Here's a look at today's AI briefing: - Microsoft VP joins OpenAI's board as a non-voting observer.
- Winter storms will test the latest AI weather prediction tools.
- Palantir's stock downgraded after analysts cite "overhyped" AI.
Thanks for reading! Beth p/beth-duckett | |
1 | Dee Templeton of Microsoft has joined OpenAI's board as a non-voting observer, Bloomberg reported. Templeton, Microsoft's VP for technology and research partnerships and operations, has gained access to meetings and confidential information but not voting rights. More: - Templeton is the first representative from Microsoft, OpenAI's biggest investor, to join its board after CEO Sam Altman's temporary ouster.
- OpenAI's board ended Altman's employment on Nov. 17 following a "breakdown of communications." Less than a week later, Altman was re-hired.
- Upon his reinstatement, Altman said Microsoft would gain a non-voting observer seat on the nonprofit board overseeing OpenAI.
- The company has a 49% stake in OpenAI's for-profit unit.
Zoom out: - After Altman's return, the company revamped its board, which now consists of former Salesforce CEO Bret Taylor, who is chairman, along with former Treasury Secretary Larry Summers and prior member Adam D'Angelo, the co-founder and CEO of Quora.
- According to a new report by The Information, OpenAI's board is seeking other new members. Among the potential candidates are Scale AI CEO Alexandr Wang and ex-GitHub CEO Nat Friedman.
- While other candidates are unknown, the board plans to include reps from firms collaborating with or involved in the AI sector, though talks are still at an early stage.
| | |
2 | AI models like Google DeepMind's GraphCast are supplementing traditional weather prediction tools and will soon be tested in winter storms across North America and Europe, according to Axios. The publication notes that the public will soon see how AI computer models can predict complex winter storms, offering a comparison to tools like the GFS and the European model. More: - AI models learn from past data to predict complex systems, while numerical models use physics equations and observations for simulations.
- Companies venturing into weather and climate AI include Google, Nvidia, IBM, and Tomorrow.io, as well as government agencies such as NASA and NOAA.
Zoom out: - Google DeepMind recently detailed GraphCast, an advanced AI weather prediction model that delivers 10-day forecasts in under a minute.
- The AI model, trained on almost 40 years of historical data, can generate 10-day forecasts at six-hour intervals for global locations.
- In September, GraphCast predicted Hurricane Lee's Nova Scotia landfall nine days ahead, surpassing traditional forecasts by three days, according to RΓ©mi Lam, a Google DeepMind research scientist.
- GraphCast also outperformed the gold-standard "European model" by around 10% on over 90% of assessed weather variables.
| | |
3 | Palantir's stock fell over 3% on Friday following a downgrade to "underperform" by Jefferies analysts, who cited the company as "overhyped" amid the AI enthusiasm. Jeffries Senior Analyst Brent Thill suggested that Palantir's technology is promising but overhyped, noting he expects slower-than-expected revenue growth from AI adoption than what the market currently anticipates. What happened: Palantir shares fell following Jefferies' downgrade from "hold" to "underperform" and a price target cut to $13 from $18, citing an "unjustifiable" valuation. In a note, the analysts cited concerns that Palantir's shares have "rallied to unsustainable valuation levels primarily on the back of AI euphoria (and retail trading momentum) with no monetization strategy." What it means: The downgrade suggests that Palantir's valuation may have been excessively inflated due to overestimated expectations of an impending AI boom. Still: Palantir's AI platform users nearly tripled during the July-September period, according to CRO Ryan Taylor. In November, Palantir reported its fourth straight quarterly profit, fueled by rising demand for its AI and data analytics services. Meanwhile, analysts recognize Palantir's long-term AI tech advantage but view its current valuation as risky, suggesting investors wait for the AI hype to die down before investing in Palantir and other AI companies at potentially lower levels. | | |
4 | A database allegedly listing thousands of artists whose styles Midjourney's AI image generator can mimic was shared publicly, sparking criticism on social media. The Google spreadsheet, called "Midjourney Style List," lists 16,000 artists, including Frida Kahlo, Walt Disney, Banksy, Yayoi Kusama, David Hockney, and KAWS. More: - The spreadsheet not only lists artists but also labels for image styles like "cyberpunk" and "zombiecore," along with periods, artistic movements, mediums, genres, and video game software.
- The database stems from a court case alleging that Midjourney's text-to-image tool illegally profits by copying creators' work without consent.
- While the spreadsheet is no longer directly accessible, it can still be viewed through the Internet Archive and a court document.
Zoom out: - The document stems from the class action lawsuit filed by several artists against Midjourney, DeviantArt, Stability AI, and Runway AI, alleging unauthorized use of their works to train AI.
- It claims that Midjourney's AI was trained on the listed artists' works and that CEO David Holz created the Google Sheet for the AI to mimic these artists' styles based on user prompts.
| | |
5 | Google's Project Green Light, which uses AI to help improve traffic flow, is now being piloted in Seattle. The project leverages Google's AI and Maps data to analyze travel patterns and suggest ways that city officials can optimize traffic lights. More: - According to project officials, minor changes in traffic light timings can greatly improve traffic flow.
- Google estimates its approach could cut stop-and-go traffic by as much as 30% and vehicle emissions at intersections by up to 10%.
- Seattle is the first U.S. city to pilot Project Green Light, which is also undergoing trials at 70 intersections across a dozen other cities worldwide.
- Seattle has already seen reduced congestion by shifting four seconds in traffic light timings from a north-south street to an east-west street.
- While the project is still in the early research stage, Google is inviting cities to join for free.
| | |
6 | Weekly roundup — The top AI stories you might have missed this week: - A leaker claims Apple is incorporating generative AI into Siri for more natural conversations and user personalization. The blogger claims Apple will preview the "new Siri" at its 2024 Worldwide Developer Conference in June.
- Intel partnered with DigitalBridge, an investment firm, to form a spinoff focused on generative AI software solutions for enterprises. The independent company, called Articul8 AI, will promote Intel's AI hardware suite and allow businesses to train and deploy AI models based on their own data.
- Samsung will unveil new Galaxy devices offering "an all-new mobile experience powered by AI." Samsung will unveil the new devices, presumably its Galaxy S24 smartphones, during its Unpacked event on Jan. 17.
- Microsoft is adding a dedicated "Copilot" key to keyboards on some new Windows 11 PCs. The key, which will summon Microsoft's AI-powered assistant with one click, is described as "the first significant change to the Windows PC keyboard in nearly three decades."
- U.S. Supreme Court Chief Justice John Roberts has predicted that AI will "significantly" impact judicial work, especially at the trial level. In his year-end report published on Sunday, Roberts said courts need to better incorporate new tech into their operations but warned that AI requires "caution and humility."
| | |
7 | QUICK HITS - Need to close more deals faster? See 6 tips on how Calendly helps sales & marketing teams automate the meeting lifecycle and grow revenue. Read the blog.*
- Italian Prime Minister Giorgia Meloni said that addressing AI's risks is a major focus during Italy's one-year G7 presidency. Meloni wants to hold a special session on AI before June.
- Yann LeCun, Meta's chief AI scientist and a Turing Prize winner, revealed that he declined a 2002 offer to be Google's research director, citing the company's then-small size and a low salary offer at the time.
- An AI doppelgΓ€nger of Vincent van Gogh fielded questions from The New York Times about his life and works, answering some of them incorrectly.
*This is a sponsored post. | | |
Upcoming Events | JAN 30 | Join TruEra for an expert discussion about the impact of the EU AI Act on US and Canadian firms | | | | | * This is a sponsored event | | | |
| AI and technology writer | Beth is a contributing editor and writer of Inside's AI and Tech newsletters. She has written for notable publications like USA Today and the Arizona Business Gazette. During her time as a public policy reporter at The Arizona Republic, she received recognition with a Pulitzer Prize nomination and a First Amendment Award for her collaborative reporting on pension cost increases. Beth also authored a book on the solar photovoltaic industry in 2016. You can reach her at Beth.Duckett@yahoo.com. | This newsletter was edited by Aaron Crutchfield | |
|
|
Comments
Post a Comment