This Week’s Read: Three stories this week share an unexpected through line: humility. Sal Khan—AI’s most prominent champion in education—admits Khanmigo was a “non-event” for most students. A veteran math educator argues that AI can’t fix what’s broken in math standards because the problem was never a technology problem. And across 31 state legislatures, 134 bills are trying to write the rules before the tools write them for us. Sal Khan Admits the AI Tutoring Revolution Hasn’t Happened—Yet Chalkbeat • April 9, 2026 • Interview/Feature In 2023, Khan Academy founder Sal Khan stood on the TED stage and declared that AI was on the cusp of delivering “probably the biggest positive transformation that education has ever seen.” Three years later, in a candid interview with Chalkbeat, Khan acknowledges that the transformation hasn’t materialized. His AI-powered chatbot, Khanmigo, was a “non-event” for most students. They simply didn’t use it much. Khan’s new conclusion is striking for someone who once predicted AI would turn average students into standouts: “AI is going to help. But I think our biggest lever is really investing in the human systems.” Creative Force Meaning: When the most visible champion of AI tutoring says the biggest lever is human systems, the entire field should take notice. Read: chalkbeat.org It’s Time to Rewrite Math Standards—and Stop Expecting AI to Do It for Us eSchool News • April 9, 2026 • Opinion/Essay James Tanton—Chief Mathematics Officer at SmartWithIt, co-founder of the Global Math Project, and one of the most respected voices in math outreach—makes a sharp argument in this essay: AI isn’t going to save math education, because the problem was never a technology problem. The crisis, he contends, is one of meaning. Students race through fragmented topics to prepare for exams while the deeper “why” behind the mathematics remains permanently out of reach. Tanton’s core provocation is that AI “personalizes, but it doesn’t humanize.” He argues that edtech companies have confused efficiency with pedagogy, generating more practice problems when what students actually need is more curiosity. When we teach mathematics without the human stories that first inspired it, he writes, we strip the subject of the very spark that drove it forward. The solution isn’t smarter software—it’s instructional materials designed like great plays, with narrative structure, emotional stakes, and a sense of purpose.

GREETING OPPOURTUNITY Hello everyone, it’s great to be part of this group. I’m a professional producer and singer, and I specialize in songwriting and music production for both personal and commercial. I’m looking forward to connecting, collaborating, and contributing value here. Glad to be here!

This week’s three stories, from my reading, share a single throughline question: what happens when the tools outpace the humans who use them? Google DeepMind’s Aletheia is now producing publishable mathematics research with zero human intervention. Meanwhile, a sweeping RAND survey reveals that students are using AI for homework at record rates—while simultaneously worrying it’s hollowing out their ability to think. And a new study out of NC State and Carnegie Mellon shows that even when teachers have AI-powered dashboards tracking every student, they still gravitate toward the same kids over and over. I hope you enjoy it if you get a moment to read and reflect. DeepMind’s Aletheia: The AI That Writes Its Own Math Papers Google DeepMind (blog + arXiv) • February–March 2026 • Research Report Google DeepMind has introduced Aletheia, a mathematics research agent that represents a genuine leap beyond AI systems that merely solve competition problems. Powered by an advanced version of Gemini Deep Think, Aletheia doesn’t just answer questions—it conducts original research. The system operates through an iterative loop: a Generator proposes candidate proofs, a Verifier flags logical gaps, and a Reviser repairs them, cycling until the proof is clean. The result is an agent that achieved 95.1% accuracy on the IMO-Proof Bench Advanced, shattering the previous record of 65.7%. This isn’t a story about AI replacing mathematicians. It’s a story about the arrival of a “tireless collaborator” that never gets stuck, never gets attached to a beautiful wrong idea, and never needs a walk to cool off. For math educators, the deeper question is pedagogical: if AI can now generate and verify original proofs, what does “learning to do mathematics” mean for the next generation of students? Read: deepmind.google Students Use AI More Than Ever—and Worry It’s Eroding Their Thinking RAND Corporation • March 25, 2026 • Research Report A new RAND Corporation survey of more than 1,200 students ages 12 to 29 has surfaced one of the most striking paradoxes in education right now: students are adopting AI tools for schoolwork at an accelerating rate while simultaneously becoming more convinced that those tools are undermining their ability to think critically. The share of middle schoolers using AI for homework rose from 30% to 46% between May and December 2025. For high schoolers, it climbed from 49% to 63% over the same period.