Axios
AI startup Anthropic said in a blog post Thursday that suspected Chinese state-backed hackers had used Anthropic’s Claude Code agent to automate cyberattacks on roughly 30 global organizations. Anthropic said the attackers jailbroke Claude by posing as a legitimate company and breaking malicious tasks into smaller steps to evade safeguards. Once compromised, Claude autonomously scanned systems, wrote exploit code, created backdoors, and exfiltrated data with minimal human oversight. Four breaches succeeded, driven by AI-enabled attack speeds far beyond human capability.
From "China-linked Hackers Used Anthropic's AI Agent to Automate Spying"
Axios (11/13/25) Sam Sabin
View Full Article
Reuters
IBM announced it has developed a new experimental quantum computing chip, which the company said constituted a milestone toward its goal of building practical quantum computers by 2029. The “Loon” chip integrates advanced quantum connections between qubits, making it more complex but potentially more powerful. IBM also unveiled a second chip, “Nighthawk,” which the company said could outperform classical computers on select tasks by 2026.
From "IBM Says 'Loon' Chip Shows Path to Useful Quantum Computers by 2029"
Reuters (11/12/25) Stephen Nellis
View Full Article
Financial Times
The U.K. plans to introduce a cybersecurity and resilience bill granting regulators powers to fine companies that fail to meet stricter cybersecurity standards. Regulators could impose penalties of up to 4% of a company’s annual turnover or £17 million (U.S.$22 million), whichever is higher, for failing to report major cyber incidents within 24 hours, or to prepare adequately for breaches. The bill expands existing regulations, requiring sectors such as healthcare, IT services, and datacenters to comply with the National Cyber Security Center’s Cyber Assessment Framework.
From "U.K. Regulators to Gain Greater Powers over Cybersecurity Failures"
Financial Times (11/12/25) Kieran Smith
View Full Article - May Require Paid Subscription
CNN
Brazilian startup re.green is using drones, AI, and satellite data to restore degraded areas of the Amazon and Atlantic forests. The company’s algorithms identify land for reforestation, recommend techniques, and model financial returns through carbon credits or sustainable timber. Re.green grows native seedlings and uses drones for planting in remote regions, aiming to recreate natural ecosystems. The company already has planted over 6 million trees.
From "Drones, AI Protect Brazilian Rainforest"
CNN (11/12/25) Nell Lewis
View Full Article
Interesting Engineering
Researchers at Germany's Jülich Supercomputing Center and Nvidia simulated a full 50-qubit universal quantum computer using the JUPITER exascale supercomputer. The 50-qubit simulation used about 2 petabytes of memory, leveraging the Jülich Universal Quantum Computer Simulator's advanced simulation software and more than 16,000 Nvidia GH200 Superchips. External researchers and companies will have access to the simulation via integration with JUNIQ, the Jülich UNified Infrastructure for Quantum Computing.
From "JUPITER Supercomputer Runs Full 50-Qubit Simulation"
Interesting Engineering (11/11/25) Neetika Walter
View Full Article
Nikkei Asia
India's Ministry of Education said the nation's AI education curriculum will be expanded from students aged 11 to 17 to those as young as 8 years old beginning next year. The ministry said AI education "will be organically embedded from the foundational stage, beginning in grade three." However, there are concerns about the plan, given that one-third of Indian schools lack computers and Internet access.
From "India to Extend AI Education to Students as Young as 8"
Nikkei Asia (11/10/25) Viren Naidu; Yuji Kuronuma
View Full Article - May Require Paid Subscription
