Dr. Grauer's Column - Panda Prepping for College and Life

Panda Prepping for College and Life
By Dr. Stuart Grauer

Disclaimer: This is Dr. Stuart Grauer's personal experience and is not an endorsement of a for-profit product. The article below is designed to illustrate one innovative way that Artificial Intelligence (AI) is being used to individualize secondary education.

Part I:  AI, SATs, and the Future of Learning

Like you, I’ve been inbox-slammed with AI pitches. AI is on fire for nearly everything: time management, investment (how did they know?), even parenting. And I understand we are all headed for bespoke coffee mugs and bargain psychotherapists, thanks to artificial intelligence. But when I got an ad promising to transform SAT prep, I paused. That one hit closer to home.

For simplicity, in the story, I’m using “SAT” to refer to SAT and ACT, the two main college entry tests—in fact, many Grauer School students prefer the ACT and a fair number opt out of college prep testing when they can. Maybe weirdly, for those who know me, I’m in favor of it (conditionally)… let’s look under the hood at this amazing topic.

As a teacher, the amount of time and stress poured into SAT prep has concerned me for decades. To me, SAT Prep is a time sink about as useful as learning the fax machine repair manual, but does it have to be?

So, when I got the AI-driven SAT prep pitch, I replied with a personal letter insinuating I had some influence over these matters. And before I knew it, I was on a video call with Piyush Kumar, co-founder of Vega AI (maker of LearnQ.ai) and his (human) assistant.

It was 10 PM in India and I imagined him on a 16-hour day, but Piyush looked focused and interested. Here was a man who creates artificial intelligence, talking to me. It is a little strange and otherworldly to talk to someone who is so smart, as though it’s a regular conversation, but then he probably could not make it in the world of surfing, much less in a classroom full of teens.

My main question: Can AI really find that elusive Goldilocks zone in learning? “Yes—90% of the time,” Piyush told me confidently. He also confirmed they were exploring many other personalized feedback mechanisms and purposes, which caused me to write this column, a two-part column this time. Big subject.

Vega AI’s platform builds any course in Duolingo style, inspired by that playful, bite-sized approach to learning. But Piyush says, “Duolingo is cool… but not cool enough.” A bit of swagger there.

His LearnQ AI SAT prep program starts out by assessing your performance in real-time. You take a mock test (as I did), and the system instantly identifies the next challenge at your learning edge. Each “game” adjusts: if you lose, the next one gets easier; if you win, the challenge steps up. It tracks your accuracy, time on task, and mastery—then recommends exactly what to work on next. You’re constantly nudged just at the edge of your comfort zone.

And yes, it knows if you’re faking. I was intellectually challenged by the verbal SAT assessment. But once on the math section, I could not get the calculator to work. So I just faked my way quickly through, and the algorithm called me out. The artificial brain wasn’t fooled.

They claim to have over 12 million questions and a chatbot named Mia that students will learn to love. While Duolingo has the green owl helper for students to enjoy, LearnQ students will be guided by a significantly cuter cartoon panda, which tracks and coaxes you through your personalized SAT “game world.” It “adapts to your tone, learning style, and communication preferences.”

The LearnQ program is gamified and grounded in adaptive learning theory. They don’t call it the Goldilocks zone, but that’s what it is. Their motto: “We track your detailed learning journey to guide your self-learning. See your progress, understand your growth, and know exactly where to focus next.”

The gamification aspect is striking—and here could be the essence of all I am expressing. On average, U.S. teens spend 1.5 to 2.5 hours a day on video games. Boys average more, girls a bit less—but across the board, this is a dominant daily activity. (Common Sense Media and Pew Research Center back this up.)

So the Big Question: What if SAT prep through AI could match even half the appeal of gaming? What if kids find the games as good or better? What would that do for our nation’s collective verbal and math IQ?

Every question in LearnQ.ai is ranked for difficulty and tagged based on how past users performed. The algorithm starts by diagnosing your current level. There are five tiers. Duolingo, (which I use almost daily), has a roughly comparable 5 learning “crowns.” You move up only by consistently scoring above 90%—though if you take too long, the problems get easier. This is because much of what we call intelligence is, for better or worse, time-based. You have to know the answer pretty much “now.” 

“After any game or test,” says the website, “we provide detailed insights into areas you’re struggling with and suggest your immediate next steps.”

At The Grauer School, some students hardly prepare for these tests at all, while others “obsess”—the whole gamut. Some of our senior high students are already discovering that simply using ChatGPT, they can generate practice questions tailored to any type of problem they get wrong—and it’s free. So in the end, when you buy into AI-assisted SAT prep, what you're really paying for is the identification of your "leading edge"—that level just hard enough to stay motivational. And, to me, that’s a huge efficiency. That’s accelerated learning.

Another twist, the most interesting of all: schools can license the Vega platform and customize it—not just for SATs, but for any subject. History? Biology? Logic? It’s all possible. This led me to Part 2 of this column, worth your waiting a week for! First, enjoy this backstory:

Backstory (Why I Even Tried This)

Many educators and parents among us have long been skeptical of SAT prep and I think we are all cautious about AI and its potential to be exploited in schools. Both have seemed like rabbit holes—potentially addictive, screen-bound distractions from “real,” non-artificial, experiential learning. So, when I saw a sleek ad promising “Smarter learning,” it felt like another fast-food pitch. And yet, there I was, clicking my way in deeper.

Why? For those of us who came up grinding through SAT books thicker than phone books (Kids, those are extremely thick paperbacks with extremely thin pages listing every known person with a phone number in a given region—they really had these!), the idea of something smarter—cleaner, adaptive, maybe even fun—sounded like redemption. There had to be a better way to prep for college entry exams.

This wasn’t just a curiosity—it was a personal experiment with ulterior motives because I don’t care all that much about the SAT. What I care about is optimizing learning, not wasting time, improving intelligence and related thinking skills, and using amazing technology in a healthy and non-time-wasting way. This all led me to some unexpected insights. In Part II. Where things get interesting. But first:

The Old Grind

I’ll never forget climbing Half Dome in Yosemite. Not everyone made it. I saw someone wearing flip-flops, holding their child’s hand, and others obviously right out of the climbing gym and geared-out. Each of them will tell you a different “truth” about Half Dome.

SAT prep used to be all brute force: repetition, thick manuals, guesswork. Students were told: Do more. Burn out slower. No clear muscle group, no clear plan.

Same for me a million years ago, same for my daughter in the 2000s. And same for most kids today. It became a negative experience for many—like being sent up the same trail in flip-flops or crampons, with no guidance. Just, “keep going.”

The Teaching Dilemma

Teachers know this mess. In test prep classes, some students are fluent, others flail. Throughout history, this same curriculum has always moved forward regardless. Some can hire a tutor—$90 an hour—but for most, the system has long rewarded standardization over growth.

Meanwhile, the real gold—curiosity, creativity, kindness, and the real goals that give us passion and purpose as teachers—gets buried under scantrons. Not part of the test. I will get to the heart of this in Part II of this story, next week... a critical turning point. But for now:

Learning at the Edge: 2 Approaches

There are two ways to approach learning: as a finish line to cross, or as a process to engage. Summative or formative.  Even early in my teaching, the finish-line model—what we call summative learning—never felt quite right. So in my English class, I handed out logic books and thinking games, encouraged experimentation, and let students work at their own pace. One year, I joined in, doing the drills alongside them a couple times a week, and my own, actual IQ jumped 10 points. Seriously. The takeaway was simple: when students consistently work at their edge, they grow. They enjoy it. That’s formative learning: focused on process, curiosity, and development, not just the final score.

This isn’t just teacher instinct. Psychologist Robert Bjork coined the term “desirable difficulties”—learning sticks better when it's just hard enough.

Mihaly Csikszentmihalyi’s “flow” theory says the same: if it’s too easy, you get bored; too hard, you get anxious. But that sweet spot? That’s where real learning happens. Cognitive scientists have put the number at 85% success. Not too easy, not too hard. Goldilocks.

Piyush’s LearnQ 90% threshold appears to lean slightly toward confidence-building and challenge, especially in a gamified setting. In gaming and behavior design, success rates closer to 90%–95% are often used to encourage learners. (Some UX researchers [like Andrzej Marczewski, Amy Jo Kim] argue that 90–92% success rates can optimize engagement and flow in game-based learning.)

Manu Kapur adds “productive failure”—struggle first, learn better. Your brain activates more deeply. This is mapped neurologically. It’s real. And now AI programmers are creating learning games and programs using this critical, Goldilocks knowledge, for the first time, broadly, for SAT prep and for almost any learning you can imagine, and accessible for almost all demographics.

What I Found

Here are a few of the SAT prep programs I sampled—not endorsements, just observations:

• Acely: Expensive, great advertising campaign, loaded with adaptive tech and 10,000+ questions, but you could easily spend $500.
• R.test: Precise diagnostics, but a clunky interface that did not allow me to hurry.
• MentoMind: A chatbot-heavy experience. I didn’t love it, but maybe Gen Z does. Too much chatbot reminds me of trying to rebook an airline flight.
• LearnQ.ai: My favorite by far. Smooth interface, smart analysis, and of course the panda mascot. What more could you want?

LearnQ was the only program that was simple for me to get into quickly and without committing my grandchild’s college savings. I took the 40-minute, free diagnostic test. Verbal? I crushed it. Math? Not so much—especially without a working calculator. After finishing the math section, LearnQ said that I ought “not set my sights on MIT. “ Ouch. Give me an AI to help me with my dreams!

For “Command of Evidence.” Whatever that means. It means I got all the “cross-text analysis” and “text structure and purpose” questions right. It means I have moral authority to recommend all this to you!

LearnQ.ai immediately sent me a personalized plan, a discount offer, and three emails a day since then.

Gamification seems like the breakthrough here. You pick your mood. Want a quick sprint, or a deep dive? The AI adjusts. You stay more motivated when you’re at your right level. You crank up your SAT or ACT score with an incredible lack of time wasting, staying on your learning edge.

And now, a spoiler to close this week’s column:

I still believe SATs are reductive as learning experiences. They flatten kids into numbers. But the formative process of prep—this new, personalized kind—is something different. It sharpens minds, builds clarity, strengthens reasoning. It’s motivational. This is all worth paying attention to. We now have a technology which enables you to program the facts and data surrounding virtually any growth area you have, gamify them, and then use those games to develop at efficiencies rarely available in the history of mankind.

I love using learning edges to optimize growth and stop wasting time. I love how AI promises to bring us all closer to our learning edges—quickly, personally, and efficiently. As it has integrated into much of what we do, I use optimized learning every day, in various areas of my life, often without my knowledge. Even in health.

This is why we might actually begin to like SAT prep. And that’s the spoiler, too: Next week, in Part II, we will open up to a future we used to only dream about: What if AI could help us grow not just smarter, but kinder, braver, and more connected to the world? In Part II: education’s new frontier—powered by empathy, resilience, and “RootedAI.”

COMMENT! Click on the "Comments" drop-down box below to share a comment.

SHARE! Click on the social media icons below or copy the link to share this column.