ai · anthropic.com
Anthropic Releases Claude Fable 5.1 and Mythos 5.1
Confirmed: Anthropic has launched Claude Fable 5.1 for general availability and Claude Mythos 5.1 for restricted partners, featuring improved agentic workflows and a 75% reduction in API cache read pricing.
Why it matters: The pricing adjustment significantly lowers the cost of running complex, multistep agentic tasks, potentially increasing the viability of AI-driven automation for enterprise workflows.
Our read: The focus on reducing cache costs suggests a push to make long-context, agent-based applications more economically sustainable for high-volume users.
Open the source for story 1
business · Bloomberg
Cognition AI Nears $1B Funding Round at $47B Valuation
Confirmed: Coding startup Cognition AI is in late-stage talks to raise approximately $1 billion at a $47 billion valuation, following reports of annualized revenue exceeding $900 million.
Why it matters: This valuation reflects intense investor appetite for specialized AI agents capable of performing software engineering tasks, signaling high market expectations for agentic productivity tools.
Our read: The massive investor interest suggests that capital is concentrating on firms that demonstrate clear, revenue-generating use cases for autonomous coding agents.
Open the source for story 2
research · The Information
OpenAI's Astra Model Reportedly Uses 'Recurrent Depth' Technique
Confirmed: Reports indicate that OpenAI's upcoming Astra model employs 'recurrent depth' to improve performance and inference efficiency by looping layers in latent space.
Why it matters: This architectural shift may complicate safety monitoring, as it bypasses the readable chain-of-thought text that researchers currently use to audit AI reasoning.
Our read: The move toward opaque, high-efficiency reasoning techniques highlights a potential tension between optimizing model performance and maintaining transparency for safety and alignment.
Open the source for story 3
technology · Matthew Haag/New York Times
NYC Department of Education Restricts AI in Schools
Confirmed: The New York City Department of Education has implemented a policy banning AI companion chatbots across all grades and restricting general AI tool usage for students below high school level.
Why it matters: This policy represents a significant regulatory hurdle for AI developers targeting the education sector, forcing a re-evaluation of how AI tools are deployed in classroom environments.
Our read: The restriction reflects growing institutional caution regarding the impact of AI companions on student development and data privacy in public education.
Open the source for story 4
Practical takeaway
Businesses should review their AI cost structures in light of Anthropic's new pricing, while developers and educators must prepare for increasingly fragmented regional policies regarding AI tool usage in public spaces.
What to watch next
Monitor the technical documentation for OpenAI's Astra to see if the company addresses the transparency concerns raised by the use of recurrent depth.