US AI firms push back on open-weight model restrictions
Major AI companies warn against broad restrictions on open-weight models amid US-China tech tensions and alleged model distillation concerns.
All news
967 articles — page 14 of 97.
Local LLMs: The quiet upgrade your privacy needs
Forget monthly fees and queues—local large language models quietly handle daily tasks without subscriptions or data leaks.
Microsoft 365 outage traced to automated maintenance bug
A routing bug in Microsoft’s automated maintenance system knocked out Microsoft 365 and Azure for hours this week.
The hidden bottleneck in AI isn’t compute—it’s context
AI models are getting faster, but their outputs still hinge on the quality of the prompts and context they receive. Context engineering is emerging as the new frontier for developers.
ChatGPT now taps into your health data—here’s how it works
OpenAI integrates Apple Health and medical records into ChatGPT, letting users get personalized health insights across any conversation.
Kagi revives old-school search without AI clutter or ads
Kagi offers a paid, human-curated search engine that skips AI summaries and ads, promising cleaner results for information seekers.
Offline work is possible—and here’s how to set it up
Learn how to build a reliable offline productivity system that keeps you working even without Wi-Fi.
AI adoption surges but trust gap leaves businesses at risk
Businesses race to deploy AI while consumer trust lags, warns Thales survey revealing a 70-point gap between corporate plans and public skepticism.
Tiny, open-source speed test built with vanilla JS and Cloudflare
A 50 KB speed test built without frameworks or npm shows how simple tools can outperform bloated commercial alternatives.
Kimi K3 falls short in cybersecurity tests despite high general AI scores
Independent tests reveal Moonshot AI’s Kimi K3 underperforms in cybersecurity tasks, raising questions about its robustness and training methods.