Hey neighbor,
Five months in — and this week's stories are some of the most surprising we've covered. AI that sees through your AirPods. A drug breakthrough from Claude. And a revenue number that will make your jaw drop. Let's get into it.
THIS WEEK IN AI
Apple is building AirPods with cameras so AI can see what you see
Apple is rumored to be developing camera-equipped AirPods that would give AI the ability to see what you see — turning your earbuds into an always-available AI assistant that can observe your environment and respond to what's happening around you.
In plain English: imagine asking your earbuds "what's that plant in my neighbor's yard?" and having AI look through the camera, identify it, and tell you — hands free, eyes free, without taking out your phone. Or asking "is this a good price for this item?" while shopping and having AI check prices instantly.
This is AI moving from your pocket to your ears — literally becoming part of how you perceive and interact with the world around you. It's not available yet, but the direction is clear: AI is getting closer and more personal every month.
Claude just designed proteins that could lead to new medicines
Anthropic published lab-validated results showing Claude designed protein binders against 14 of 15 targets tested, hitting a 22-35% success rate versus the typical 10-15% industry rate — more than double the standard success rate for drug discovery.
Here's what that means in plain English. Proteins are the building blocks of life — and designing new proteins is one of the most important challenges in medicine. New proteins can become new drugs, new treatments, and new cures. The typical success rate for designing proteins that bind to specific targets is 10-15%. Claude hit 22-35% — more than double.
This is AI being used for something that could genuinely save lives. The cancer-detecting AI from a few months ago. The protein-designing AI this week. These are the stories that remind me why AI matters beyond productivity apps.
Anthropic's revenue just hit $65 billion annualized — up from $10 billion at the start of the year
Anthropic's annualized revenue surged to $65 billion — up from roughly $10 billion at the start of 2026. That's a 6.5x increase in revenue in less than a year.
To put that in perspective: most companies take decades to reach $65 billion in annual revenue. Anthropic did it in about four years from founding. The company that makes Claude — the AI you can use for free at claude.ai — is growing at a pace that has almost no historical precedent.
What this means for you: Anthropic is not going anywhere. The free tier of Claude is going to keep improving. And the company is now large enough to take on OpenAI and Google in a serious way.
PLAIN ENGLISH EXPLAINER
What is a "protein binder" — and why does it matter that AI can design them?
Your body is made of proteins. Thousands of different kinds, each with a specific shape and job. Some proteins cause disease when they malfunction. Others can fight disease if you design them correctly.
A protein binder is a molecule designed to attach to a specific protein — like a key fitting a lock. If you can design a protein that binds to a disease-causing protein, you might be able to block the disease, slow it down, or help the body fight it.
This is the foundation of many modern medicines — antibodies, targeted cancer therapies, and more are all essentially protein binders. Designing them is hard, expensive, and slow. Most attempts fail.
Claude's 22-35% success rate — compared to the industry standard of 10-15% — means AI is now finding the right keys for biological locks at more than double the normal rate. That translates directly into faster, cheaper drug discovery. Which translates into new treatments reaching patients sooner.
TOOL OF THE WEEK
Try this: Google Gemini 3.7 Flash — a powerful free upgrade you might have missed
Google launched Gemini 3.7 Flash this week, advancing its Flash series with improved performance in software engineering, knowledge work, and web development. The model delivers higher code accuracy, better document comprehension, and new automation benchmarks — at half the price of the previous version.
If you use Google's AI tools, you just got a meaningful upgrade. Here are three things to try:
→ Upload a document — a lease, a report, a long article — and ask Gemini to summarize the key points and flag anything you should pay attention to
→ Ask it to help you understand something complex you've been putting off — a tax situation, a medical decision, a home repair question
→ Try it for planning — a trip, a home project, a major purchase — and see how it handles multi-step thinking
Go to gemini.google.com — free, no download required.
NEIGHBOR'S THOUGHT
Five months in — and I keep thinking about the protein story
Twenty issues. Five months. We've covered governments shutting down AI, billion-dollar lawsuits, chatbots hitting a billion users, and robots playing table tennis.
But the story I keep coming back to this week is the protein one.
Claude designed proteins that work at more than double the industry standard rate. That's not a benchmark. That's not a press release. Those are lab-validated results from real scientists at real labs who tested real molecules.
Somewhere in those 14 successful protein designs might be the beginning of a drug that helps someone. Maybe someone with a disease that's been resistant to treatment. Maybe someone's parent or child.
AI finding molecules that fight disease is the reason I think this technology matters beyond productivity. Beyond the email assistants and the image generators and the chatbots. This is AI being used to do something genuinely good in the world.
That's worth paying attention to. And that's why this newsletter exists.
UNTIL NEXT WEEK
That's your week in AI — Issue #20. Five months of plain-English AI news, every Thursday.
If this newsletter has helped you feel more informed and less overwhelmed, share it with one person who'd appreciate the same. That's how we grow.
See you next Thursday. ☀️
— The AI Neighbor Team
theaineighbor.com