Bill Gates was one of the investors in a $105M funding round for supercomputer startup Luminous, which is developing a light-based rather than electron-based AI accelerator chip. The Mountain View, California-based startup is said it's using its own silicon photonics technology to get rid of "data movement bottlenecks" and reimagine how AI computers are built. More: - The startup is developing a supercomputer based on silicon photonics optimized to run advanced AI models, which are difficult and costly to train.
- The hardware hasn't kept up, and training big AI models is mostly relegated to Big Tech, through stagnation in AI progress is imminent due to the hardware limitations, CEO Marcus Gomez told VentureBeat.
- Luminous' technology, however, can result in "order-of-magnitude" improvements in computing performance and simplifications to the programming model, according to the startup.
- In addition to Gates, other investors in the round included Gigafund, 8090 Partners, Neo, Third Kind Venture Capital, Alumni Ventures Group, Strawberry Creek Ventures, and others.
- The funding will help Luminous double its engineering team, build out its chips and software, and prepare for commercial-scale production.
- It's now hiring for positions including a chief ML scientist and director of AI applications engineering.
SILICON VALLEY BUSINESS JOURNAL 
| |
 Fewer than 30% of drivers polled said they would be comfortable riding in their own fully autonomous vehicle, though the percentage was higher for people under age 30, according to an AutoPacific study. Among those polled, 32% said they trust Tesla to develop a safe and reliable fully autonomous vehicle, followed by Toyota (19%) and BMW (18%). More: - The automotive marketing research surveyed 600 U.S. drivers aged 18-80 about self-driving cars.
- 29% of respondents said they would be comfortable being automatically driven in their own fully autonomous vehicle.
- 26% said they would be comfortable as passengers in someone else's AV.
- Broken down by age range, 40% of people aged 18-29 said they would be comfortable versus 20% of those aged 60+.
- 41% of those 18-29 said they think AVs would result in fewer accidents compared to 24% overall.
THE AUTO CHANNEL
| |
 A message from MALWAREBYTES Secure success for 2022 with AI-powered cyberprotection. Remote workforces continue to raise security concerns for businesses. Remote workforces make easy targets for cybercriminals. - 1 out of 5 businesses experienced a remote-worker related security breach
- 28% of people admitted using their personal device for work more than their work-issued device
Secure success in 2022 Prepare your team—and all the devices they use—against ransomware, security breaches, and other cyberattacks with easy-to-use tools designed for growing businesses. Outsmart cyberattacks and avoid downtime with Malwarebytes Endpoint Detection and Response with 72-hour ransomware rollback recovery. Get started today with 25% off your first year of EP, EDR and server protection. Learn More
| |
Cedars-Sinai's medicine department launched an AI division to research how technology can solve difficult medical problems and improve health care. The Artificial Intelligence in Medicine Unit will be led by Dr. Sumeet Chugh of Cedars-Sinai's Smidt Heart Institute, who has experience using AI to address issues linked to sudden cardiac arrest. More: - The division will tap into the Los Angeles-based medical center's network of data to help design new AI approaches to health challenges.
- Its current focuses are on COVID-19, cardiac imaging, sudden cardiac arrest, and clinical genetics.
- Recently, Cedars-Sinai physician-scientists developed an AI tool that can read ultrasounds and flag patients suspected of having two potentially deadly heart conditions.
- The AI division also recently employed machine learning to identify patients at risk for heart attacks. The study was published in the Journal of Nuclear Medicine.
HIT CONSULTANT
| |
 C3.ai CEO Tom Siebel Enterprise AI software provider C3.ai reported quarterly revenue of $69.8M, an increase of 42% from the year prior. The company, led by billionaire technologist Thomas Siebel, expects revenue for fiscal 2022 to range from $251M to $252M, up 38% from the previous year. Siebel said the company's revenue was driven by a sales refocus, higher customer count, and more industry diversification of its AI products. More: - The company reported its results that exceeded analysts' expectations for fiscal Q3, which ended Jan. 31.
- Subscription revenue was $57.1 million, a 34% jump from a year ago.
- It lost $0.07 a share on $69.8M in revenue, exceeding expectations of a $0.26 loss per share on $67.16M in revenue.
- The U.S. Defense Department recently awarded a $500M contract to C3.ai to use its AI suite and defense/intelligence applications and accelerate modeling and simulation research.
- In related news, Siebel disclosed that its CFO, Adeel Manzoor, recently resigned due to reasons "highly personal and sensitive in nature," which involved a personal relationship. It's C3.ai's third CFO to leave in 17 months.
BARRON'S
| |
 A message from LABELBOX Labelbox's on-demand webinar series, ML Unboxed, offers a deep dive into essential ML techniques and emerging concepts. In this episode, you will learn how to dramatically reduce labeling time and associated cost with model-assisted labeling. Teams using model-assisted labeling have seen a savings of 50-70% in terms of annotation costs and can dramatically improve overall project efficiency. We’ll explore workflows to increase labeling velocity whether you’re actively iterating on production models or are just getting started on your AI journey. Learn how to: - Use an open source model to accelerate your first batches of training data
- Use your own model to pre-label data so labeling teams can focus on adjusting annotations vs. starting from scratch
- Use the model of your choice to apply annotations for use in weak supervision workflows
Watch Now
| |
Dartmouth researchers created an AI model that they say can detect mental disorders in Reddit conversations. The non-peer-reviewed pre-print study was published on arXiv.org. More: - Their model focuses on emotional transitions, such as moving from anger to joy, rather than the content of the text the post, as is often used in other mental-health screening tools.
- The system was trained to label emotions expressed in Reddit posts and then map the so-called emotional transitions between posts.
- The model can detect emotional disorders, such as major depressive and bipolar disorders, through a person's emotional "fingerprint" that's compared to signature patterns of emotional transitions.
- In their study, the researchers also stressed the importance of vetting models that analyze social media data to screen for mental health, which can have biases that could widen health inequalities.
DARTMOUTH NEWS
| |
🍋 FRESH FUNDING 🍋 - Sanctuary (San Francisco, Calif.), general-purpose robots with a humanlike appearance and intelligence: $58.5M A from Bell, Evok Innovations, Export Development Canada, Magna, SE Health, Verizon Ventures, Workday Ventures.
- Qventus (Mountain View, Calif.), AI-powered software for care operations: $50M from led by Thomas H. Lee Partners, Premier, participation from Bessemer Venture Partners, Mayfield Fund, Norwest Venture Partners.
- nSure.ai (Tel Aviv, Israel), fraud prevention intelligence platform for virtual goods: $18M A led by MoreTech Ventures, participation from DisruptiveAI, Gryffin Ventures, Moneta Seeds.
- EdgeCortix (Singapore), AI-focused fabless semiconductors: $8M A led by Vajra Direct, participation from Monozukuri Ventures, Futu, et al.
- TrueCircle (U.K.), data-driven AI-based computer vision for the recycling industry: $5.5M Pre-Seed led by Lowercarbon Capital, participation from Passion Capital, Giant Ventures, firstminute Capital, et al.
For the latest on AI startup funding and more in the VC community, subscribe to our newsletter Inside Venture Capital.
| |
QUICK HITS - Skip the dentist and save up to 80% with Remi custom night guards and teeth whitening kits sent directly to you.*
- The Russian invasion has forced Ukraine's Ministry of Digital Transformation to halt efforts to encourage major tech firms to open AI R&D centers in the country.
- Elon Musk invited the United Auto Workers union to hold a vote on worker unionization among Tesla's employees without intervention from the EV maker.
- A UBS machine-learning model has predicted how the S&P 500 will respond to three Russia-Ukraine scenarios.
- Get world-class wine tailored to your taste & delivered to your door with Firstleaf! Try 6 wines for $29.95 + free shipping. Satisfaction guaranteed.*
*This is sponsored content.
| |
Upcoming Events: - March 02 - March 03 - Actionable tips for your ecommerce. Join 50+ inspiring DTC leaders (feat. TV personality Marcus Lemonis, The Profit) (Register Here) *
- March 08 - Digital Identity in the Age of Virtual Worlds with David Lucatch (Liquid Avatar) (Register Here)
- March 08 - Venturing in VC - Episode #10 with Alda Leu Dennis (Initialized Capital) (Register Here)
- March 10 - NoCode's impact on the Gig Economy with Sharath Kuruganty (Register Here)
- March 17 - Understanding Million Dollar Virtual Real Estate with Neil Mandt (Metaverse Rights) (Register Here)
- March 23 - Need to Know NoCoder: Nichole Bestman (Bubble) (Register Here)
- March 31 - How To Build In Web 3.0 (Register Here)
- April 04 - Blockchain and the Law (Register Here)
- April 06 - Entertainment in the Metaverse (Register Here)
- April 14 - Shopping In The Metaverse (Register Here)
- April 28 - Alternative Investments 2: Crypto Safety Standards (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. | |
| |
Labelbox is a training data platform designed to increase model performance faster and with less data. | |
Comments
Post a Comment