Anthropic Accuses 3 Chinese Companies of Harvesting Its Data
Anthropic reveals DeepSeek, Moonshot, and MiniMax used 24,000 fake accounts for 16 million Claude interactions to distill and improve their models, amid US debates on AI chip exports.
Monitor
Archived Intelligence
American AI Industry Trembles as DeepSeek Prepares V4 Release
DeepSeek's upcoming V4 model threatens US labs with superior efficiency, amid distillation accusations.
VC loyalty dead in AI: Dozen OpenAI backers now fund Anthropic
Investors jumping ships freely between top AI labs signals cutthroat competition and blurred allegiances.
What's behind the OpenClaw ban wave from AI providers
Anthropic and Google are banning flat-rate accounts connected to OpenClaw over API overuse and policy violations.
AI News Roundup: Nvidia-OpenAI deal shrinks, ByteDance video model alarms Hollywood
Latest roundup covers Nvidia and OpenAI scaling back investment, India's AI summit, and concerns over ByteDance's new video-generation model impacting Hollywood.
Pentagon summons Anthropic CEO Dario Amodei amid dispute over military AI limits
Defense Secretary Pete Hegseth is meeting Anthropic's leader over restrictions on using Claude for weapons, risking contracts and labeling it a supply chain risk.
What to know about job market and AI job fears
Amid strong job market, fears grow over AI displacing workers; experts weigh current realities and future impacts.
Forget Nvidia: This stock for AI boom's next phase
Analysts highlight optical circuit plays like Coherent (COHR) as next big AI infrastructure bet over Nvidia.
76% of finance firms plan 2026 AI investments, but only 6% advanced
New report reveals automation gap as most plan agentic AI spending but few have implemented sophisticated systems.
Uber launches Autonomous Solutions division for robotaxi survival
Uber creates new AV unit to integrate autonomy across ride-hailing and delivery amid intensifying competition.
IBM shares tank 11% as Anthropic's Claude masters COBOL modernization
Anthropic unveils Claude's ability to streamline legacy COBOL code, flipping economics of modernization and slamming IBM's consulting business; worst day for IBM since 2000.
Anthropic accuses Chinese labs of distilling Claude with 16M+ queries via fake accounts
DeepSeek, MiniMax, Moonshot allegedly scraped Claude outputs en masse to train rivals; Google warns of similar attacks on Gemini.
Google Cloud AI lead outlines three frontiers of model capability: intelligence, speed, cost
In interview, Google exec Michael Gerstenhaber details push toward smarter, faster, cheaper models for real-world apps like code gen and moderation, emphasizing infrastructure gaps holding back agentic AI.
The toll and opportunity of AI data centers
AI data centers guzzle electricity and water but spur economic growth, sparking debate in Colorado and beyond. Episode explores environmental costs vs. job creation from AI infrastructure boom.
2026, the Year of AI Victims
Litigation surges against AI firms post-hype, with lawsuits over harms mounting rapidly.
Guide Labs open-sources interpretable AI model Steerling-8B
Guide Labs releases Steerling-8B, a new open-source LLM designed for interpretability, aiming to make AI decision-making more transparent.
Department of War prioritizes AI in new defense agenda
Secretary Hegseth outlines AI-first strategy for unified innovation in military tech.
OpenAI partners with top consultants Accenture, Capgemini for enterprise AI push
OpenAI enlists major consulting firms to accelerate enterprise adoption and customization of its models in 2026.
Dow plunges 800 points on AI disruption and tariff fears
Markets tumbled as investors worry about AI's impact on jobs/software and Trump's tariff plans, pressuring tech.
Anthropic uncovers industrial-scale distillation attacks on Claude by DeepSeek, Moonshot, and MiniMax
Anthropic revealed Chinese AI labs created 24k+ fake accounts for 16M+ interactions to distill Claude's capabilities into their models, violating ToS. Attacks targeted reasoning/coding gains, with rapid pivots to new Claude releases. Raises alarms on safeguards stripping for military/surveillance uses.
Particle's AI news app listens to podcasts for interesting clips
New app from ex-Twitter engineers uses AI to extract key podcast moments, expanding news aggregation beyond text to audio.
Future AGI vs Arize AI: Battle for best LLM evaluation tool in 2026
In-depth comparison of leading platforms for evaluating LLMs, crucial for teams building reliable AI systems.
Malicious OpenClaw skills distribute Atomic macOS stealer malware
Trend Micro warns of bad actors using OpenClaw's skill marketplace to spread info-stealing malware at scale.
Intel launches OpenVINO 2026 with enhanced LLM and NPU support
Updated toolkit boosts large language model optimization and Core Ultra NPU performance for edge AI inference.