🧠 Health integrations, identity signals, and the rise of physical AI

🔥🔥🔥 Three big updates

1. Claude can now securely connect to your health data

Claude just rolled out new health-data integrations in beta, which is a big shift from “chatbot that answers questions” to “assistant that understands your context.”

The new connectors include Apple Health (iOS), Health Connect (Android), HealthEx, and Function Health. And importantly, these are private by design, meaning you have to explicitly opt in, and Anthropic says your health info won’t be used for training.

Once connected, Claude can summarize your medical history, explain test results in plain language, and even spot patterns across fitness metrics. The practical value here is simple: less time decoding your data, more time actually acting on it.

Rollout note: it’s currently in beta for Pro and Max users in the US.

2. ChatGPT can predict your age

CNBC covered a new angle on ChatGPT that’s equal parts fascinating and unsettling: age prediction.

On paper, this sounds like a fun party trick. In reality, it’s a reminder that identity is easier to infer than most people think, especially when AI systems pick up on patterns in how we talk, what we ask, and what we assume.

Features like this sit right at the intersection of personalization and privacy. If AI can guess your age accurately, it can also tailor outputs, recommendations, or even risk profiles, whether you asked for that or not.

3. Dragonfly” turns raw video into real-time operations intelligence

Anisoptera AI just launched Dragonfly, a no-code “physical AI” platform that converts raw video footage into real-time business operations intelligence.

This is a genuinely interesting direction because most ops teams already have the raw material: cameras everywhere: warehouses, stores, loading docks, production floors. But the footage usually just sits there until something goes wrong.

The bigger implication is that operations intelligence may become less about dashboards you build manually, and more about systems that observe the real world and surface what matters automatically.

🔥🔥 Two Tools Worth Trying

1. 🧩 Prompt Genie (the “GitHub for prompts” idea)

This Instagram clip nails a problem most AI teams quietly suffer from: prompt chaos.

We have tools for docs, version control for code, and testing for software. But prompts are basically the backbone of many AI workflows, and most teams still manage them like random notes floating around in Slack and Notion.

The pitch here is simple: collaborate on prompts, version-control them, test them, and ship them into workflows via an API. Best for teams scaling AI use internally, especially non-technical teams that still need structure.

2. 🎤 InteractPitch.ai

InteractPitch is built to help you create and deliver pitches more effectively, with AI support. If you do sales, partnerships, podcast guest pitches, speaking applications, or even internal stakeholder presentations, this kind of tool helps you tighten the message fast without sounding robotic.

Basically: fewer “what should I say?” moments, more clean, confident delivery.

🔥🔥 Things You Didn’t Know You Can Do With AI

Turn Claude into a high-signal “thinking partner” (without writing code)

This workflow is built around a Claude prompt you can reuse whenever you want better output, better structure, and fewer mediocre first drafts.

Here’s what it helps you achieve: it pushes Claude to slow down, think deeper, plan before answering, and focus on elegant solutions instead of quick ones. Great when you want clarity, strategy, or a real upgrade in quality.