mentioned that after upgrading to GPT-4 Turbo, the character limit will be further expanded, with the current maximum input of 4000 characters.
Microsoft Introduces DALL-E 3 Integration for Drawing App on Win11
According to IT Home's report on November 28, Microsoft has now started rolling out the "Drawing" app with integrated DALL-E 3 to all Windows 11 users, which can be accessed through the "Cocreator" button. However, this feature was previously limited to testing by some users in the United States, France, the United Kingdom, Australia, Canada, Italy, and Germany, and it is currently unclear whether it will be open to all users.
The DALL-E model was developed by OpenAI and has been used in Microsoft Bing's image generation function.
Taotian Group Establishes Large Model Team, Led by Daishan, Ruohai, and Zhengbo
According to Leiphone's report on November 27, Taotian Group is currently establishing a large model research team and has started recruiting. It is reported that Taotian Group's large model research will mainly focus on two scenarios: advertising and contentization of window shopping. The team building work is jointly led by Daishan, CEO of Taotian Group, Ruohai, CTO of Taotian Group, and Zhengbo, CTO of Alibaba's Tmall. A technical talent related to large models stated that the large model positions at Taotian Group are still very attractive. On the one hand, Taotian Group has a very strong risk resistance in the current environment; on the other hand, the application scenarios for large models at Taotian Group are relatively clear, with higher certainty compared to general large models.
Meta's New Research Automatically Shields Irrelevant Information for Large Models, Increasing Accuracy by 27%
According to QbitAI's report on November 27, Meta recently announced a new study on attention mechanisms. The research team named this new attention mechanism "System 2 Attention" (S2A), which can shield irrelevant interference and further improve the accuracy of large models. According to test results, this mechanism does not require fine-tuning or training and can increase the accuracy of large models by 27%.
Investment and Financing:
Biotechnology and AI Startup Cradle Raises $24 Million in Series A Financing
According to TechCrunch's report on November 28, after completing a $5.5 million seed round of financing last year, biotechnology and artificial intelligence startup Cradle has raised $24 million in Series A financing, with lead investor Index Ventures, Kindred Capital (also a seed round investor), and individual investors Chris Gibson and Tom Glocer participating. The new funds will be used for team development and sales.
According to reports, Cradle's technology can significantly shorten the manufacturing time of functional proteins and reduce the number of experiments. Currently, the company has used its software to produce an alternative version of T7 RNA polymerase, an RNA production enzyme.
AIGC Recommended Reading:
"Hugging Face CEO Predicts Six Major Changes in the AI Industry in 2024!"
OpenAI Chairman Greg Brockman predicted on the last day of last year that 2023 would make 2022 look like AI development had not yet awakened. Sure enough, in 2023, the AI industry experienced a comprehensive outbreak. Yesterday, Clement Delangue, CEO of Hugging Face, the world's largest AI open-source community, made six specific predictions for the industry in 2024:
https://www.aixinzhijie.com/article/6839692
"The AI Story of Meitu, Still 'Old Wine in New Bottles'"
If we count from the official release of chatGPT, in less than a year, this global AI large model storm is overturning everything, including Chinese enterprises at the moment. In less than a year, Chinese enterprises have rushed to release hundreds of industry large models, and in various evaluation rankings and articles, "the strongest capability" seems to have become the official unified language. Although many analytical articles have already reached some consensus, such as the current domestic large model craze has tended to calm down, and the competition pattern of basic large models has been determined, and the struggle at the application layer has begun. But if we leave the context of geeks' excitement and return to a company or an industry itself, we may have a more practical perspective to solve some more practical and direct problems: What is the AI strategy of a company? How will it affect the future of this company? Is it a follower of the trend or a trendsetter?
https://www.aixinzhijie.com/article/6839674
"Nvidia's GPU Becomes a God in the Battle! Huang Renxun Bets on Artificial Intelligence and Builds a Trillion-Dollar Graphics Card Empire"
The popularity of AI has made Nvidia's market value soar, successfully entering the trillion-dollar club. As the soul of the company, what stories has Huang Renxun experienced along the way? How did this leather-clad man build his graphics card empire step by step?
https://www.aixinzhijie.com/article/6839680
免责声明：本文章仅代表作者个人观点，不代表本平台的立场和观点。本文章仅供信息分享，不构成对任何人的任何投资建议。用户与作者之间的任何争议，与本平台无关。如网页中刊载的文章或图片涉及侵权，请提供相关的权利证明和身份证明发送邮件到support@aicoin.com，本平台相关工作人员将会进行核查。
