Bloom proved in 1984 that one-on-one tutoring beats everything else in education, by miles. We just couldn't afford it. AI finally changes the economics, but only if teachers are trusted to use it.
LAUSD's AI policy is not quite as bad as NYC's in one critical area.
New York City is calling its new AI policy a balance, but for the 600,000 students in Pre-K through 8th grade, it's actually a ban, just a smarter one than 2023's.
Hiding freshman grades treats the symptom of a student mental health crisis while ignoring what's actually causing it, and strips away the incentive to push through hard classes in the process.
From personalized instruction to removing the blank page for kids who've never had a way to show what's in their heads, this is where AI actually earns its place in the classroom, and why that matters most for the kids with the least.
Cheating, plagiarism, screen time, equity, privacy: the fears about AI in schools are real, but most of them aren't new. AI didn't break education, it just shone a spotlight on problems that were already there.
A room full of educators went quiet when I suggested assignments shouldn't have a word count. That reaction is the real subject of this post: the "well, this is how we've always done it" mindset that's slowing AI adoption in the classroom, and why it's worth pushing past, carefully.
Last week I presented a keynote at the Adobe ColdFusion Summit 2026. The team created this (admittedly somewhat over the top) intro video for me.
What's behind the spiraling cost of higher education? Administrative bloat? Athletics spending? Fancy dorms? All of those are real. But they're symptoms, not the cause.
I am sitting on a flight heading from Las Vegas, and I am feeling utterly overwhelmed.