💭 Claude's agents now dream to self-improve 💭
Anthropic's Managed Agents analyze past sessions to self-improve, adding outcomes and multiagent orchestration.

Anthropic rolled out major new capabilities for Claude Managed Agents, expanding access to production-grade autonomous agent deployments. The headline feature is "dreaming," where agents analyze past sessions between runs to identify patterns and self-improve over time — a meaningful step toward persistent agent learning. Outcomes tracking lets agents self-correct based on predefined success criteria, creating a built-in feedback loop, while multiagent orchestration enables complex task delegation to specialized subagents. Companies like Harvey, Netflix, Spiral by Every, and Wisedocs are already leveraging these features. Early enterprise adoption signals growing demand for agent infrastructure that learns autonomously.

⚖️ Murati testifies Altman lied about AI safety protocols ⚖️
OpenAI's former CTO says under oath that Altman bypassed deployment safety reviews.

Mira Murati, OpenAI's former CTO, testified under oath during the Musk v. Altman trial that CEO Sam Altman lied to her about safety standards for a new AI model. Murati said Altman falsely claimed OpenAI's legal department had determined a model did not need to go through the company's deployment safety board. She confirmed a "misalignment" between Altman's statements and those of general counsel Jason Kwon, and made sure the model went through the board anyway. Murati also agreed with descriptions of Altman as undermining executives and pitting them against each other. The testimony adds internal credibility to ongoing concerns about OpenAI's governance and safety oversight at a critical regulatory moment.

⚡ Anthropic taps SpaceX Colossus for 220K GPUs ⚡
Anthropic secures massive compute access via SpaceX, doubling Claude Code limits immediately.

SpaceX announced a landmark deal granting Anthropic access to its Colossus 1 data center in Memphis, Tennessee, delivering over 300 megawatts of power and more than 220,000 NVIDIA GPUs. Anthropic is expected to begin using the compute within the month. As an immediate benefit, the company will double Claude Code's 5-hour usage limits for Pro, Max, and Team users while increasing API rate limits. With annualized revenue reportedly surpassing $30 billion, this deal reshapes Anthropic's capacity for competitive scaling. The partnership also carries notable irony — Elon Musk's SpaceX now powers a direct rival to his own xAI. Musk stated SpaceX will rent compute to "AI companies taking the right steps for humanity."

INTERESTING TO KNOW

💻 Codex overtakes Claude Code with GPT-5.5 💻

OpenAI's Codex has reportedly overtaken Anthropic's Claude Code following the integration of GPT-5.5 and significant app performance improvements. Industry practitioners highlight practical new use cases: Austin Tedesco uses Codex to create strategy documents from diverse sources, while Dan Shipper applies it for recruiting by analyzing career trajectories. The shift signals intensifying competitive pressure in the developer tooling market, where Claude Code had previously held the lead.

💰 DeepSeek eyes $50B from China's state fund 💰

DeepSeek is in talks with China's National Artificial Intelligence Industry Investment Fund — a government-backed fund with approximately $8.8 billion in capital — to raise a few billion dollars at a $45–50 billion valuation. The pricing would place DeepSeek among the most valuable AI startups globally and cement its role as a cornerstone of China's domestic AI strategy. This funding round signals intensifying state-backed competitive pressure against U.S. AI leaders amid deepening geopolitical rivalry over AI leadership.