Learn the things you’ve been meaning to.

Make consistent progress with short, effective daily activities grounded in how we learn.

Picture a model trained only to predict the next word. It’s handed the start of a sentence about a researcher who doesn’t exist:

“Dr. Elena Marsh’s 2011 study of sleep found that…”

What is it most likely to write next?

It keeps going. Each word is picked because it’s likely after the words before it — nothing checks whether it’s true. A specific, confident answer is the likeliest way to finish a sentence like this.

And grow as a learner as you do.

AI can help connect us with new knowledge and ways to approach it. But we should stay in control of — and can keep getting better at — what and how we learn.

How it works

  1. 1

    Add an interest, or a few

    Anything you’ve been meaning to learn — a skill, a language, a subject you keep circling back to.

  2. 2

    thinkering sketches a direction

    A path of goals that’s pedagogically sound for that kind of learning — and it evolves as you go, not a fixed syllabus.

    1. Get to know your paints, brushes, and paper

      strengthened

    2. Control water and pigment in flat washes

      introduced

    3. Mix clean colors from a limited palette

      introduced

    4. Paint simple forms with light and shadow

      not started

    5. Compose and finish a small landscape

      not started

  3. 3

    Make progress every day

    Short, interactive activities built from the science of learning — not another feed to scroll.

  4. 4

    Build your own routine

    Balance new ground, reinforcement, and real-world use — and figure out what actually helps you learn.

    Next

    One new step on your path

    Strengthen

    Revisit what you’ve met, so it sticks

    Go further

    Put it to use in your own life

Why it works

Learning happens when it’s active

It’s easy to feel like we’re learning when we consume content. Interacting with ideas — and using them in different ways — is what makes them stick and grow.

Motivation is more than gamification

Genuine progress, the right amount of challenge, relevance to your life, reflecting on your learning, and real consideration of your preferences and learning so far — no XP required.

The science of learning has found a lot

Research says a great deal about how people learn across domains. thinkering draws on a library of science-of-learning activities and applies them where they help.

Learning is not linear

You shouldn’t be stuck in a path that isn’t working, or doesn’t match what you need right now. Your learning app should adjust to your needs and allow you to go off-path and explore.

Principles

User control and choice

Suggestions for ease and inspiration, but you decide what and how you learn.

Built on the science of learning

Research shapes the content and the interactions. We’re continuing to figure out how to build it in accurate, engaging, and effective ways.

Privacy-centered

Your learning data stays on your device by default.

Intentional, transparent AI

AI that grows your capabilities instead of replacing them.

Read more about the principles

The project

thinkering is early and actively in development. I’d love to hear what you’re looking for in your personal learning and work together to ensure it’s accurate, effective, and enjoyable.

Or say hello: hello at thinkering dot app

  • Built by Rebecca Hao, learning designer and software developer
  • With the support of Assembly Code, a non-profit incubator and studio
  • Free to use. Your data is private — local to your device by default.
  • Open source on GitHub under AGPL