Summary: Baidu will officially launch an AI native application development product - “Qianfan AI Native Application Development Workbench” to help developers greatly lower the threshold for AI native application development and accelerate the implementation of enterprise AI native applications. Alibaba Cloud open-sourced the 72 billion parameter model Qwen-72B, the 1.8 billion parameter model Qwen-1.8B, and the large audio model Qwen-Audio. Microsoft President Brad Smith said in an interview that super AI is unlikely to be born in the next 12 months, warning that the technology could be decades away.
Image source: Generated by Unbounded AI
Hot Info:
Microsoft will support the United Nations in building a new AI platform and a global climate data center
The United Nations and Microsoft announced a partnership to support the United Nations Framework Convention on Climate Change (UNFCCC) in building a new artificial intelligence platform and global climate data center to monitor and analyze progress in global emissions reductions. This collaboration will significantly simplify the validation and analysis of climate data submitted by the 196 Parties to the Paris Agreement.
Baidu will release “Qianfan AI Native Application Development Workbench”
Baidu will hold the 2023 Baidu Cloud Intelligence Conference and Intelligent Computing Conference on December 20, when it will officially launch an AI native application development product - “Qianfan AI Native Application Development Workbench” to help developers greatly lower the threshold for AI native application development and accelerate the implementation of enterprise AI native applications.
Amazon has launched a number of AI tools, including the Titan series of AI models
On November 30, it was reported that following the launch of a new chatbot called Amazon Q, the upgraded AI system processor Trainium2, and the expansion of the partnership with NVIDIA, Swami Sivasubramanian, vice president of data and artificial intelligence at Amazon AWS, announced a series of new AI tools at the re:Invent conference yesterday. These include three generative AI models from the “Titan” family: Titan Image Generator, Titan Text Express, and Titan Text Lite. In addition, Amazon Bedrock has been upgraded to provide enterprise customers with access to most models on the market, including Jurassic for AI21, Claude 2.1 for Anthropic, Llama 2 for Meta, and Stable Diffusion for Meta.
Google’s DeepMind uses AI tool GNoME to discover 2.2 million new crystalline materials
On November 30, Google’s DeepMind presented the AI tool GNoME in the journal Nature and introduced the relevant applications of AI in materials science. It is reported that DeepMind has discovered 2.2 million new crystals using GNoME, of which 380,000 crystals are stable materials that can be manufactured in the laboratory and are expected to be used in batteries or superconductors. DeepMind claims that it would take 800 years to calculate these materials by humans alone.
Mastercard launches Shopping Muse, an AI service that provides personalized recommendations to consumers
On December 1, it was reported that Mastercard launched the Shopping Muse service, which uses generative AI to help consumers search and discover products in retailers’ digital catalogs.
It is reported that Shopping Muse is powered by Dynamic Yield, a personalization platform and decision-making engine acquired by Mastercard from McDonald’s in April 2022. To date, Dynamic Yield has helped more than 400 brands deliver personalized product recommendations, offers, and content based on a range of factors, such as purchase history, page views, customer affinity profile information, and AI algorithms.
The world’s first international standard for financial risk control models was released, which was led by Tencent
On November 30, the IEEE Financial Risk Control Large Model Standard Kick-off Meeting was held in Shenzhen. Initiated by Tencent, the standard is the world’s first international standard for large models in the field of financial risk control, aiming to provide reference and guidance for financial institutions to apply AI large model technology in risk control modeling, so that financial institutions can efficiently predict, measure and manage business risks in an increasingly complex and data-driven financial environment.
openKylin has launched an AI framework installation assistant, which can deploy relevant environment dependencies with one click
The AI4 OS SIG group of the openKylin community has developed an AI framework installation assistant, which claims to help relevant developers install AI frameworks with one click. According to reports, the AI framework installation assistant adopts a front-end and back-end separation architecture, which is claimed to improve development efficiency, maintainability, and scalability. The front-end is responsible for detecting the system hardware, automatically recommending the installation method, providing installation options, obtaining installation options, displaying installation information and progress, and the back-end performs specific tasks and provides timely feedback to the front-end.
Former Apple employees create startups with plans to reimagine the way desktop computers work with generative AI
On December 1, it was reported that former Apple employees Ari Weinstein, Conrad Kramer and Kim Beverett have teamed up to create a startup called Software Applications Incorporated, which plans to make AI more humane and wants to reimagine how desktop computers work with generative AI.
It is reported that Ari Weinstein and Conrad Kramer created the iOS automation app Workflow, which was acquired by Apple in 2017, which is the prototype of Apple’s “shortcuts”.
Kunlun Wanwei officially released the “Tiangong SkyAgents” platform
On December 1, Kunlun Wanwei officially released the “Tiangong SkyAgents” platform. According to reports, “Tiangong SkyAgents” is based on Kunlun Wanwei’s “Tiangong Model”, which has the ability to learn independently and think independently from perception to decision-making, from decision-making to execution. Users can build their own single or multiple “personal assistants” through natural language. In addition, different tasks are modularized and executed through the operating system module, including question preset, specified response, knowledge base creation and retrieval, intent recognition, text extraction, HTTP request and other tasks.
Alibaba Cloud open-source Tongyi Qianwen 72 billion parameter model
On December 1, it was reported that Alibaba Cloud open-sourced the 72 billion parameter model Qwen-72B, the 1.8 billion parameter model Qwen-1.8B and the audio large model Qwen-Audio. It is reported that in addition to the pre-trained model, Alibaba Cloud’s open-source model has also launched a corresponding dialogue model, providing a 4bit/8bit quantitative version of the model for 72B and 1.8B dialogue models, which is convenient for developers to inference and train.
Brad Smith, President of Microsoft: Super AI is unlikely to be born in the next 12 months
Microsoft President Brad Smith said in an interview that super AI is unlikely to be born in the next 12 months, warning that the technology could be decades away. “In the next 12 months, you’re absolutely not going to see the so-called AGI, where computers are more powerful than people,” he said. It will take years, if not decades, but I still think it’s time to focus on security. ”
Asked if these kinds of risks were the reason for Altman’s dismissal, Smith said, "I don’t think that’s the case at all. I think there’s clearly a split between the board and others, and fundamentally, it’s not because of concerns like that. ”
Investment & Financing:
“Haina AI” Completes Tens of Millions of Yuan in Series A Financing, Exclusive Investment by Lenovo Venture Capital
It was reported on December 1 that “Haina AI” recently completed tens of millions of yuan in Series A financing, which was exclusively invested by Lenovo Venture Capital, and this round of financing funds will be used for talent recruitment, AI model research and development, and marketing system construction. It is understood that “Haina AI” is an AI product in the vertical field of talent recruitment under Beijing Qunxing Shining Technology Co., Ltd., which was first launched in 2019, specializing in AI interview services and helping enterprises complete recruitment interviews with the help of AI technology.
Microsoft to invest £2.5 billion in UK AI industry over the next 3 years
On November 30, Microsoft announced that it will invest £2.5 billion over the next three years to expand the UK’s next-generation AI data centre infrastructure. It is reported that this is Microsoft’s largest investment in the UK’s 40-year history. Microsoft has pledged to more than double its data center footprint in the UK, bringing more than 20,000 state-of-the-art graphics processing units (GPUs) to the UK by 2026 and training more than a million people for the AI economy. Microsoft will also support the UK’s growing AI security and research efforts through partnerships with the government and top universities.
Good Article:
“Human-Machine Relationship in the Era of Large Models: Don’t Gently Enter the Era of Symbiosis with AI”
There is no doubt that we are at an inflection point in the technological era: when artificial intelligence becomes the technical background of daily life, when AI technology enters the era of large models, when ChatGPT opens up the transformation of language interaction… The technological revolution may be another great wave in the field of life. To some extent, however, all philosophers are reminding us not to gently enter the age of symbiosis with AI, and not to be overly seduced by technological optimism and developmentalism.
“A year after the release of ChatGPT, Musk’s large model Grok will also be officially launched”
A year after the release of ChatGPT, generative AI has become a definitive wave of technology, and Elon Musk’s xAI plans to bring its large model Grok, which was released in early November, to the X platform (Twitter) this week, open to X Premium+ subscribers.
