https://humphreytheodore.com/writing/ai-last-48h-28-may-2026 Humphrey Theodore K. Ng'ambi en-GB 2026-05-28T15:27:50.702Z Twelve AI Stories from the Last 48 Hours AI, AI news, weekly roundup, OpenAI, Anthropic, xAI, MiniMax, Tencent, 2026 elections, AI safety, Claude Security, frontier AI, AI policy, AI governance https://humphreytheodore.com/writing/anthropic-coding-agents-social-sciences-adoption-gap Humphrey Theodore K. Ng'ambi en-GB 2026-05-28T16:12:59.203Z Only 20% of Social Scientists Use AI Coding Agents AI, Anthropic, AI adoption, social science, AI research, Claude, AI access, AI economics, AI in academia, education https://humphreytheodore.com/writing/anthropic-opens-korea-kiyoung-choi-apac-enterprise-stack Humphrey Theodore K. Ng'ambi en-GB 2026-05-28T16:12:44.768Z Anthropic Opens Korea with the Enterprise Stack Already There AI, Anthropic, Korea, APAC, Seoul, enterprise AI, KiYoung Choi, Claude, AI business, global AI https://humphreytheodore.com/writing/anthropic-talent-magnet-karpathy-boyd-2026 Humphrey Theodore K. Ng'ambi en-GB 2026-05-28T16:12:46.747Z Anthropic Hired Karpathy. The Talent Flow Is the Story. AI, Anthropic, Andrej Karpathy, Eric Boyd, AI talent, OpenAI, frontier AI, product-market fit, Simon Willison, AI business https://humphreytheodore.com/writing/claude-security-public-beta-anthropic-opus-47-defender-stack Humphrey Theodore K. Ng'ambi en-GB 2026-05-28T16:12:42.959Z Claude Security Ships to Enterprise Public Beta on Opus 4.7 AI, AI security, Anthropic, Claude, Claude Security, Opus 4.7, cybersecurity, defender stack, AI policy, Mythos, Glasswing https://humphreytheodore.com/writing/fujitsu-signs-openai-anthropic-dual-vendor-japan-enterprise Humphrey Theodore K. Ng'ambi en-GB 2026-05-28T16:12:48.545Z Fujitsu Signed Both OpenAI and Anthropic on the Same Day AI, Fujitsu, OpenAI, Anthropic, Japan, enterprise AI, multi-vendor AI, critical infrastructure, AI business, APAC https://humphreytheodore.com/writing/lab-split-existential-risk-anthropic-vatican-altman-walkback Humphrey Theodore K. Ng'ambi en-GB 2026-05-28T16:12:41.304Z The Lab Split on Existential Risk Is Now Public AI, AI safety, AI policy, Anthropic, OpenAI, Chris Olah, Sam Altman, Pope Leo XIV, existential risk, AI personhood, AI ethics, Vatican https://humphreytheodore.com/writing/minimax-m3-sparse-attention-china-frontier-architecture Humphrey Theodore K. Ng'ambi en-GB 2026-05-28T16:12:55.536Z MiniMax M3 Previews a New Sparse-Attention Architecture AI, MiniMax, M3, sparse attention, GQA, China AI, frontier AI, AI architecture, long context, AI research https://humphreytheodore.com/writing/openai-2026-elections-us-brazil-ai-safeguards Humphrey Theodore K. Ng'ambi en-GB 2026-05-28T09:06:04.098Z OpenAI Puts AI Inside the 2026 US and Brazil Elections AI policy, AI governance, OpenAI, 2026 elections, election integrity, Associated Press, deepfakes, C2PA, SynthID, ChatGPT, Brazil, United States https://humphreytheodore.com/writing/openai-ai-math-breakthrough-plays-to-strengths Humphrey Theodore K. Ng'ambi en-GB 2026-05-28T14:21:33.473Z OpenAI's AI Math Breakthrough Plays to Its Strengths ai, artificial intelligence, openai, mathematics, understanding ai, kai williams, erdos, frontier ai, ai reasoning, emergent intelligence https://humphreytheodore.com/writing/openai-codex-windows-sandbox-engineering-deep-dive Humphrey Theodore K. Ng'ambi en-GB 2026-05-28T16:12:50.429Z OpenAI Built Codex's Windows Sandbox the Hard Way AI, OpenAI, Codex, Windows, sandbox, security, AppContainer, Mandatory Integrity Control, AI engineering, software engineering https://humphreytheodore.com/writing/openai-self-improving-tax-agents-thrive-codex Humphrey Theodore K. Ng'ambi en-GB 2026-05-28T16:12:52.230Z OpenAI Built Self-Improving Tax Agents With Thrive AI, OpenAI, Codex, tax, Thrive Holdings, agent self-improvement, enterprise AI, AI evaluation, AI engineering, continuous learning https://humphreytheodore.com/writing/tencent-hunyuan-world-models-hy-world-2-mirror Humphrey Theodore K. Ng'ambi en-GB 2026-05-28T16:12:57.350Z Tencent Hunyuan Pushes Two World Models in a Day AI, Tencent, Hunyuan, world models, 3D reconstruction, ICML 2026, China AI, multi-modal AI, AI research, generative AI https://humphreytheodore.com/writing/xai-grok-kilo-code-ide-agentic-coding Humphrey Theodore K. Ng'ambi en-GB 2026-05-28T16:12:53.877Z xAI Puts Grok Inside Kilo Code AI, xAI, Grok, Kilo Code, agentic coding, VS Code, JetBrains, IDE, developer tools, MCP