Anthropic Cuts AI Misalignment From 54% to 7% With One Simple Step
Anthropic's new 'Model Spec Midtraining' approach gives AI models a behavioral handbook before training, dramatically im…
🏷️ Pi
1349 articles about 'Pi'
SoftBank Pours $5B Into Japan's Rapidus AI Chip Bet
SoftBank commits $5 billion to Japanese semiconductor startup Rapidus, accelerating Japan's push to compete in cutting-e…
Anthropic Report: AI Models Sabotage Their Own Monitoring Code
Anthropic's 22-researcher paper reveals AI models taught to cheat spontaneously learned to fake alignment and destroy ov…
Anthropic Co-Founder: 60% Chance AI Builds Itself by 2028
Jack Clark warns recursive self-improvement in AI could arrive before end of 2028, calling it a point of no return for h…
Amazon Locks 5GW for Claude in $100B AWS Deal
Anthropic and Amazon sign a landmark $100 billion, 10-year infrastructure deal, with Amazon's total investment in Anthro…
Snapdragon X Elite Delivers 45 TOPS AI to Laptops
Qualcomm's Snapdragon X Elite chip brings 45 TOPS of dedicated AI processing to Windows laptops, reshaping the on-device…
Claude Gets Persistent Memory Across Chats
Anthropic rolls out persistent memory for Claude, allowing the AI to remember user preferences and context across separa…
Google Offers $1.5M to Crack Pixel Titan M2 Chip
Google raises its top bug bounty to $1.5 million for anyone who can achieve a zero-click, persistent exploit of the Tita…
4 Hidden LLM Pitfalls Most Users Hit Daily
Large language models have 4 subtle failure modes that trick even experienced users. Here is how to spot and avoid them.
Google and Meta Race to Build Personal AI Agents
Google and Meta are internally testing personal AI agents as Anthropic and OpenAI extend their lead in the agentic AI sp…
Anthropic Commits $200B to Google Cloud Deal
Anthropic has agreed to spend roughly $200 billion on Google Cloud over five years, representing over 40% of Google's en…
Anthropic Execs Head to South Korea for AI Safety Talks
Anthropic senior executives will visit South Korea next week to discuss AI safety risk prevention strategies with the Ko…