Brain-Computer Interface Industry Alliance 2024 First Plenary Meeting to be Held Soon to Release Brain-Computer Interface Industry Alliance Innovations
The first plenary meeting of Brain-Computer Interface Industry Alliance 2024 will be held on March 11th in Beijing. Achievements in Academic/Hardware/Software & Platform/Industry Application categories will be released. The Chinese Expert Consensus on Clinical Research of Brain-Computer Interfaces for Neurological Diseases will also be re-released.
ChatGPT consumes over 500,000 kWh of electricity per day, more than 17,000 times more than U.S. households.
ChatGPT consumes more than 500,000 kilowatt-hours of electricity per day to respond to approximately 200 million requests from users. In comparison, the average U.S. household uses about 29 kWh of electricity per day By 2027 the entire AI industry will consume between 85 and 134 terawatt hours per year.
Chip and AI stocks follow Nvidia lower, profit-taking momentum after big gains
Asian chip and artificial intelligence (AI)-related stocks fell, with NVIDIA and other tech stocks falling heavily in the U.S. market on Friday, with significant profit-taking momentum after recent big gains NVIDIA was down 5.61 TP3T on Friday, and the Philadelphia Stock Exchange Semiconductor Index was down 41 TP3T.
Philosopher 10,000 words to analyze the nature of Sora, Yang Likun praise forwarding AI video is how far away from the world simulator?
Meta Chief Scientist Likun Yang liked and shared a 10,000-word blog post, "Is a Video Generator a World Simulator? The article asks questions and explains how Sora works, the simulation hypothesis, intuitive physics, the definition of a world model, and image generation.
Jen-Hsun Huang's next-gen DGXAI system will be liquid-cooled
NVIDIA CEO Jen-Hsun Huang has confirmed that its next DGXAI system will be liquid-cooled. This opens up new opportunities in the data center space. The data center and AI server space has been accelerating the pace of liquid cooling, investing huge sums of money in liquid cooling equipment.
OpenAl Palace Fights Investigation Results Released, Altman Returns to Board of Directors
OpenAI announces the results of an investigation into the previous CEO ouster; CEO Sam Altman is innocent and will return to the board.OpenAl Board Announces Overhaul of OgenAl Governance Structure, Announces Election of Three New Board Members, Including:Former CEO of the Bill and Melinda Gates Foundation, a former...
Windows moves into the AI era! Microsoft's first AIPC is here
Microsoft recently announced the launch of two new "AI PC" products, which will be equipped with Intel Core Ultra or Qualcomm Snapdragon X Elite processor these processors are built-in the latest neural processing unit NPU.
GPT-3.5 May Raise Risk of Mass AI Hiring Discrimination
Recently the large model GPT-3.5 developed by OpenAl appeared to discriminate against certain black names when serving recruiters. According to a Bloomberg investigation, using generative AI for recruiting poses the risk of large-scale discrimination.
Zhu Songchun, member of the National Committee of the Chinese People's Political Consultative Conference (CPPCC) and president of the Beijing Institute of General Artificial Intelligence: general artificial intelligence should meet 3 basic conditions
General artificial intelligence to meet the basic conditions: ① can complete unlimited tasks; ② can be in the scene of the active, autonomous discovery of the task, that is, "eyes in the work", ③ by the value of the autonomy of the drive." To win this technological competition, the key is still in the talent."
AI startup Astera seeks $534 million IPO valued at $5 billion
According to Bloomberg, Astera Labs, an Intel-invested "AI unicorn," recently sought a $534 million IPO, valuing the company at about $5 billion.
Ant Group Launches SkySense, a Multimodal Remote Sensing Model
Ant Group launched 2 billion parameter multimodal remote sensing base model SkySense, which is the latest research and development achievement of Ant Bailing Big Model in the field of multimodal, and the paper has been received by the world's top meeting of computer vision, CVPR 2024.SkySense exceeds the indexes of the international similar products in 17 test scenarios, and it can be used for geomorphology and crop observation.
Komo CEO: AI is far from a passing hype, it's a truly useful tool
According to Jamie Dimon, CEO of JPMorgan Chase, Artificial Intelligence (AI) is definitely not short-term hype, but a great tool with a wide range of applications. JPMorgan Chase is looking internally at how to use the new technology, as AI will eventually be used for almost all jobs.
Li Want to talk about Apple's abandonment of car-making and shift to artificial intelligence: absolutely right
Apple gave up building cars and chose to focus on artificial intelligence is absolutely the right strategic choice, and the timing is also appropriate. Li Want said, "made the car, a big success, Apple will increase the market value of 2 trillion dollars, but the car is a big success necessary conditions are still artificial intelligence.
Ali Releases Portrait Video Generation Framework EMO
A team from Alibaba has released an audio-driven portrait video generation framework, EMC (Emote Portrait Alive), which was published simultaneously in arXiv. Inputting a reference image and sound audio, the framework is able to generate audio portrait videos with rich facial expressions and head poses.
Pika launches 'lip-sync' feature
AI video generation platform Pika has announced the launch of LipSync, a lip-syncing feature that voices the characters in the generated video and "lip-syncs" them at the same time, which is currently only available to Pro users. The audio generation is powered by AI voice cloning startup ElevenLabs.
Google Releases 11B Base World Model Generating Interactive Virtual Worlds
Google releases Genie, a base-world model trained on Internet video with a parameter scale of 11B, which can generate a myriad of motion-controllable interactive environments from composite images, photographs, and even sketches.Genie can be cued with images it has never seen before, the paper's address:https://arxiv.org/abs/2402.15391
Nanjing University Offers AI Liberal Studies Core Curriculum System
Nanjing University held a new semester work arrangement meeting, announcing that it will open "Artificial Intelligence General Knowledge Core Curriculum System" for all undergraduates. Nanjing University will build a "1+X+Y" three-level "Artificial Intelligence General Knowledge Core Curriculum System".
NVIDIA launches RTX 500, 1000 Ada AI graphics cards
NVIDIA introduced RTX500 and RTX1000 series laptop graphics cards for laptops and mobile workstations with RTX500 and 1000 GPUs and other next-generation mobile workstations will include NPUs, CPU components and NVIDIA RTX GPUs.
Chip god Jim Keller's startup partners with Japanese firm to design AI chips
Rapidus, a Japanese government-funded semiconductor R&D group, will partner with Canadian chip startup Tenstorrent to design its first AI chip.Tenstorrent will license a portion of its design for a Japanese AI gas pedal and co-design the entire chip with RapidusTenstorrent was founded in 2016 and Rapidus was founded in 2022.
Huawei Releases First Big Model of Communication
Huawei unveils the first Big Model in the communications industry at MWC 2024, Huawei's Communications Big Model will provide two types of application capabilities, Role-based Copilots and Scenario-based Agents, which will ultimately enhance network productivity across the board.
Li Zhifei calls Sora a world simulator
Li Zhifei, the CEO of ChatGPT, analyzed the article "Why Sora is the simulator of the world?". ChatGPT can simulate the scenes and characters in the virtual world through the carrier of "language" thinking, and become the "simulator" of the virtual world.Sora is able to reproduce the physical reality in the virtual environment, and has the potential to become the leading simulator of the next generation of the physical world. Sora has the potential to be the next generation of physical world simulators.
UC Berkeley's Big World Model Tops GitHub Hot List
Large WorldModel (LWM) proposed by UC Berkeley researchers tops GitHub list, has now garnered 3.2K starred papers Address:https://largeworldmodel.github.io/
Musk: Tesla could generate real-world video a year ago
Musk said on social media platform X that Tesla was able to generate real-world video with precise physics a year ago, but the training data came from the car so generating the video wasn't interesting, so even though there was dynamically generated content that looked like it came from a Tesla video.
AGI and Sora are not greatly related, says Foucault
Fu Sheng, chairman of Orion Star, said that the official launch of Sora by OpenAl is a milestone-level event:The emergence of Sora is the result of better engineering and greater arithmetic power.Sora is equivalent to opening a door, and soon a lot of people will flock to this fieldFu Sheng also mentioned that the emergence of general-purpose AI and Sora are not very much related.
NVIDIA Senior Research Scientist Responds to Sora Video "With Sound"
NVIDIA Senior Research Scientist Linhee Fan posted a response to AI audio generation startup EleverLabs adding audio to videos generated by OpenAlSora. Currently this audio is cued by text, but proper adjustments should be made for both text and video pixels.
Using light, not electricity, to perform operations New chip opens door to light-speed AI computing
Engineers at the University of Pennsylvania have developed a new type of chip that uses light rather than electricity to perform the complex mathematical operations necessary to train artificial intelligence (AI). The chip could radically speed up computer processing while also reducing energy consumption.
Samsung gets order for 2-nanometer AI chips, sources say
Samsung Electronics has received an order from Japanese AI startup Preferred Networks (PFN) to produce AI gas pedals and other 2-nanometer process AI chips.
Meizu announces 'All in AI' will stop traditional smartphone program
Meizu has decided that it will be All in AI, stopping new projects for traditional smartphones and fully investing in Al For New Generations, the device of tomorrow. 2024 Meizu's new mobile-side operating system for the AI era will be updated with a system update.
Merchants Strategy: Sora Drives AI Video Generation into a Whole New Era
The China Merchants Strategy research report believes that a number of visual artists, designers and filmmakers (as well as OpenAl employees) have already gained access to Sora, and it is expected that as the technology matures further, it will be pushed further to all users.
OPPO founder Chen Mingyong: 2024 is the year of AI phones
Chen Mingyong, founder of OPPO, published a letter to all employees at the beginning of the year, saying:2024 is the first year of AI cell phones, and the era of AI cell phones will become the third stage of the cell phone industry after feature phones and smartphones.AI can redo the experience of the cell phone all over again.
UBS says NVIDIA dramatically shortens AI GPU delivery cycle: from 8-11 to 3-4 months
According to a recent memo shared with investors by UBS analysts, NVIDIA has dramatically shortened its delivery cycle for AIGPUs from 8-11 months at the end of last year to 3-4 months now. NVIDIA may already be in a position to meet future orders for volume production.
Sony and Seagate partner on hard disk drives for AI
Sony Semiconductor Solutions will work with Seagate to develop mass-produced hard disk drives (HDDs), and Sony will begin mass-producing the lasers used in the drives in May and will invest 240 million yuan in a new production line.
OpenAl rumored to close new deal, valuation may exceed $80 billion
OpenAl has closed a deal to bring its valuation to $80 billion or more, nearly tripling the startup's valuation in less than 10 months. The existing shares will be sold through a tender offer led by venture capital firm Thrive Capital.
Google exposed to internal AI big language model Goose
Google has quietly launched an internal AI big language model called "Goose," which has 25 years of engineering expertise at Google" and is available only to Google employees "to answer questions about specific Google technologies, write code using an internal technology stack, and support novel features such as editing code based on natural language hints. It can answer questions about specific Google technologies, write code using an internal technology stack, and support novel features such as editing code based on natural language hints.
SoftBank is seeking $100 billion for AI chip program
SoftBank Group founder Masayoshi Son is seeking as much as $100 billion to fund a chip joint venture to compete with NVIDIA Corp. and supply AI-essential chips SoftBank is considering raising money from Middle Eastern investors. The project is codenamed lzanagi.
AI startup Meshy launches Meshy-2 Vincennes 3D image models
AI startup Meshy released its second generation product, Meshy-2. Meshy-2 supports four text-to-3D graphics creation styles: realistic, cartoon, low polygon and voxel. In terms of efficiency, it takes only 25 seconds for Meshy-2 to produce a graphic preview and 5 minutes to produce fine results.
AI cracks 2,000-year-old Greek scrolls on Nature
Participants in the Vesuvius Challenge trained learning algorithms to unveil the contents of a 2,000-year-old scroll in Greek text with the help of AI, and the results of their research appeared in Nature Researchers say it could have a revolutionary impact on our understanding of the ancient world.
Fudan launches Al Agent visually impaired assistant
FudanNLP, the natural language processing laboratory of Fudan University, announced on its WeChat public platform that it has launched the "Hear the World" Agent based on the multimodal large model - MouSi - to help visually impaired groups to perceive the world again.
Ikea launches AI home assistant on GPT Store
Home furnishings giant IKEA has announced on its website the launch of the new IKEA AI Assistant an AI-powered, all-in-one tool for home design, inspiration and shopping, which has gone live on the OpenAlGPT Store.Users can use the GPT to articulate specific home design needs and get customized advice from the AI.
Tsinghua University Updates Deep Learning Framework Keitu User Documentation
The Graphics Laboratory at Tsinghua University announced that it has updated the user documentation of Jitto, a deep learning framework. The new user documentation provides more detailed descriptions of features, richer examples, and is much more user-friendly for Chinese users.