Natalia Pereldik

Love mathematics,chocolate and adventures. Co-founder of Funexpected Math.

[Aakash Gupta @aakashgupta]

A 15-year-old girl immigrates to New Jersey from China. Doesn’t speak English. Her parents, both educated engineers back in Chengdu, are now working as cashiers and restaurant cooks. She gets a job washing dishes at a Chinese restaurant to help the family survive.

She gets into Princeton on a full scholarship. Her reaction is so disbelieving she asks two different advisors to verify the acceptance letter is real. Then her mom gets sick, so the family opens a dry cleaning shop in Parsippany. Every weekend for seven years, Fei-Fei Li leaves Princeton’s physics department to run the register, handle inspections, talk to customers, manage billing. Monday through Friday: quantum mechanics problem sets. Saturday and Sunday: sorting other people’s laundry. She later called herself the “CEO” of the dry cleaning business. She kept running it remotely through half of her PhD at Caltech.

In 2007, she proposed building an image dataset so massive her own mentor told her she’d taken the idea “way too far.” Pre-ImageNet, the entire AI field was working with datasets containing a few hundred images. She built one with 15 million. Most researchers at the time believed algorithms were the bottleneck. She bet on data when nobody else would.

By 2012, a team ran a neural network on that dataset and halved the existing error rate overnight. AlexNet on ImageNet became the moment the deep learning era started. Every computer vision product shipping today traces its lineage back to that dataset.

Fast forward to 2024. She starts World Labs. Four months in, $230 million raise, $1 billion valuation. Today, $1 billion more at roughly $5 billion.

The bet investors are making: that the woman who gave AI its eyes with 2D image recognition is about to give it spatial awareness of the 3D physical world. Her new model, Marble, generates persistent 3D environments from text or images. Unlike video generators that fake depth frame by frame, Marble creates actual geometric space where objects stay where you left them.

The investor list tells you everything. AMD and NVIDIA both wrote checks. When the two biggest competing chipmakers both fund the same startup, they’re telling you this workload is coming whether their competitor funds it or not. Autodesk put in $200 million and signed on as strategic advisor, which means they see spatial AI integrating directly into CAD and design workflows within 18 months.

From dry cleaner to ImageNet to a $5 billion spatial intelligence company. Fei-Fei Li has now placed two bets that the rest of the field thought were too early and too big. The first one created modern computer vision. The second one is trying to give machines the ability to understand physics.

If she’s right again, this is the last major unlock before embodied AI actually works.

[Natalia Pereldik @perenat00]

@MariaPosicel @HOUmanitarian @houstonlibrary @AliefISD @sarahgish @Intl_District @ThomasDTiffany One of my kids’ favourites events in Funexpected Math!

[Natalia Pereldik @perenat00]

@MariaPosicel @sciencegirl I love this math task!

[Natalia Pereldik @perenat00]

@MariaPosicel That’s wonderful, I should try it out with kids

[Natalia Pereldik @perenat00]

@MariaPosicel Yeah, they also had a Nobel prize in math in their script, so I guess they do not take much time to check themselves 😃

[Natalia Pereldik @perenat00]

Lack of basic math understanding is striking. I just watched a short vertical film in which the character said the door has a six-digit code, which means there are 900 million possible variations.

[Natalia Pereldik @perenat00]

Competitive math teaches perseverance, working on a problem in steps, and working on a problem long enough. It also often helps to meet like-minded people and to make friends for life. As well as generating ideas and checking your own hypothesis. That’s a great foundation for any future career.

[Natalia Pereldik @perenat00]

@juliapintar Manage

[ℏεsam @Hesamation]

investing enough time to learn math in the earliest age possible will give you insane ROI. first step of learning AI isn’t reading “attention is all you need.

it’s MATH.

[Natalia Pereldik @perenat00]

One of my favourite sets of questions in the app. It is a great prompt for various discussions with kids

[Maria Positselskaya @MariaPosicel]

It’s crucial to talk to children about the predictive power of science.

In our problems, we ask kids to predict which chair will fall, or which teapot will pour water.

At school, I once had an exam on electricity with just one task: put on rubber gloves, stand on a rubber mat, hold two nails, and stick them into a socket. You were alone. You had to decide for yourself. Rubber is an insulator—but at the same time, parents teach us never to put nails into sockets.

We study science to be confident that 2+2 is always 4, the Earth is spherical, and rubber does not conduct electricity.

[Anna Stokke @rastokke]

🎅The best Christmas gift ever. Free access to the Coursera course by Barb Oakley and John Mighton, Making Math Click: Understand Math Without Fear. Shared this in my latest episode.
🎁 https://t.co/cj3sxsuOiu

[Natalia Pereldik @perenat00]

This is our Christmas quest postcard this year in Funexpected Math. Wish you all the best!

[こにゃん。 @pilot_P623]

やったぁ！さんすうパズルをすべてといたよ！ネコちゃんの クリスマスじゅんびも、てつだったよ！🐱🎄^_^（ネネさんが）
うつくしいさんすうのせかいをひらいてみよう！
https://t.co/3mwSqVn8z0

[Natalia Pereldik @perenat00]

@funexpectedmath That’s so cool that families are sharing

[Natalia Pereldik @perenat00]

We know that in our app, kids sometimes get stuck when our tutor Aika asks a question. One of the simplest UX ideas we want to try is this: after a moment of hesitation, Aika gently says, “Let’s just pick an answer and see what happens.”

Learning starts with trying something, not with being sure.

[Natalia Pereldik @perenat00]

A common difficulty for kids when solving problems is simply getting started. If they aren’t sure which method to use, they can freeze. School often makes this problem even stronger: in class, kids always know what topic they’re studying and which method the teacher expects them to apply.

But in real life, you often have to try different approaches before you find one that works. And that first step — trying something — is one of the hardest things to learn.

For example, you can start by plugging in a few numbers just to see what happens.

[Natalia Pereldik @perenat00]

My favorite video so far!

[Funexpected Math @funexpectedmath]

One of our best videos yet. We count how many candles one needs for all days of Hanukkah using an addition strategy. Happy Hanukkah!
https://t.co/82uQk4d9tS

[Natalia Pereldik @perenat00]

Let’s count how many candles one needs for all the days of Hanukkah! #k... https://t.co/0tfgFylks3 via @YouTube

[Natalia Pereldik @perenat00]

Just finished watching this interview.

“I think I am a scientist at heart. I always thought I was going to become this math or CS professor and work on trying to understand a universe and language and the nature of communication. Like it’s kind of funny, but I always had this fanciful dream where if aliens ever came to visit Earth and we need to figure out how to communicate with them, I wanted to be the one to go with and I’d use all this fancy math and computer science and linguistics to decipher it.”

What an ambitious dream to have as a teen. What a goal!

My son is 12. He’s a great kid — curious, really good at math — but for now, his main driver is social. I wonder what kind of life experiences create such a complex motivational goal so early.

[Lenny Rachitsky @lennysan]

The fastest company in history to $1B did it with no VC money and fewer than 100 people.

@HelloSurgeAI has become the secret weapon behind Anthropic and Google's best models. Founder Edwin Chen (@echen) built it without playing the Silicon Valley game—no viral posts, no fundraising treadmill.

In my powerful conversation with Edwin Chen, we discuss:
🔸 How Claude got so good at coding and writing
🔸 Why the AI industry is optimizing for "dopamine instead of truth"—and delaying AGI
🔸 The problems with AI benchmarks and leaderboards
🔸 Why RL environments are the next frontier in AI training
🔸 Why taste and human judgment determine which AI models win
🔸 Why Edwin believes we’re still a decade away from AGI

Listen now 👇
• YouTube: https://t.co/RkA8ONNsKE
• Spotify: https://t.co/gluDhTNssd
• Apple: https://t.co/lSM0XIxCM2

Thank you to our wonderful sponsors for supporting the podcast:
🏆 @TrustVanta — Automate compliance. Simplify security: https://t.co/JHcQhNsK42
🏆 @WorkOS — Modern identity platform for B2B SaaS, free up to 1 million MAUs: https://t.co/XH3bKmeEkH
🏆 @coda_hq — The all-in-one collaborative workspace: https://t.co/sP4vEcSaUY

[Natalia Pereldik @perenat00]

So many kids do well in early elementary math, but then suddenly lose confidence when they move into middle-school math. They spend years building fluency in number sense and learning different strategies for addition and subtraction. And then, very abruptly, they’re expected to switch to abstract thinking: working with variables, solving geometry problems that require mental visualization, using the coordinate plane in algebra.

Early-grade skills don’t really prepare them for this jump.

That’s why in our program we intentionally include games on coordinates, spatial-thinking challenges, tasks with maps and projections, and much more — even for our youngest learners. We want them to build these foundations early, so the transition doesn’t feel so overwhelming later.

[Maria Positselskaya @MariaPosicel]

Before kids meet graphs in algebra or physics, it helps to let them feel graphs with their whole body. I always dreamed of a simple device that plots your position over time. Imagine a child running with it and watching the graph of their own motion appear live. You could even run competitions: kids get a target graph and try to run so that their own motion matches it as closely as possible.

We built a different game instead: the position of a point on the screen depends on your right hand on the vertical axis and your left hand on the horizontal axis. Kids have to coordinate their hands to steer a little ship toward a treasure — a full-body introduction to functions and coordinate planes.

https://t.co/XzsO2g7JwW

[Natalia Pereldik @perenat00]

And still teaching math now often doesn’t use even half of these strategies

[Justin Skycak @justinskycak]

Numerous effective learning strategies have been identified and researched extensively since the early to mid-1900s, with key findings being successfully reproduced over and over again.

At a glance, here are some of the highlights: