# Morning AI TL;DR — 2026-05-12

Fresh frontier-model news was light over the last 24–48 hours, but several practical agent/developer updates are worth tracking:

- **OpenAI launched the OpenAI Deployment Company**, a new enterprise implementation venture with 19 partners to help companies build and deploy AI systems in real operations ([OpenAI](https://openai.com/index/openai-launches-the-deployment-company/), [Bain](https://www.bain.com/about/media-center/press-releases/2026/bain-company-openai-a-new-venture-to-deploy-ai-at-enterprise-scale/)); **why it matters:** frontier labs are moving deeper into services/implementation, which could reshape the SI and Salesforce partner ecosystem.
- **Claude Code added “agent view” for managing parallel coding sessions** from a single CLI dashboard ([Claude Code docs](https://code.claude.com/docs/en/agent-view), [TestingCatalog](https://www.testingcatalog.com/anthropic-adds-agent-view-for-claude-code-for-parralel-work/)); **why it matters:** coding agents are becoming orchestrated work queues rather than one-off chat sessions.
- **Anthropic recently raised Claude Code and Opus API usage limits** after a compute expansion, including doubled Claude Code five-hour limits and no Pro/Max peak-hour reductions ([Anthropic](https://www.anthropic.com/news/higher-limits-spacex)); **why it matters:** higher ceilings make long-running agentic dev workflows less brittle.
- **No major model releases were broadly tracked for May 9–11**, according to one model-release tracker ([AI Flash Report](https://aiflashreport.com/topics/new-ai-model-releases.html)); **why it matters:** this morning’s signal is more about deployment/tooling than benchmark jumps.
- **GitHub’s recent Copilot-in-VS Code update improved agent-session UX**, including incremental chat rendering, sortable agent sessions, and synced Copilot CLI session titles ([GitHub Changelog](https://github.blog/changelog/2026-05-06-github-copilot-in-visual-studio-code-april-releases/)); **why it matters:** agentic IDE workflows are converging on session management and handoff primitives.
- **Salesforce’s TDX 2026 agent stack remains strategically relevant**, with Headless 360, Agentforce Vibes 2.0, and Salesforce capabilities exposed as APIs/MCP tools/CLI commands ([Salesforce Ben](https://www.salesforceben.com/salesforce-headless-360-and-agentforce-vibes-2-0-revealed-at-tdx-2026/), [Computer Weekly](https://www.computerweekly.com/news/366641628/TDX-2026-Salesforce-depicts-Saas-as-in-agentic-evolution)); **why it matters:** Salesforce builders should design reusable capabilities as agent-callable tools, not just UI components.

**Worth watching:** Expect the next competitive edge to come from agent operations—permissions, observability, queues, and enterprise deployment playbooks—more than raw model announcements this week.