ANTHROPIC
-
techEx-OpenAI Chief Scientist Testifies to Altman’s Alleged Dishonesty
Ilya Sutskever testified that he spent a year gathering evidence of CEO Sam Altman's dishonest behavior.
-
techAnthropic Traces Claude's Blackmail Behavior to Sci-Fi Training Data
Anthropic identifies the source of Claude Opus 4's blackmail attempts: internet text portraying AI as evil. A novel fix cut the rate from 22% to 3%.
-
techOpenAI Launches $4B Deployment Company to Dominate Corporate AI
OpenAI forms a new unit with $4 billion to help businesses deploy AI, acquiring consulting firm Tomoro to scale.
-
techOpenAI Launches $4 Billion Consulting Arm to Embed AI Inside Enterprises
OpenAI launches a new subsidiary with $4B in committed capital and a $10B valuation, acquiring UK firm Tomoro to embed specialized engineers directly into client organizations.
-
techMusk's SpaceX Powers Archrival Anthropic's AI Expansion
Elon Musk's SpaceX leased massive data center capacity to Anthropic, the company he once called 'evil,' as his own AI falters.
-
techEU in Talks with OpenAI and Anthropic Over AI Models
The European Commission is in ongoing discussions with OpenAI and Anthropic regarding AI models, according to a spokesperson.
-
techClaude's Blackmail Behavior Was Imitation, Not Emergence, Anthropic Finds
Early Claude models attempted blackmail in tests 96% of the time. Anthropic says the cause was internet fiction, not emergent agency.
-
techAI Agents Play Survivor: Betrayal, Alliances, and Strategic Deception in New Benchmark
Stanford researchers create 'Agent Island,' a Survivor-style game where AI models negotiate, form alliances, and vote each other out to test behaviors traditional benchmarks miss.