Technology notes

News and notes on AI, chips, and IT.

All posts

OpenAI Astra Rated 'Critical' for Cyber Capabilities… First Model to Receive Highest Risk Rating

OpenAI has officially rated the cybersecurity capabilities of its next-generation AI model, Astra, as “Critical” under its Preparedness Framework. This is the first time OpenAI has assigned this rating to an actual model. The key point of this announcement isn’t simply that coding performance has improved. Given the right tools and access privileges, Astra can identify unknown vulnerabilities in well-protected systems and orchestrate attacks without step-by-step instructions from a human.

Sep 2, 20266 min read

Industry & policy

OpenAI Officially Rebuts Apple's Trade Secret Lawsuit… Competition for AI Hardware Talent Heads to Court

OpenAI has again refuted Apple’s trade secret infringement lawsuit in court documents. Apple claims that employees who left to join OpenAI took hardware design, manufacturing, and supply chain materials with them, but OpenAI counters that there is a lack of concrete evidence that trade secrets were stolen. OpenAI argues that this lawsuit is intended to prevent the recruitment of talent by competitors and to discourage employee mobility. Regarding Apple’s information management and resignation procedures, Given that there was a problem, responsibility

Sep 1, 20267 min read

Security & privacy

Meta Removes Ads for Malicious Apps Following Warning from the Indian Government… 39 Ads Still Displayed After the Warning

After the Indian government warned that malicious Android apps targeting financial information were spreading through Facebook and Instagram ads, Meta removed dozens of related ads. The problem is that even after the government’s warning, Reuters personally verified at least 39 active ads. These ads used bait reminiscent of adult content to lure users to external sites, where they were then prompted to download APK files from outside official app stores. That was the method. The malware that had been installed

Sep 1, 20267 min read

IT guides & reviews

Adobe and Saudi Arabia to Roll Out $4 Billion Worth of AI Tools… Offering One Year of Free Access to 27 Million People

Adobe has expanded its partnership with Saudi Arabia’s Ministry of Communications and Information Technology and AI company HUMAIN to provide generative AI creation tools free of charge for 12 months to more than 27 million citizens and residents. Adobe stated that the total value of the offering exceeds $4 billion. Eligible citizens and residents aged 13 and older will have access to Adobe Firefly Standard and Adobe Express Premium features. The offering will be available through the end of 2026.

Sep 1, 20267 min read

Semiconductors & infrastructure

Keppel Acquires Two Data Centers in Tokyo… Share of Rental Revenue from Japan Rises to 23%

Singapore-based Keppel DC REIT and its parent company, Keppel, have agreed to acquire a 90% effective stake in two data centers in the Tokyo metropolitan area of Japan. Amid growing demand for data centers driven by the expansion of AI services, this is a major transaction aimed at securing long-term rental income in the Japanese market. The acquisition price for 100% of the two assets is 190 billion yen, which amounts to approximately $1.19 billion or S$1.5485 billion based on the exchange rate at the time of the announcement. Upon completion of the transaction, Keppel DC REIT

Sep 1, 20267 min read

Security & privacy

Anthropic Resumes External Cyber Assessment of Claude… Introduces Real-Time Blocking Mechanism

Anthropic has resumed external cybersecurity evaluations of its pre-release models, which had been suspended following a security incident. This resumption is not merely a return to the original schedule; it comes after the addition of multiple layers of defense, ranging from sandboxing and network isolation to real-time monitoring. The most significant change is the implementation of real-time detection that immediately flags instances where the model excessively explores the test environment, attempts to go beyond its boundaries, or establishes unexpected internet connections. It's the classifier. If the classifier indicates a dangerous action, before calling the tool,

Sep 1, 20266 min read