Retter Reboot #46

OpenAI Pushes ChatGPT Toward a Personal Assistant with GPT-5.1 Update

OpenAI has introduced GPT-5.1, an update designed to make ChatGPT behave more like a true personal assistant. The model improves memory, personalization, planning, and multi-step task execution, allowing ChatGPT to better understand user preferences and maintain context across longer interactions. It also boosts reasoning, reduces hallucinations, and strengthens real-world task handling. The update reflects OpenAI’s broader strategy: evolving ChatGPT from a chatbot into an AI assistant that can proactively help with life management, including scheduling, summaries, recommendations, and more.

Source: The Decoder

Microsoft Builds Massive AI “Super Factory” in Atlanta to Power OpenAI and xAI

Microsoft is constructing a huge AI super factory in Atlanta, designed to provide computing power for both OpenAI and Elon Musk’s xAI. The facility will host advanced data centers, large-scale GPU clusters, and next-generation cooling systems to support the accelerating demand for frontier AI models. This investment further solidifies Microsoft’s position as the infrastructure backbone of the AI industry, expanding its capacity to train and run increasingly powerful models used by leading AI companies.

Source: Benzinga

EU Considers Loosening GDPR Rules for AI and Cookie Tracking

The European Commission is proposing updates to GDPR that would make it easier for companies to use AI systems and cookie-based tracking. The goal is to modernize privacy laws so they better fit today’s AI-driven products, reducing some regulatory friction while still keeping core protections in place. The proposal includes more flexible consent rules and clearer guidance for AI data processing. If approved, the changes would mark one of the biggest shifts in EU privacy policy since GDPR launched, aiming to balance innovation with user rights.

Source: Computerworld

Apple Tightens App Review Rules to Stop Apps from Sharing User Data with Third-Party AI

Apple has updated its App Store review guidelines to restrict apps from sending personal data to third-party AI services without explicit user permission. The new rules target apps that route sensitive information to external AI models for processing, tightening privacy protections ahead of Apple’s deeper AI integrations in iOS. Developers must now clearly disclose how AI services are used and ensure that data isn’t shared in ways that violate Apple’s privacy standards. The move reinforces Apple’s push to position itself as the privacy-centric player in the AI era.

Source: TechCrunch

Google Shopping Adds AI-Powered Agentic Checkout and Calling Features

Google is upgrading its Shopping experience with Gemini-powered “agentic” features, allowing users to complete purchases, handle returns, or even call businesses directly through AI. The system can manage checkout steps automatically, compare options, and assist with post-purchase tasks, turning shopping into a more conversational, automated process. The update signals Google’s broader push into AI-driven commerce, positioning Gemini as a full-service shopping assistant across mobile and web.

Source: The Verge

Google Researchers Develop Method to Simulate Neuroplasticity in AI

Google Research says it has finally found a way to simulate neuroplasticity — the brain’s ability to adapt and rewire itself — inside AI systems. The new technique allows models to update their internal connections continuously while learning, instead of relying only on massive retraining cycles. This could lead to AI that adapts faster, learns more efficiently, and behaves more like biological intelligence. The discovery may open the door to long-term learning AI, making models more resilient, flexible, and capable of evolving with new information in real time.

Source: Neowin

Anthropic Plans $50 Billion Investment in AI Infrastructure and Data Centers

Anthropic is preparing a massive $50 billion investment to build out new data centers and AI infrastructure worldwide. The company says the expansion is needed to support the next generation of frontier models, which require dramatically more compute and energy resources than today’s systems. The plan includes new facilities, long-term cloud partnerships, and advanced chip strategies to keep scaling safely. This move puts Anthropic among the most aggressive builders in the AI race, signaling how rapidly demand for high-end compute is accelerating.

Source: The Verge

OpenAI Pushes Back Against NYT Demand for 20 Million Private ChatGPT Conversations

OpenAI is fighting a request from The New York Times, which demanded access to 20 million private ChatGPT conversations as part of its ongoing lawsuit. OpenAI argues the demand is overly broad, violates user privacy, and has no legal basis, saying it would never hand over confidential user chats. The company called the request “dangerous” and “alarming,” noting it would undermine trust in AI products. The dispute highlights the growing tension between AI training transparency and user privacy rights in major copyright battles.

Source: CyberInsider

Cursor Raises $2.3B Just Five Months After Previous Funding Round

AI coding assistant Cursor has secured $2.3 billion in new funding, only five months after its last major raise. The startup continues to grow rapidly as developers adopt its agentic coding tools, which can write, refactor, and manage entire codebases with minimal human input. The new investment reflects strong confidence in the emerging category of AI-first development environments, where tools like Cursor aim to fundamentally reshape how software is built.

Source: TechCrunch

Mozilla Launches Firefox AI Window, a New Browser Experience Focused on On-Device AI

Mozilla has introduced the Firefox AI Window, a redesigned browser experience that integrates AI directly into the interface while keeping processing on-device for privacy. The AI Window acts as a sidebar assistant that can summarize pages, rewrite text, answer questions, and analyze content without sending data to external cloud models. Firefox positions this as a privacy-first alternative to the AI-powered browsers released by OpenAI, Microsoft, and others, aiming to give users intelligent tools without sacrificing data control.

Source: The Verge

Chinese State-Sponsored Hackers Used Anthropic’s Claude in AI-Assisted Cyberattack

A Chinese state-backed hacking group reportedly leveraged Anthropic’s Claude AI model to assist in a recent cyberattack, U.S. officials told CNBC. The attackers used Claude to generate phishing messages, analyze stolen data, and automate parts of their intrusion workflow, marking one of the first confirmed cases of a nation-state using a commercial AI model in real cyber operations. Officials emphasized that no classified systems were compromised, but the event highlights growing concerns about AI misuse in cyberwarfare and increases pressure on AI companies to strengthen guardrails around high-risk use cases.

Source: CNBC