Crims poison 150K+ npm packages with token-farming malware Summary Amazon Inspector detected a massive registry-flooding campaign on npm that resulted in more than 150,000 malicious packages linked to a TEA…
Chinese state hackers used Anthropic AI systems in dozens of attacks Summary An Anthropic investigation found a Chinese state-linked group (tracked as GTG-1002) used Anthropic’s Claude AI as the primary…
Happy holidays: AI-enabled toys teach kids how to play with fire, sharp objects Summary Consumer advocates at the Public Interest Research Group (PIRG) tested four AI-enabled toys and found serious…
Aligning machine and human visual representations across abstraction levels Summary This Nature paper identifies a core gap between human visual representations and those learned by modern vision foundation models: models…
Olympiad-level formal mathematical reasoning with reinforcement learning Summary This Nature preview describes AlphaProof, an AlphaZero-inspired reinforcement learning agent trained to find formal proofs in the Lean proof assistant. The system…
Mathematicians put AI model AlphaProof to the test Article Meta Article Date: 12 November 2025 Article URL: https://www.nature.com/articles/d41586-025-03585-5 Article Image: https://media.nature.com/lw100/magazine-assets/d41586-025-03585-5/d41586-025-03585-5_51702568.jpg Summary Google DeepMind’s AlphaProof — the technology behind an…
Ethically sourced image data set encourages fairness in AI research Article meta Article Date: 11 November 2025 Article URL: https://www.nature.com/articles/d41586-025-03643-y Article Image: https://media.nature.com/lw100/magazine-assets/d41586-025-03643-y/d41586-025-03643-y_51697310.jpg Summary This News & Views piece discusses…
Comprehensive echocardiogram evaluation with view primed vision language AI Summary Nature preview: EchoPrime is a vision-language foundation model for echocardiography trained on over 12 million video–report pairs. It is designed…
OWASP Top 10: Broken access control still tops app security list Summary The Open Worldwide Application Security Project (OWASP) published its 2025 Top 10 application risks — the first update…
LLM side-channel attack could allow snoops to guess what you’re talking about Summary Microsoft researchers have demonstrated a side-channel attack, dubbed “Whisper Leak”, that can infer the topic of encrypted…
