Messy breakup? Pentagon and Anthropic

Good morning. While everyone else scrolls mindlessly, we’ll scroll with purpose. 💪

What’s going on? The Pentagon and Anthropic are breaking up. In a formal statement, Anthropic CEO Dario Amodei said the company cannot agree to let the Department of Defense use its AI models for “all lawful purposes” without limits. As a result, President Trump has directed every federal agency to ”immediately cease all use of Anthropic’s technology.”

Want the whole story? Anthropic, the AI startup behind the Claude chatbot, has a $200 million contract with the Pentagon and already provides AI tools for some military workflows. But the company wants a written guarantee that its technology will not be used for fully autonomous weapons or for mass surveillance of Americans. The Defense Department says it has no intention of using the models that way and notes that mass domestic surveillance would be illegal. Still, it wants the freedom to use Anthropic’s AI for any lawful purpose, including in war. Defense Secretary Pete Hegseth has reportedly warned that Anthropic could lose its contract or face other penalties if it refuses to agree. In response to all of this, President Trump has decided to outlaw Anthropic’s tools at the federal level, stating that Anthropic’s “Terms of Service … puts our National Security in jeopardy.”

Why does this matter? This dispute highlights a growing tension between AI companies and the government over how powerful AI systems should be used. As AI becomes more integrated into military and national security work, companies are tentative to give up control over how their tools should be used. The outcome of this situation could shape how much control tech firms have over their tools once they enter government hands, and how AI is used for national security and defense going forward.

What now? It looks like the Department of War is working out a deal with OpenAI to replace Anthropic’s role and deploy their own models on the classified network, so long as human responsibility is assigned within the Department and not to OpenAI. CEO Sam Altman posted some details on X last night. Check it out →

What’s new? Claude’s Cowork feature now lets paid users create scheduled tasks that run automatically on a recurring basis.

How does it work? With scheduled tasks, you write the instructions once and Claude follows them on the schedule you choose — hourly, daily, weekly, weekdays, or only when you trigger it. These tasks can pull information from tools you’ve connected, like Slack, Google Drive, email, calendars, and spreadsheets. For example, you can set up daily briefings that summarize messages, weekly reports that compile data into clean summaries, recurring research that tracks competitors, or automatic file organization. Each scheduled task runs in its own session, and you can review, pause, edit, or delete tasks anytime. One important detail: tasks only run when your computer is awake and the Claude Desktop app is open. If your device is asleep, the task will run once you reopen the app.

Why should you care? This turns Claude into a lightweight automation assistant. Instead of repeating the same reporting or research work every week, you delegate it once and let it handle the routine — saving time and reducing busywork.

Hi — I’m Hunter, a PhD candidate whose work has appeared in major academic journals and popular tech outlets. I founded FryAI to make staying ahead of AI clear, accessible, and fun.