ACM TechNews, Monday, September 23, 2024

Welcome to the September 23, 2024 edition of ACM TechNews, providing timely information for computer professionals three times a week.

U.S. to Convene Global AI Safety Summit in November

The International Network of AI Safety Institutes will hold its first meeting Nov. 20-21 in San Francisco to discuss priority work areas and "advance global cooperation toward the safe, secure, and trustworthy development of artificial intelligence." The meeting will involve technical experts from the AI Safety Institutes, or equivalent government-backed safety office, of member nations, which include Australia, Canada, the EU, France, Japan, Kenya, South Korea, Singapore, the U.K., and the U.S.
[ » Read full article ]

NSF Investing $48M to Broaden Participation in Computing

The U.S. National Science Foundation (NSF) awarded $48 million to the NSF Broadening Participation in Computing Alliances to boost participation of underrepresented groups in computer and information science and engineering. Recipients include the Computing Research Association, Morehouse College, the New Jersey Institute of Technology, the New York City Foundation for Computer Science Education, the University of Colorado at Boulder, the University of Washington, the University of Chicago, the University of Texas at Austin, and the University of Texas at El Paso.
[ » Read full article ]

Tech Jobs Have Dried Up — and Aren’t Coming Back Soon

Indeed.com reported a more-than-30% decline in postings for software development jobs since February 2020, and Layoffs.fyi has tracked about 137,000 layoffs by tech companies since the start of 2024. Tech workers are scrambling for fewer positions as tech firms cut their recruiting teams, pull back on entry-level hires, and concentrate on revenue-generating projects and services. Said Cornell University's Jason Greenberg, "This is not 2012 anymore. It's not the bull market for college graduates."
[ » Read full article *May Require Paid Registration ]

Cat's Eye Camera Can See Through Camouflage

A camera developed by researchers at South Korea's Gwangju Institute of Science and Technology can distinguish objects from their background regardless of the time of day thanks to modifications inspired by cats' eyes. The camera mimics cats' vertical pupils and the reflective structure behind their retinas, with a 3D-printed vertical slit serving as an aperture and a silver reflector placed under each of the camera's photodiodes.
[ » Read full article ]

Instagram Makes Teen Accounts Private

As of Sept. 17, youth under age 18 in the U.S., U.K., Canada, and Australia will receive restricted teen accounts when they sign up for Instagram. Existing accounts for those under 18 will be migrated to restricted accounts within 60 days. Additionally, "sensitive content" will be limited, and notifications will be sent to teens using the app for more than an hour.
[ » Read full article ]

C++ Alliance Takes Aim at C++ Memory Safety

The C++ Alliance said it plans to submit its Safe C++ Extensions proposal to the ISO for inclusion in the C++ standard, with the goal of advancing a superset of C++ with a "rigorously safe subset" to ensure C++ code has the same safety guarantees as code written in Rust. The White House in February had called on developers to stop using C++ and the C language due to concerns regarding memory safety.
[ » Read full article ]

Robot Skeleton Helps Paralyzed Woman 'Walk Again'

The Exomotus M4 exoskeleton is helping patients at the Morello Clinic in Newport, Wales, U.K., walk again. Sophie Shuttleworth, who is paralyzed from the chest down after contracting viral meningitis, was able to stand and walk for the first time in two years thanks to the exoskeleton and hopes it will help her to regain her independence.
[ » Read full article ]

How an Unknown Chinese Phonemaker Took Over Africa

Chinese manufacturer Shenzhen Transsion Holdings Co. accounts for around half of Africa's smartphone market. Transsion is known for its focus on African consumer habits, having rolled out handsets with multiple SIM card slots, as African consumers often switch telecommunication providers to obtain better service. Transsion's Itel, Tecno, and Infinix brands are available only in Africa, but the company also is gaining a foothold in India, where its budget 5G phones are popular.
[ » Read full article *May Require Paid Registration ]

'There's a War for the Top 1%': Inside French Tech's Fierce Battle for the Best AI Talent

AI startups in Paris are courting top engineers from big tech firms. Said Mathias Frachon of tech recruitment firm The Product Crew, "There's a war only for the top 1%, but they are superstars and everyone is fighting over them." Paris is the focus of the latest AI talent war, since France is home to several prestigious universities known for producing top AI talent, prompting big tech firms like Facebook and Google to open research labs in the city.
[ » Read full article ]

Brain Implant Lets People Control Alexa with Their Minds

Synchron said its brain-computer interface (BCI) can allow users to control Amazon's Alexa with their minds. One of the 10 people with Synchron's implanted BCI has been given access to Alexa via Bluetooth. The Amazon integration is important because people with disabilities generally operate virtual assistants with voice commands that can be overheard. Said Case Western Reserve University's Emily Graczyk, "Restoring any amount of independence is really important to people, but restoring independent private use is even better."
[ » Read full article ]

Researchers Uncover RCE Exploit in Google Cloud

Tenable researchers identified a significant vulnerability in Google Cloud Platform (GPC) that could have enabled threat actors to run malicious code on millions of GPC servers remotely. The researchers said the "dependency confusion" flaw known as CloudImposer was located in GCP's Composer dependency installation process, and would have allowed hackers to upload a malicious package to PyPI that would be preinstalled on all Composer instances with high permissions. The flaw has been fixed.
[ » Read full article ]

Ban Warnings Fly as Users Probe the 'Thoughts' of OpenAI's Latest Model

OpenAI reportedly has sent warning emails threatening to ban users who attempt to determine how its newest "Strawberry" AI model works. With the o1 model, users can see a filtered interpretation of its chain-of-thought-process in the ChatGPT interface, but its raw chain of thought is hidden from users. Marco Figueroa, manager of Mozilla's GenAI bug bounty programs, said the move prevents positive red-teaming safety research from being performed on the model.
[ » Read full article ]