And.... this December @ThePocketLab Voyager will join that list!! Stay tuned for some crazy science in micro-gravity!!!

Schools Are Pushing Back on AI. Los Angeles has tightened screen-time rules. Portland paused expansion of generative AI programs. Katy ISF bars generative AI chat tools for K-6. Chicago put a contemplated Gemini rollout on hold. And NYC is reportedly preparing to prohibit student-facing generative AI through 8th grade. Schools are still using AI for teachers, administration, translation, and planning. The pushback is increasingly about putting open-ended AI directly in front of younger students. And the evidence is supporting a broad rollout is still thin: Stanford reviewed 800+ papers on AI in K-12 education and found only 20 high-quality casual studies, none examining student AI use in US K-12 classrooms. My view: AI is a tool like a calculator that students eventually need to use. But first they need reading, writing, math, reasoning, and the ability to think without AI.

Remarkable data here. In reality, even flat student enrollment strains budgets as costs increase. But for most of the cities at the bottom of the list they have little choice other than to close school buildings.

Congratulations Christine!! @CCgirt

New York science scores are improving. But 47% proficiency is still 47%. There is genuinely good news in New York's latest science results: Grade 5: 35% → 45% → 46% Grade 8: 34% → 44% → 49% Overall: 35% → 44% → 47% That's meaningful progress in just two years. But fewer than half of students reaching proficiency should also make us uncomfortable. Here's the part I keep coming back to: Science education doesn't have an ideas problem. It has an implementation problem. Pull a handful of experienced science teachers out of random classrooms and ask them what would improve student outcomes. I suspect you'd hear remarkably similar answers: More hands-on investigation More inquiry and experimentation More learning around real phenomena More measuring, collecting data, analyzing and explaining Less passive instruction and test preparation We largely know what better science instruction looks like. The harder question is: How do we make that experience the normal science classroom — rather than the exception? Progress matters. But we shouldn't normalize 47%. Source: New York State Education Department, 2026 preliminary assessment results.

Whew... I was really beginning to worry 🤪

Finally, someone at the NY DOE took a few minutes to click through the Realbotix site to see the full product line...

Setting up at NOAA! @thepocketlab

And... will be receiving a @ThePocketLab Voyager sensor in a couple months!! Stay tuned... @voyagertech_

10s of thousands of teacher clicks in our AI newsletter. Here's what they're actually interested in. A year ago we started a weekly AI-in-education newsletter. No big strategy, just the most interesting AI news for teachers, every week. 52 issues later, 8,000+ teachers open it, and I can see exactly what they click. Not what they say they care about. What they actually click. Something changed this spring. For most of the first year, the clicks followed the hype. New tutor, new model, new tool. Everyone was figuring out what this stuff could do. Then around May, the top stories flipped. A professor who thinks his whole class cheated with AI. The "dopamine dilemma" of instant answers and what it does to real thinking. And right beside those, hands-on science, kids doing the experiment instead of watching a screen do it. Those two themes went from a fifth of the clicks to about half. Teachers aren't anti-AI. They still click the tool stories. But the question is shifting from "what can this do?" to "what is this costing my students?"

AI regulation in education is moving quickly at the state level. There are now 134 AI-in-education bills are being tracked across 31 states. Major themes are becoming clear: student data privacy, classroom-use boundaries, and AI literacy. Source: MultiState, AI in Education Legislation: 2026 State Policy Trends.

Mahalo @BigIslandNow for the shout out. More than 12,000 teachers from around the world registered and eager to learn about the world class science happening on your beautiful islands! 🏝️

Can't wait to talk to Leon and so many amazing scientists at Daniel Inouye Regional Center in Honolulu!! #ScIC16 @VictorVescovo

Excited to have Hawaiian-born @GuyKawasaki keynote our live stream from Hilo. Join the more than 11,000 teachers already signed up (it's free!) to learn about the amazing science happening in Hawaii, all the way from volcanos, wave energy, astronomy, marine debris, and even a session on Wayfinding using just a star compass!

Language is the interface to intelligence. This chart should worry all of us. Not because AI is causing the decline in reading. Causation isn't the point. The reading slide started years before ChatGPT existed. The alarming part is the timing: reading skills are eroding at the exact moment they're becoming more valuable, not less. Here's why: LLMs essentially mirror the clarity of your thinking. Vague, scattered prompts yield generic or wrong answers. Clear, specific, well-structured ones produce far better results. You've surely seen this yourself. So as AI spreads into work, education, healthcare, and everyday decisions, the ability to read carefully, write clearly, and organize your thoughts matters more than ever. And a growing share of students are leaning on AI precisely when the underlying skills it rewards are weakening. The students who struggle here aren't just falling behind in school. They're at a fundamental disadvantage for the rest of their lives.

More than 9,000 teachers signed up already!