Latest in AI
Trucked Gas Faces a Hard Limit in the AI Datacenter Race
Trucked natural gas may help power emerging AI datacenters, but SemiAnalysis says fuel deliveries, local supply and logistics could limit the largest campuses before they scale.
AI Companies Are Preparing for the Political Fallout of a Catastrophic Failure
Anthropic, OpenAI and other AI companies are privately planning for the political backlash, regulation and public distrust that could follow a catastrophic AI-linked failure affecting essential services.
AI firms are preparing for the political fallout of a catastrophic event
Anthropic, OpenAI and other AI firms are privately planning for the political and commercial fallout of a catastrophe, raising questions about accountability, regulation and industry power.
A 5.9 GB AI model could make private assistants feel ordinary
PrismML’s 5.9 GB Bonsai 2 27B brings compressed AI closer to laptops and phones, promising private offline assistants while independent tests question its benchmark performance, compatibility and reliability.
Can an AI Learn to Raise Its Hand Before It Goes Rogue?
OpenAI researchers are testing whether AI agents can report their own scheming before acting. The approach sharply reduces undetected attacks in synthetic trials, but raises difficult questions about honesty, incentives and deception.
Claude Opus 5.5 Trades Old AI Tells for New Ones
Claude Opus 5.5 may have reduced familiar AI writing tells, but Graphite’s research suggests new fingerprints are emerging in its vocabulary, punctuation and communication style.
Google’s Gemini Agent Wants to Become the Office’s Long-Term Operator
Google is building a persistent Gemini workplace agent designed to work across applications for days. The system could coordinate other agents, but raises urgent questions about memory, permissions, accountability, cost and enterprise control.
PolicyLM Targets Content Moderation’s Cost and Control Problem
Musubi’s open-weight PolicyLM-1.7B aims to make content moderation faster, cheaper and easier to update by translating written rules into constrained decisions, though real-world reliability remains unproven.
Google’s offline meeting assistant tests whether local AI can beat the cloud
Google’s AI Edge Foresight tests whether a local Mac meeting assistant can rival cloud tools by keeping transcripts, notes and document-grounded answers offline and private.