 Facebook parent Meta announced its AI supercomputer, the world's fifth-fastest for training machine learning systems. The AI Research SuperCluster (RSC) will help Meta build better AI models, analyze text, images, and video, develop augmented reality and metaverse tools, and accelerate AI research and development. More: - The RSC is helping Meta train models in natural-language processing and computer vision for research.
- It's expected to be the world's fastest supercomputer for AI tasks once it's fully built out in mid-summer.
- The machine is made up of 760 Nvidia A100 modules with 6,080 graphics processing units. Meta plans to expand that to 16,000 GPUs later this year, boosting its AI performance by 2.5x.
- At that time, the machine will be capable of "performing at nearly 5 exaflops of mixed-precision compute."
- It will be able to train large natural language processing models 3x faster and run computer vision workflows 20x faster than Meta's 2017-era Nvidia machine.
Zoom-out: - According to CEO Mark Zuckerberg, the RSC will enable AI models that can understand hundreds of languages and learn from trillions of examples.
- It will fuel experiences for the company's metaverse plans, which require compute power of quintillions of operations/second, he noted.
- The supercomputer took two years of work and was built with partners Nvidia Inc., Penguin Computing Inc., and Pure Storage Inc.
- Meta didn't say how much the machine cost to build or where it's located.
CNET 
| |
 SenseTime launched operations at its Shanghai-based AI data computing center, set to become one of the largest in all of Asia. The firm's computing power base can train ultra-sophisticated AI models and lower the barrier for industries to pursue AI applications. More: - The $880M center in Lin-gang Free Trade Zone contains 5,000 computer racks and a designed computing capacity of 3.74 exaFLOPS.
- An exaFLOP is equal to one quintillion floating-point operations per second. By comparison, Tesla's supercomputer, the fifth most powerful in the world, achieved 1.8 exaFLOPS of performance last year.
- AI computing can lower the energy costs and usage of traditional algorithm methods. The center uses "energy-saving power consumption systems" that are expected to save 45 million kilowatt-hours of power consumption each year, according to SenseTime.
Zoom-out: - SenseTime is building data centers using a portion of the funds from its recent $700M initial public offering on the Hong Kong Stock Exchange.
- The center is expected to contribute to the growth of China's smart cities as well as local R&D in Shanghai, southern Jiangsu, and northern Zhejiang.
- A Wuhan Artificial Intelligence Computing Center, sponsored by Huawei and the city's government, launched in May 2021.
GLOBAL TIMES
| |
 A message from SPLINTERLANDS Splinterlands is leading the charge as Pay To Earn gaming quickly becomes the industry standard. The world’s most popular blockchain game saw record sales with their latest Chaos Legion card packs launch on January 17th. Fans raced to purchase card packs that they can then battle with to earn crypto. - Earn daily airdrops just from playing the game or holding assets. Players are experiencing up to 40% APR back from their staked tokens.
- In Splinterlands, your cards are NFTs that can be bought, rented, or sold at any time–just like physical trading cards!
- Rare and powerful gold foil cards are up for grabs whenever you open your new packs.
- The tokens you earn in this trading card battler are yours to trade and swap on global markets.
Sign up to Splinterlands for free and join millions of other players! It's time to start your Summoner’s journey and enter the fray. Sign Up Now
| |
 Zoom and three of its early investors took part in a $4.7M seed funding round for AI notetaking app Fathom. Zoom investors Maven Ventures, Bill Tai, and Matt Ocko participated along with the CEOs of Reddit, Twitch, Cruise, Mercury, People.ai, Snapdocs, and Shogun. More: - Fathom, which announced its official launch today, was founded in 2020 and was a member of the Y Combinator Winter ‘21 cohort.
- Its AI-based notetaking app, exclusive to the Zoom app marketplace, automatically records, transcribes, and creates summaries of Zoom calls.
- The summaries can be emailed or shared across Slack channels and CRM platforms such as Salesforce.
- The round also included participation from Zoom’s own Apps Fund and early-stage VC firms Character, Active Capital, Global Founders Capital, Rackhouse.vc, and Soma Capital.
- The fresh funding will help Fathom grow its go-to-market team and launch integrations with sales, customer success, and productivity tools.
TECHCRUNCH
| |
The Brookings Institution is accepting questions for a Jan. 31 virtual event that will discuss responsible AI in the military. The think tank's "Responsible AI from principles to practice" event will delve into the ethical issues of integrating AI into military applications in the U.S. and allied nations. More: - According to Brookings, responsible AI refers to ethical and accountable artificial intelligence that is geared through frameworks to be "transparent, fair, secure, and inclusive."
- Next week's event will focus on different countries' AI principles and the lessons learned from putting them into practice.
- Viewers can submit questions at events@brookings.edu or on Twitter using the hashtag #ResponsibleAI.
Zoom-out: - The U.S. Department of Defense adopted its five principles to support safe and ethical AI in 2020.
- In November, the Pentagon's Defense Innovation Unit issued its own responsible AI guidelines that third-party contractors will have to follow when they develop AI technologies for the military.
- The Pentagon's Joint Artificial Intelligence Center is currently seeking a permanent head of responsible AI after Alka Patel departed in October.
THE BROOKINGS INSTITUTION
| |
 A message from MODERN TREASURY Why you need to care more about payment operations. If your business moves money, you have payment operations. Payment operations include processes to initiate, approve, track, and reconcile payments. Just like the money they manage, payment operations are critical to a functioning business. In fact, survey data show that 86% of finance leaders view upgrading their payment operations as a priority. To better understand the current state of payment operations, Modern Treasury partnered with the Harris Poll to survey 300 leaders and uncover their top payment operations problems and priorities. In our first-ever report on payment operations, we reveal: - The costliest payment-related challenges companies face
- The 5 upgrades to payment operations leaders believe are most important
- The 7 most common barriers to upgrading payment operations
And more. Download the Report
| |
American Financial Group (AFG) said it has acquired Verikai, developer of a machine learning tool for assessing insurance risk, for $120M in cash. Verikai's predictive data and risk tool is geared toward health insurance underwriters and brokers. More: - American Financial subsidiary Great American Insurance Group is expected to gain the risk tool and Verikai's marketplace platform.
- The group plans to enter the medical stop-loss business with a focus on small and underserved risks.
- Verikai, founded in 2018, will remain a standalone company and continue to be led by President and CEO Jeff Chen.
- The San Francisco-based startup last raised $6M in Series A Funding in 2020. Its co-CEO, Carl H. Lindner III, stated that more insurers are adopting technologies including AI and data analytics, robotics process adoption, chatbots, and robo-advisory.
INSURANCE JOURNAL
| |
AI Events and Conferences: - World Artificial Intelligence Cannes Festival: Feb. 10-12, Cannes, France and digital.
- Deep Learning Summit: Feb. 17-18, San Francisco and online.
- European Chatbot & Conversational AI Summit: March 1-2, Edinburgh, Scotland.
- International Conference on Intelligent User Interfaces: March 22-25, Helsinki and online.
- MLconf 2022 New York City: March 31, New York City.
- World Summit AI Americas 2022: May 4-5, Montréal, Canada.
- ICLR 2022, the International Conference on Learning Representations: April 25-29, virtual.
- International Conference and Expo On Robotics and Artificial Intelligence: May 16-18, 2022, London.
- CVPR 2022, IEEE Conference on Computer Vision and Pattern Recognition: June 19-24, New Orleans.
- Deep Learning World: June 19-24, Las Vegas.
- International Conference on Machine Learning and Data Mining: July 16-21, New York City.
- ICML 2022, International Conference on Machine Learning: July 17-23, Baltimore, in-person conference with virtual elements.
- AI & Big Data Expo Global: Dec. 1-2, London.
| |
Quick Hits - Drive more sales with partnerships. Learn how to discover, recruit, and manage affiliates, influencers, and more to build trust and grow revenue.*
- RadNet acquired AI software developers Aidence and Quantib, which will enable the imaging services firm to launch AI-based screenings for prostate and lung cancer.
- Microsoft, Tesla, and Apple are set to report earnings this week.
- Osara and Innotech collaborated to develop an AI-based picking robot system for Rohto Pharmaceutical's upcoming factory production line.
- Nearly 60% of CFOs surveyed by Workday said they're now seeking new hires with AI/ML skills.
- Create immersive virtual, hybrid, and in-person event experiences for your audience, no matter where they are — View a Demo.*
*This is sponsored content.
| |
Upcoming Events: - January 25 - Need To Know NoCoder - Leon Young (Cogniss) (Register Here)
- January 25 - Venturing in VC - Episode #4 with David Cowan (Bessemer Venture Partners) (Register Here)
- January 26 - Join Ketch & Kelley Drye's data privacy experts for a joint webinar. Privacy Priorities for 2022: Legal and Tech Developments to Track and Tackle. (Register Here) *
- January 27 - Understanding the Metaverse (Register Here)
- February 01 - Venturing in VC - Episode #5 (Register Here)
- February 03 - Blockchain Gaming & Play-to-Earn Economies (Register Here)
- February 10 - The Business of Wellness (Register Here)
- February 17 - Efficient Marketing Strategies in a Post iOS 14 World (Register Here)
- February 23 - Need to Know NoCoder: Derya Sousa (Kianda Technologies) (Register Here)
- February 24 - Alternative Investments (Register Here)
- March 03 - The $12B Appetite for Amazon Aggregators (Register Here)
- March 10 - NoCode's Impact on the Gig Economy (Register Here)
- March 23 - Need to Know NoCoder: Nichole Bestman (Bubble) (Register Here)
*This is a sponsored listing.
| |
 |  | Beth is an Inside Analyst and a former public policy/investigative reporter. She has covered AI, VR, technology, and e-commerce for Inside. When she's not busy writing, she's hiking in Arizona's desert, playing pickleball, or trying out new Mexican food recipes. |   | | Editor | Aaron Crutchfield is based in the high desert of California. Over the last two decades, he has spent time writing and editing at various local newspapers and defense contractors in California. When he's not working, he can often be found looking at the latest memes with his kids or working on his 1962 Ford. | |
Splinterlands is the world’s #1 most played blockchain game. Play Splinterlands for free today. | |
| |
Comments
Post a Comment