Google is folding in former moonshot robotics software company Intrinsic as Alphabet tries to streamline its ambitions in the space.After five and a half years developing its technology at Alphabet's "moonshot factory" X, Intrinsic became an independent Alphabet company in its "Other Bets" division in July 2021. The Google parent company announced Wednesday that it will be folded into Google itself.While hardware like robotic arms has become cheaper, programming them remains incredibly complex, often requir ...

Core Insights - Anthropic and Atlassian Williams F1 Team have announced a multi-year partnership, designating Claude as the team's Official Thinking Partner [1] - Claude will be integrated throughout the Williams organization to enhance race strategy, car development, and operations, especially during significant regulatory changes in Formula 1 [2] - The partnership aims to leverage Claude's advanced AI capabilities to improve decision-making and performance under pressure [3] Company Overview - Atlassian Williams F1 Team is a historic Formula 1 team, founded in 1977, with a record of nine Constructors' Championships and seven Drivers' Championships, making it one of the most successful teams in F1 history [7] - The team is known for its ingenuity and innovation, focusing on maximizing performance through meticulous attention to detail [4] Partnership Details - Claude's branding will be featured on the FW48 cars, drivers, and team kit, starting with the 2026 livery reveal on February 3 [5] - The partnership is expected to demonstrate the value of AI in high-stakes environments, combining human talent with advanced AI models [6][9] - Team Principal James Vowles expressed enthusiasm for the partnership, highlighting the potential for success through collaboration with Anthropic [6]

Model Performance & Benchmarks - Kimmy K2 Thinking outperforms GPT-5 on the "Humanity's Last Exam" benchmark with a score of 44.9% compared to GPT-5's 41.7% [1] - In agentic search for Browse Comp, Kimmy K2 Thinking scores 60.2% versus 54.9% for GPT-5 [1][2] - Kimmy K2 Thinking achieves 83.1% on Live Codebench v6, a competitive programming benchmark [1] - The model can execute 200 to 300 sequential tool calls without human interference [1][2] - Kimmy K2 Thinking significantly outperforms the human baseline of 29.2% on browse comp with a score of 60.2% [2] Model Architecture & Training - The base Kimmy K2 model used 2.8 million H800 hours with 14.8 trillion tokens, costing approximately $5.6 to $6 million [3] - Kimmy K2 Thinking has a trillion parameters with 384 experts, while 32 billion parameters are active during inference [5][6] - Kimmy K2 Thinking has a vocabulary size of 160,000 [5] Market & Industry Impact - China is emerging as a key player in open-source, open-weights frontier AI models [9][10] - The cost of training frontier models is decreasing rapidly [3][4] Use Cases & Capabilities - Kimmy K2 Thinking can solve PhD-level mathematics problems using 23 tool calls in its chain of thought [1] - The model can create component-heavy websites and math explainer visualizations from single prompts [1] - Kimmy K2 Thinking can analyze the relationship between population density and healthcare facility accessibility, generating interactive maps and charts [11][12][13][14][15]

Government AI Expansion - xAI expands its services to the United States Federal Government [1] - All federal agencies and departments gain access to Frontier AI models (Grok 4, Grok 4 Fast) [1] Pricing and Commitment - The cost is $0.42 per department for 18 months [1] - xAI commits a team of Grok Engineers to assist the government in utilizing AI [2] Team Growth - xAI is expanding its team and hiring engineers [2]