Anthropic's AI strategy, Apple executive exodus, critical React flaw, server price hikes...

The Information reports that Anthropic, under CEO Dario Amodei, is increasingly seen as a more measured and disciplined alternative to OpenAI, focusing on profitability and risk management. Amodei’s recent appearance at The New York Times’ DealBook summit highlighted his thoughtful approach, contrasting with Sam Altman’s "YOLO" expansion strategy at OpenAI, which now spans hardware, data centers, and even exploring rocket company acquisitions, as noted by The Wall Street Journal. This divergence, according to The Information, could position Anthropic as the preferred partner for business-oriented AI deployments, especially as it eyes profitability years ahead of its larger rival. The Information

The Verge details a major leadership overhaul at Apple, with environmental and policy chief Lisa Jackson and general counsel Kate Adams both set to retire in 2026. Meta’s current chief legal officer Jennifer Newstead will be joining Apple, assuming a combined role of general counsel and head of government affairs. These changes follow recent high-profile exits including the COO, head of AI, and lead designer. TechCrunch further emphasizes the significance, underlining the company’s ongoing organizational restructuring and the shifting of key responsibilities to new leaders such as Sabih Khan. The Verge TechCrunch

A critical security flaw in the React JS framework has been uncovered, with TechSpot warning of an almost certain exploitation risk. The vulnerability is described as a “perfect” issue—both easy to exploit and capable of causing catastrophic damage to web servers and private cloud infrastructure. Security researchers urge immediate mitigation, underscoring the severity for both web developers and enterprises relying on React-based applications. TechSpot

The Register reveals that major OEMs are preparing to raise server prices by up to 15% and PC prices by 5%, driven by soaring DRAM and NAND memory costs. The shortage, exacerbated by booming AI demand, is prompting double-digit component price hikes across the hardware supply chain. This trend is expected to impact enterprise IT budgets and cloud providers alike. The Register

TechSpot and Tom's Hardware report that AMD is increasing memory module prices for Radeon GPUs supplied to board partners, translating to expected retail price hikes in early 2026. However, Ryzen CPUs are not affected for now. Tom’s Hardware adds that new Ryzen 7 9850X3D chips will maintain the same power draw as their predecessors, offering improved performance without higher energy consumption. TechSpot Tom's Hardware

TechSpot highlights that Amazon is preparing to reduce or potentially end its reliance on the US Postal Service (USPS) for last-mile deliveries, a move that could deprive USPS of over $6 billion in annual revenue. While The Information clarifies that Amazon has not confirmed a complete contract termination, both outlets agree that Amazon is rapidly building out its own logistics network, signaling major changes in U.S. parcel delivery dynamics. TechSpot The Information

TechSpot warns that over 4 million devices have been infected by rogue Chrome and Edge browser extensions, some of which were weaponized by the China-linked group ShadyPanda. The malicious extensions are spreading spyware via popular plugins, underscoring rising risks in browser security. TechSpot

The Register reports that Proxmox has launched its stable Datacenter Manager, providing robust software-defined datacenter management across multiple clusters and sites. Seen as a credible private cloud alternative and a migration path from VMware, Proxmox’s open source approach could reshape IT infrastructure for cost-conscious enterprises. The Register

Computer World explains that the SAMOSA bill (H.R. 5457) aims to overhaul federal software asset management, mandate automation, and crack down on shadow IT by requiring CIO approval for all software acquisitions. While analysts praise the intent, many criticize the lack of operational realism, unfunded mandates, and potential compliance chaos, especially in the rapidly evolving AI landscape. Computer World

MIT Technology Review reports on new studies showing that politically biased AI chatbots can shift voter opinions more effectively than traditional advertising. Experiments in the US, UK, Canada, and Poland demonstrated that just one conversation with a language model could meaningfully move partisan attitudes—sometimes by more than 10 points. However, increased persuasiveness correlated with more misinformation, raising alarms about AI’s impact on democratic processes. MIT Technology Review