Anthropic sues the Pentagon (yes, really)

Andreessen Horowitz

ChatGPT leads with 900M weekly users but faces accelerating competition from Gemini and Claude as the AI assistant market fragments into distinct ecosystems—consumer super-apps versus professional power tools—while standalone products increasingly bundle into platforms and agentic AI emerges as the next frontier beyond chat interfaces.

TechCrunch

Anthropic sued the DOD in two federal courts after being designated a supply-chain risk and losing all government contracts for refusing to allow its AI to be used for mass surveillance and autonomous weapons, claiming unconstitutional retaliation for its AI safety positions.

WIRED

Nvidia is launching 'NemoClaw,' an open-source AI agent platform targeting enterprise customers through partnerships with major tech companies, signaling the company's strategic pivot from LLMs to autonomous AI agents that can perform complex tasks independently.

Martin Alderson

Claims that Anthropic loses $5,000 per Claude Code user mistake retail API pricing for actual compute costs; analysis comparing OpenRouter pricing suggests real inference costs are ~10% of API prices, meaning Anthropic likely breaks even or profits on inference despite overall unprofitability from training costs.

Anthropic

Anthropic's labor market analysis finds no significant unemployment impact from AI yet, though hiring has slowed 14% for young workers (ages 22-25) entering highly exposed occupations like programming and customer service. The gap between theoretical AI capability and actual workplace adoption remains substantial, with real usage covering only 33% of theoretically automatable tasks in computer/math fields.