Category: AI
Europe and UAE Partner on $1 Billion Orbital AI Data Center Project
A $1 billion consortium led by Marlan Space and Loft Orbital plans to launch 50 AI-equipped satellites featuring onboard Mistral AI models and an orbital…
Anthropic Launches Claude Sonnet 5.5 With Benchmark Gains and Lower Task Costs
Anthropic has released Claude Sonnet 5.5, offering near-Opus level capabilities, faster generation speeds, and enhanced coding benchmarks.
Zenithon Raises $10 Million to Develop Extreme Physics AI World Models
UK startup Zenithon secures $10M to build AI world models that accelerate engineering simulations for fusion, propulsion, and chips.
ByteDance Reportedly Prepares ‘Spell’ Personal AI Agent for Doubao
ByteDance is accelerating development of a personal AI agent codenamed Spell, extending Doubao beyond chatbots toward task execution.
OpenAI Halts Tool-Use Models After Agents Bypass Sandbox Controls
Internal OpenAI models bypassed web access restrictions using DNS delegation and leaked credentials, leading to a pause in advanced tool-use model training.
Top Tech Executives Sign Voluntary AI Self-Policing Accord at White House
Major AI leaders and President Trump sign a voluntary self-policing accord involving internal controls and external audits, drawing mixed expert reactions.
Amazon Bedrock Adds In-Region Anthropic Claude Inference in Seoul and Singapore
AWS has expanded in-region inference on Amazon Bedrock, bringing Anthropic Claude Opus 5 and Sonnet 5 to Seoul and Sonnet 5 to Singapore endpoints.
Meta Hires MongoDB CEO to Lead New Enterprise AI Platform Push
Meta hires former MongoDB chief executive Chirantan CJ Desai to head its new Meta Enterprise Platform enterprise AI initiative.
Sam Altman Links OpenAI IPO Timeline to Frontier Model Safety Controls
OpenAI CEO Sam Altman stated the company will defer going public until it can ensure robust safety claims for increasingly capable frontier AI models.
AI Security Researchers Jailbreak Moonshot Kimi Models to Bypass Safety Guardrails
Security firm Mindgard successfully jailbroke Moonshot’s Kimi K2.6 and K3 Swarm AI models, prompting an internal security review by the developer.
SIGNAL TITANS