🤖 OpenAI ships Presence to power enterprise AI agents 🤖
OpenAI rolls out its most ambitious enterprise agent platform to date.

OpenAI has launched Presence, a new enterprise product designed for deploying controlled AI agents across customer support and internal operations. The platform bundles model reasoning with permissions, policies, evaluations, escalation rules, and post-deployment improvement tools into one production-ready package. Notably, Presence already powers OpenAI's own phone support line, serving as a live proof of concept. This is a direct challenge to enterprise agent platforms from Salesforce, Microsoft, and others — OpenAI competes head-on for enterprise agent infrastructure revenue. The release signals a clear strategic shift toward monetizable, production-grade agent tooling beyond raw API access, making it one of OpenAI's most significant enterprise launches this year.

💰 AMD invests $5B in Anthropic for MI450 chips 💰
Anthropic diversifies its compute supply with a massive AMD integration deal.

AMD and Anthropic have entered a strategic partnership worth tens of billions of dollars in AI servers. Anthropic will purchase up to 2 gigawatts of AMD's next-generation Instinct MI450 chips starting in the first half of next year. In return, AMD will invest up to $5 billion in Anthropic, contingent on deployment milestones. The companies will also collaborate on data center locations, and AMD plans to use Claude to improve its chip design processes. This deal reshapes AI chip supply chains — Anthropic reduces its exclusive Nvidia dependence with one of the largest AI infrastructure commitments ever announced.

🎥 Claude Cowork learns tasks from screen recordings 🎥
A new capability gives non-technical users direct access to custom AI automation.

Anthropic has introduced a new Claude Cowork capability called "Record a skill," which lets users record their screen while completing a task and narrate each step aloud. Claude then automatically converts the demonstration into a reusable skill it can perform autonomously — no prompts or code required. This approach mirrors how humans learn through observation, lowering the barrier to workflow automation for non-technical users. It opens Claude's agent capabilities to a far wider adoption base. With four independent reports this cycle, Claude Cowork's screen recording feature is one of Anthropic's most widely covered updates, positioning Claude as a uniquely intuitive agent platform for everyday productivity.

INTERESTING TO KNOW

⚡ Cursor Router cuts AI coding costs by 60% ⚡

Cursor has rolled out Cursor Router, an intelligent routing system that automatically selects the optimal AI model for each coding task — delivering frontier-quality results at 60% lower cost compared to routing everything through a single top-tier model like Opus 4.8. Pricing drops without quality loss signal real competitive pressure in AI-assisted development. Early access customers reported no quality drop-off and lower cost per commit. The feature is now available across all surfaces on Teams and Enterprise plans, with admin controls for per-team routing and model management.

🏛️ Altman briefs White House on GPT-6 ahead of launch 🏛️

OpenAI CEO Sam Altman is scheduled to brief the Trump administration and members of Congress next week on the upcoming GPT-6 model family, covering new capabilities, workforce impact, and AI safety — a rollout now requiring high-level political engagement. The meetings coincide with U.S. officials finalizing a framework for reviewing frontier AI models, suggesting GPT-6 may be approaching its public release window. Frontier models now require government-level review before launch.