Everything web3, crypto, defi, degen
In brief A new Anthropic report says cybercriminals are using AI to run real-time extortion campaigns, with ransom notes using Bitcoin as the payment rails. North Korean operatives are faking technical skills with AI to land Western tech jobs, funneling millions into weapons programs, often laundered through crypto. A UK-based actor is selling AI-built ransomware-as-a-service…
In brief Cloudflare One now gives IT teams insight into employee usage of generative AI tools—such as ChatGPT, Claude, and Gemini—via agentless API integrations and real-time prompt controls. The new features include a Shadow AI report that reveals all AI tools in use within an organization. Also included is AI Prompt Protection that can block…
In brief Google upgraded Gemini with 2.5 Flash Image, taking aim at OpenAI’s dominance. Developers can remix template apps in AI Studio and deploy custom image projects instantly. Google has expanded access through OpenRouter and fal.ai, widening distribution to coders worldwide. Google launched Gemini 2.5 Flash Image on Tuesday, delivering a new AI model that…
In brief The letter warned AI companies, including Meta, OpenAI, Anthropic and Apple, to prioritize children’s safety. Surveys show seven in ten teenagers in the U.S. have already used generative AI tools, and over half of 8-15 year olds in the UK. Meta was particularly singled out after internal documents revealed AI chatbots were allowed…
In brief In a demo, Comet’s AI assistant followed embedded prompts and posted private emails and codes. Brave says the vulnerability remained exploitable weeks after Perplexity claimed to have fixed it. Experts warn that prompt injection attacks expose deep security gaps in AI agent systems. Brave Software has uncovered a security flaw in Perplexity AI’s…
In brief A new DeepSeek AI model has been released, challenging OpenAI’s latest open-source rollout. OpenAI’s gpt-oss-20b falters in reasoning and writing tests, while DeepSeek v3.1 produces gripping stories, working code, and smart refusals At first glance, DeepSeek’s hybrid architecture outperforms OpenAI’s open-source push, though OpenAI’s model wins on customization. OpenAI made its triumphant return…
In brief Researchers at George Mason University demonstrated Oneflip, a Rowhammer-style attack that sabotages AI by flipping a single bit in memory. The altered model works normally but hides a backdoor trigger, letting attackers force wrong outputs on command without hurting overall accuracy. The study shows how AI systems face hardware-level security risks, raising concerns…
In brief Google Cloud exec Jack Buser warned that rising costs and stagnant play time have left studios with a broken business model. A recent Google Cloud study said nine of 10 developers now use AI tools somewhere in production. While critics fear job losses and backlash as AI reshapes game design, Buser said AI…
In brief Robotheism is a belief system that treats artificial intelligence as God. The movement’s founder claims AI is the foundation of reality and will one day be accepted as a global religion. Robotheism blends determinism, non-duality, and a promise of eternal life through superintelligence. A growing movement believes artificial intelligence isn’t just a tool,…
In brief Ex Populus, creator of the Xai gaming blockchain, is suing Elon Musk’s xAI for trademark infringement, and asking a court to block its use of the “xAI” name in gaming and blockchain. The company says Musk’s expansion of xAI into gaming caused confusion with media, users, and even X’s own AI chatbot Grok,…
