All Updates
Page 1737 of 1959
March 5, 2026
Meta AI Glasses Leak Intimate Footage
Meta's AI smart glasses reportedly send sensitive videos to human reviewers in Kenya. Footage includes bathroom visits, sex, and intimate moments, per Swedish investigation. A class action lawsuit has emerged over privacy violations.
AI Data Center Boom Sparks Land Rush
Surging demand for AI data centers is driving prospectors to hunt for land and power supplies. Former Microsoft executive Brian Janous's firm Cloverleaf packages electricity and land deals for data centers. This reflects growing infrastructure challenges for AI expansion.
Smart Home AI Reboot with Alexa+ Gemini
Traditional smart homes have underperformed, but AI assistants promise a reboot. Amazon's Alexa+ and Google's Gemini aim to accelerate setup processes. However, fundamental problems in the ecosystem persist.
Osmo Revives AR Edutainment Push
Osmo is making another attempt to succeed in AR edutainment. The newsletter recalls a 2014 experience with a tangram puzzle system using physical wooden pieces detected by an iPad camera to match on-screen shapes, triggering animations and sounds upon completion. Full details are in the complete Verge article.
Google's February 2026 AI Updates
Google's AI Blog features a roundup of the latest AI news and updates announced throughout February 2026.
Lenovo Adopts CXMT LPCAMM2 in ThinkBook
Amid ongoing global DRAM shortages, Lenovo is incorporating LPCAMM2 memory from Chinese vendor CXMT into select ThinkBook laptop models. Orders demand one-hour decisions and prepayments to secure supply. This reflects Chinese suppliers rapidly penetrating the PC industry chain.
ChatGPT Trump Alignment Sparks Claude Surge
Backlash against ChatGPT over perceived alignment with Trump has driven soaring downloads for rival AI Claude. The controversy coincides with a feud involving the Pentagon. This highlights emerging risks in AI companies' political stances.
Lex Multi-Developer CI/CD Pipeline
This post introduces a multi-developer CI/CD pipeline for Amazon Lex, providing isolated development environments, automated testing, and streamlined deployments. It includes step-by-step setup instructions and real-world results from teams adopting the approach.
Netflix Acquires Affleck's AI Filmmaking Firm
Netflix has acquired InterPositive, the AI filmmaking company founded by Ben Affleck. Affleck stated he aims to preserve human judgment in storytelling amid AI integration. This deal positions Netflix to advance AI in content production.
Custom Provider for Strands Agents on SageMaker
This post shows how to create custom model parsers for Strands Agents using LLMs on SageMaker that lack Bedrock Messages API support. It covers deploying Llama 3.1 with SGLang via awslabs/ml-container-creator and implementing parsers for integration.
US Weather Agency Plans AI Overhaul
The US agency overseeing National Weather Service prioritizes advanced AI weather models. Plans include reorganizing into a leaner structure. Aimed at overhaul in coming months.
BYD Ultra-Fast Charge Targets NIO
BYD launched second-gen blade battery with 5min 10-70% charge at room temp, 12min 20-97% at -30°C, plus 1500kW flash stations network. This erodes NIO's battery swap advantage amid northern cold challenges. NIO responds with Shenji AI chip funding at 22.57B RMB, NX9031 boasting 4x Nvidia Orin-X compute.
Haobo A800 Packs 896-Line LiDAR
Haobo A800 launches at 18.98万 RMB with Huawei's 896-line LiDAR and ADS 4.1, matching luxury Zunjie S800 specs at half price. Features advanced chassis, efficient increaser powertrain, and HarmonyOS 5 with AI agents. Emphasizes safety with 14 airbags and battery gun-test pass.
PyTorch ExecuTorch Hits Micro-Edge on Arm
PyTorch introduces ExecuTorch for deploying models on micro-edge devices using Arm architecture. This enables AI inference on palm-sized, low-memory systems previously limited by resource constraints. It expands PyTorch's reach beyond cloud to tiny embedded hardware.
US Lawmakers Probe Device Eavesdropping Risks
Two US lawmakers urged GAO to investigate if modern computers and phones remain vulnerable to Cold War-era physical radiation eavesdropping. This revives focus on longstanding side-channel attacks. The request spotlights potential risks in current hardware.
AI Set to Worsen Africa's $10B Cybercrime Costs
Cybercrime inflicts $10 billion annual losses on Africa. HP Wolf Security's Threat Insights Report reveals attackers pivoting from precise exploits to high-speed, high-volume tactics. AI is expected to accelerate this dangerous shift.
OpenAI Develops BiDi for Natural Voice Interruptions
OpenAI is building a bidirectional voice model named BiDi to make ChatGPT conversations more human-like by handling interruptions seamlessly. Unlike current turn-based voice mode, BiDi adjusts responses in real-time without stopping. Prototype faces stability issues, with release possibly delayed to Q2 or later.
China's Plan Prioritizes Orderly AI Growth
China's 15th five-year plan stresses safe and orderly AI development amid global tech concerns. It elevates the AI Plus initiative as a top national priority. The plan received final approval at the 'two sessions'.
Hisense Launches AI World Cup Home Appliances
Hisense released a full lineup of AI-powered appliances for the 2026 World Cup, including UX2026 TVs with AI match viewing, smart ACs with human-sensing airflow, and multi-function washers with fabric recognition. Fridges feature AI ingredient detection for 800+ items, while the Big Mint 2.0 suite enables whole-home AI control. This launch emphasizes proactive AI understanding user needs in smart homes.
Anthropic Rejects Pentagon AI Demands
Anthropic refused US Defense Department's demand for unrestricted Claude access, citing bans on autonomous weapons and mass surveillance. DoD retaliated by labeling Anthropic a national security supply chain risk, barring government contracts. This clash highlights AI ethics versus military control, with Anthropic planning a lawsuit.