Introducing Claude Duet

This link leads to one of the most interesting things I’ve seen in AI recently.

What’s cookin’? Anthropic has introduced Claude Duet, a new AI agent designed to proactively handle work across your connected apps instead of waiting for you to ask. It can scan your email, calendar, Slack, accounting tools, CRM, and other services, then surface what actually needs your attention. You can also hand it bigger jobs like preparing forecasts, scheduling meetings, drafting follow-ups, building dashboards, or managing parts of a business. Duet keeps working even when your computer is off, but it asks for approval before taking sensitive actions like sending messages, making payments, or signing documents. The research preview is starting with U.S. Pro and Max subscribers.

🤔 Hunter’s take: The interesting part isn’t that Claude can complete tasks. We’ve seen that before. It’s that the AI is starting to decide when you need help before you even ask. That changes the relationship completely. We’re moving from “open an AI tool” to having an AI coworker quietly watching the work pile up and tapping you on the shoulder when something matters.

What’s cookin’? Websites and government services around the world were temporarily shut down after researchers uncovered two serious security flaws in PHP, the programming language that powers a large part of the web. The bugs could potentially let hackers remotely break into servers, access files or databases, and install malicious software. Cybersecurity agencies warned organizations that couldn’t install the emergency fix quickly enough to take affected systems offline instead. That led to shutdowns involving government websites in Finland and the Netherlands, parts of Wikipedia, and other online services. The vulnerabilities had reportedly gone unnoticed for more than a decade before researchers using advanced AI tools found them.

🤔 Hunter’s take: This is kind of nuts. We’re talking about governments literally taking websites offline because AI found a hole in the internet that had been sitting there for years. This goes to show that AI isn’t just making cybersecurity teams a little more efficient. It’s finding stuff humans have completely missed, on the world’s most critical infrastructure. Great if you’re the one finding it first. A lot less great if attackers do.

Here’s your chance to get a straightforward answer, from an expert in the field.

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.