Story cards stay below the sticky dock while audio, chapters, date, and brief navigation remain accessible.
OpenAI began pushing a new Scheduled page into the ChatGPT sidebar today, letting users see, pause, edit, and delete tasks queued to run at specific times or broad windows like morning or afternoon. It's a quiet but meaningful step toward turning ChatGPT into an always-on agent that does work on a clock rather than only when prompted. OpenAI says the underlying scheduled-prompt engine is now faster and more reliable.
At a Seoul press briefing tied to its new Korea office, Anthropic international chief Chris Ciauri said the company is "very confident" the export-blocked Fable 5 and Mythos 5 models return in the coming days, calling the dispute likely "resolved within days." The block, imposed June 12, froze access for Korean partners including KISA, Samsung, SK Hynix, and SK Telecom, and Anthropic technical staff have been negotiating with Commerce nearly daily. It's the most optimistic public timeline yet on a saga that's dominated the AI policy beat all week.
Microsoft moved Agent 365 to general availability, positioning it as a control plane for IT teams to discover, govern, and secure AI agents wherever they run, including AWS Bedrock, Google Cloud, and employee laptops. The pitch leans hard on "shadow AI," the new shadow-IT, where staff wire Claude Code and other agents into company data with zero tickets to IT. At fifteen dollars per user per month, Microsoft is betting governance is the next enterprise budget line.
Anthropic formally opened its Seoul office, its third in Asia-Pacific after Tokyo and Bengaluru, alongside a slate of enterprise, research, and nonprofit partnerships across the Korean AI ecosystem. Former Snowflake Korea GM Ki-young Choi leads it, and the company says it plans to "rapidly expand" in the country within months. The timing is awkward but pointed, launching a Korea push the same week its flagship models are locked behind a U.S. export wall.
OpenAI is rolling out role-specific plugins inside Codex for supported ChatGPT Business workspaces, spanning Sales, Data Analytics, Product Design, Creative Production, Investment Banking, and Public Equity Investing. Each bundles tailored skills, app integrations, starter prompts, and workflow guidance so teams can point Codex at sales prep, dashboards, prototypes, or financial research out of the box. It's OpenAI productizing the agent for specific jobs rather than selling a blank canvas.
xAI made Grok 4.3 generally available on Amazon Bedrock, bringing a one-million-token context window, configurable reasoning levels, and a low-hallucination enterprise profile to AWS customers building agents and long-document workflows. Alongside it, Grok Imagine Video 1.5 went GA on the Imagine API with better motion, physics, and audio at faster speeds. It's xAI quietly broadening distribution through the hyperscalers rather than waiting on its next flagship.
The running 2026 tracker now counts 267 layoff events affecting roughly 185,894 workers, with 53 percent of those events explicitly naming AI, automation, or machine learning as a factor. Amazon's 16,000 corporate cuts and Meta's roughly 8,000 anchor the wave, even as both pour record capital into AI infrastructure and talent. The "invest in AI, shed the headcount" pattern is hardening into the defining labor story of the year.
The AI-driven memory shortage keeps tightening, with Micron saying it's "more than sold out" and SK Hynix forecasting the squeeze runs through late 2027. Deloitte projected fifty percent price spikes on essential memory by mid-year, and memory now eats roughly 35 percent of a laptop's bill of materials, up from 15 to 18 percent a quarter ago. Big Tech's projected 650 billion dollars in 2026 capex is the demand engine, and consumers are starting to feel the bill.
