Can Flexible AI Data Centers Save the Power Grid Billions?
AI data centers may be able to shift electricity use during grid stress. Here is what Reuters found about potential savings and the hurdles to scale.
AI data centers may be able to shift electricity use during grid stress. Here is what Reuters found about potential savings and the hurdles to scale.
Australia is developing frontier-AI safety standards that would require developers to show safeguards work. What would proof of safety mean?
Microsoft and Nvidia are expected to introduce a Surface Laptop Ultra for local AI-agent work. Reuters reported the plan ahead of an October 7 event; performance, price and availability were not yet confirmed.
Anthropic is expanding access to its most cyber-capable AI models for vetted security teams. Project Glasswing partners reported at least 129,000 verified vulnerabilities, and the revamped Cyber Verification Program adds three access tiers.
Hark Pro can use your computer and connected services. We examine its privacy claims, the access an AI assistant needs, and questions users should ask.
Kevin Roose says The AGI Chronicles is human-written, even though AI helped him organize research, check facts and get feedback. Here is the distinction he drew in WIRED.
Reflection AI has launched Beam, a 501B-parameter open-weight model for coding and AI agents. Here is how its claims stack up against DeepSeek, Kimi, GLM and Qwen.
Deutsche Telekom projects roughly €2.5 billion in AI-related indirect-cost savings by 2030. We examine its network and customer-service examples, the targets behind the headline, and what remains to be proven.
AI agents can read email and act across connected accounts. Here is how prompt injection can turn ordinary content into a security risk, and what users can do.
An OpenAI safety employee says the company’s culture needs to change. Here is what David Robinson alleged, how OpenAI responded, and what to watch next.