A few things happened this week regarding AI safety that should make us pay attention.

1) Anthropic dropped its flagship safety rule. Anthropic was founded by researchers who LEFT OpenAI because safety wasn't being taken seriously enough. Until this month, they had a core commitment to never train an AI system unless it could guarantee in advance that the company’s safety measures were adequate. That commitment is now gone. The new policy says they'll keep developing unless they believe they are leading the AI race and the risk of catastrophe is significant. Their reasoning? Pausing while others race ahead could make the world LESS safe. I get the logic, but 😳.

2) The Pentagon gave Anthropic an ultimatum over military AI and the deadline is today. Separately, Defense Secretary Hegseth told Anthropic to remove Claude’s military safeguards (Claude is currently the AI model used in the Pentagon’s classified systems), or face consequences. Anthropic has refused to budge on two things: autonomous weapons without a human in the loop, and bulk surveillance of American citizens. The Pentagon has threatened to blacklist them as a "supply chain risk" and even invoking the Defense Production Act to commandeer the technology. The deadline is 5pm today.

Side note: A researcher at King's College London just published a study where he ran 21 simulated war games using today's top AI models, including Claude. In 95% of the games, at least one model chose to deploy a tactical nuclear weapon - and no model ever chose to surrender.

My thoughts: I still think Anthropic is one of the more responsible companies in this space (the fact that they're pushing back on the Pentagon says something). But this really concerns me… even if Anthopic does hold their line, which I hope they do (and as of last night, they say they are), other leading AI companies are in negotiations with the Pentagon to replace them. Given this administration's aggressiveness, this is seriously unsettling.

I have to confess, February is not my favorite month. The holidays are gone, spring still feels far, and for whatever reason, it feels like my boys are moodier and restless.

I came across something recently that I think might help, and it's backed by research: awe walks. The concept is super simple: go outside for 15 minutes and try to look at the world with fresh eyes. Pay attention to things that feel vast, or beautiful, or surprising. The way light hits a building. The pattern of bare branches against the sky. The size of a cloud formation. That's it. That's the whole thing.

Turns out, this really works: compared to people who just took regular walks, the awe walkers reported more joy, more compassion, and less daily stress.

You can do this on your own, or here are a few ways to try as a family:

Give it a try… it’s free, right outside your door, and brings instant calm and joy. 🍂 ❤️

AI can be an incredibly powerful learning tool given it’s ability to personalize learning for you in just the right way at the right time. Like, bitesized daily lessons about something you’ve been wanting to learn about but keep putting off.

You can schedule tasks with ChatGPT, Gemini, but more powerfully, Claude Cowork also just announced this ability. Try a prompt like this in Cowork:

Google has quietly been adding more and more AI tools to Google Photos and one of the coolest features I’ve discovered is its AI editing abilities. You can instantly remove sunglasses, change backgrounds, open someone’s eyes who happened to be blinking, etc. right inside the photos app.

To try:

Scientists have developed a lab-grown human spinal cord that successfully healed after a simulated injury - a breakthrough that could eventually help paralyzed patients walk again.

Actual human spinal tissue - built in a lab - and deliberately injured then responded to "dancing molecules" therapy by regrowing nerve fibers and shrinking scars. The treatment, which has already enabled paralyzed mice to walk, marks the first time researchers can test spinal injury treatments on actual human tissue before clinical trials.

For people living with severe spinal injuries, this could be life-changing.