Last Week's Tech Highlights

Pentagon Accelerates "Kill Chain" with AI

The Pentagon announced it is leveraging generative AI technologies from companies like OpenAI, Anthropic, and Meta to speed up its "kill chain" process by enhancing threat detection, tracking, and assessment. Dr. Radha Plumb, the Pentagon’s Digital and AI Chief, emphasized that these systems are only used in planning stages, with human oversight retained throughout decision-making processes. Partnerships with companies like Lockheed Martin and Palantir demonstrate AI's potential in improving military planning scenarios. However, the Pentagon maintains a policy against fully autonomous weapons due to ethical concerns.

Trump Signs Executive Order to Delay TikTok Ban

President Donald Trump signed an executive order delaying the enforcement of TikTok’s U.S. ban by 75 days. The ban, triggered by national security concerns, required ByteDance to sell TikTok by January 19 or face shutdown. Trump suggested the U.S. should have a stake in TikTok and warned of imposing tariffs on China if no deal is reached.

Crisis Apps: A Lifeline for Millions

Apps like Watch Duty in Los Angeles provide real-time wildfire updates, including evacuation orders and weather reports, sourced from 911 data and police scanners. In Ukraine, Diia offers air raid alerts and access to vital documents, while military apps like EPPO track enemy movements. In Israel, Red Alert warns citizens of rocket attacks. These crisis apps have become indispensable tools for survival during emergencies worldwide.

UBTech and Foxconn Use Humanoid Robots for iPhone Production

UBTech has partnered with Foxconn to deploy Walker S1 humanoid robots in iPhone production. Standing 1.72 meters tall and weighing 76 kg, these robots handle quality checks, screw tightening, assembly, and load carrying. After training at Shenzhen factories, the robots will undergo feasibility tests at Foxconn's Zhengzhou facility. This partnership marks a significant step in reducing production costs and advancing industrial robotics.

HPE Investigates Alleged Data Breach by Hacker

Hewlett Packard Enterprise (HPE) is investigating claims by hacker IntelBroker, who allegedly stole data, including Zerto and iLO source codes, GitHub repositories, and private API access credentials. While HPE activated cyber-response protocols and disabled compromised credentials, it stated that customer data remains unaffected.

Google Unveils Titans Architecture for Human-Like AI Memory

Google introduced Titans, a new AI architecture that surpasses traditional Transformer models by integrating short- and long-term memory modules. Titans improves contextual understanding in large datasets and outperforms existing systems in fields like language modeling, DNA analysis, and time-series forecasting. This architecture aims to bring AI closer to human-like reasoning, but challenges in scalability and ethical use remain.

Couriers Struggle Under Algorithmic Management

Delivery drivers for platforms like Uber Eats, Deliveroo, and Just Eat report difficulties caused by non-transparent algorithms that determine job allocation and pay. Workers face risks such as sudden deactivation from facial recognition errors or working on multiple platforms. Activists are calling for greater algorithmic transparency to protect workers' rights and livelihoods.

Malware Hidden in Images: A New Cyber Threat

Cybercriminals are embedding malware into images hosted on popular file-sharing sites, bypassing traditional security systems. Malware like VIP Keylogger and 0bj3ctivityStealer spreads through fake emails and HTML smuggling techniques. Experts emphasize the importance of updated security software and caution when downloading files, even from trusted platforms.

AI Could Surpass Human Intelligence by 2027

Anthropic CEO Dario Amodei predicts AI will exceed human intelligence within the next 2-3 years. Amodei highlighted growing demand for their AI chatbot Claude and plans to deploy over a million chips by 2026. He warned of potential negative impacts and emphasized the need for careful monitoring as AI continues to evolve.

Trump’s $500 Billion AI Plan Boosts Oracle Shares

President Donald Trump announced a $500 billion investment in Stargate, a partnership with Oracle, OpenAI, and SoftBank to develop AI infrastructure. Oracle shares rose 9%, and Nvidia gained 2.4% on the announcement. Analysts believe this investment marks the beginning of large-scale U.S. AI initiatives and highlights the growing demand for energy in data centers.

Nordic Countries Lead Europe in 5G Coverage

Denmark, Finland, and Norway are at the forefront of 5G coverage in Europe, with Denmark leading at 83.4% accessibility. Early spectrum allocation, operator collaborations, and rural coverage priorities have propelled these countries ahead. Meanwhile, the UK lags behind with only 42.2% 5G coverage, despite high urbanization. Experts suggest that Nordic policies can serve as a model for countries aiming to improve 5G infrastructure.

UK Launches Fiber Internet Rollout Under Project Gigabit

CityFibre has begun deploying full-fiber internet in nine regions under the UK government’s £5 billion Project Gigabit program. Supported by £782 million in government funding, the initiative aims to connect 1.3 million rural homes and businesses with high-speed internet, reducing digital inequality. The project focuses on enhancing connectivity for small businesses and preparing the UK’s digital infrastructure for the future.

Record-Breaking 5.6 Tbps DDoS Attack Neutralized

Cloudflare mitigated a record 5.6 Tbps DDoS attack, which targeted an ISP in East Asia using 13,000 sources. The attack lasted 80 seconds, involving SYN, DNS, and UDP flood methods. Cloudflare reported a 53% rise in DDoS attacks in 2024, with telecommunications and internet service providers as primary targets. Indonesia led in initiating attacks, while China was the most frequently targeted nation.

Ransomware Attackers Target Organizations Through Microsoft Teams

Ransomware groups are leveraging Microsoft Teams to carry out phishing attacks. Attackers first bombard victims' inboxes with spam emails to distract them, then impersonate IT support staff via Teams to gain remote access. Sophos’ report highlights the growing sophistication of these tactics, emphasizing the need for employee training and robust security protocols to counter such threats.

LinkedIn Faces Lawsuit for Sharing Data to Train AI Models

LinkedIn is facing a proposed class-action lawsuit from Premium users accusing the platform of sharing their private messages without consent to train AI models. The lawsuit, filed in California, claims LinkedIn silently updated its privacy policy in September 2024 to permit such data usage retroactively. LinkedIn denies the allegations and maintains that it complies with data protection laws.

Optical Memory Technology Revolutionizes Data Storage

Researchers have developed a programmable photonic latch, a new optical memory unit that stores and processes data at the speed of light. Built on a silicon photonics platform, the device offers ultra-fast and energy-efficient data handling for AI applications, large language models, and optical computing. The technology could pave the way for a new era of high-speed data processing.

AI Detects Lung Diseases with 97% Accuracy

Charles Darwin University has developed an AI model, TD-CNNLSTM-LungNet, that detects lung diseases like pneumonia and COVID-19 with 96.57% accuracy using ultrasound videos. Combining convolutional neural networks (CNN) with long short-term memory (LSTM), the model outperforms existing diagnostic tools. It also offers transparency by visually explaining its decisions, aiding clinicians in diagnoses.

VR-Based Balance Test Reveals New Insights

A study at New York University used VR to simulate NYC subway environments, finding that sounds like white noise and metro sounds worsen balance for individuals with vestibular hypofunction. Healthy participants remained unaffected. Researchers suggest incorporating sound into balance rehabilitation programs, with VR as an effective tool.

Wearable Material Generates Electricity from Human Skin

Researchers at North Carolina State University have created a wearable material that generates up to 300 volts of electricity from human movement while ensuring comfort and low friction. Modified with amphiphilic molecules, the material combines energy efficiency with user comfort, potentially revolutionizing wearable tech.