This is a short summary and review of the book Ultralearning, by Scott H. Young.
The book offers a series of tips on "ultralearning" — an intense, efficient, and self-directed method of learning.
The author's most famous project was completing the four-year MIT Computer Science curriculum in just twelve months, studying entirely on his own, without ever enrolling at the university. After that, he learned four languages in one year by living in different countries under a strict rule: no English allowed.
If you feel like doing something similar, the book's tips will be quite useful, even if some of them seem vague or even obvious. Either way, I think it's worth reading — especially if you enjoy this summary, since the book includes countless other examples and tips not covered here.
One caveat: the book was published in 2019 — there's already a follow-up, which I haven't read yet — and it isn't updated to make good use of apps and artificial intelligence. So rather than just writing a review, I'll also share a series of tips for applying the book's principles with the tools available in 2026.
This article was written with some help from AI (especially the last section), and by an author who read the book three times, summarized it with pen and paper, read the summary out loud, tried to put some examples into practice, and is hoping he hasn't become obsolete yet!
Summary
Ultralearning is an intense, self-directed study strategy that requires real effort. Using it is a choice — for some skills, you may prefer a gentler approach — but if you decide to try it, it serves to learn quickly, reduce time and costs, and advance in a skill, hobby, or career (your current one, or a new one!).
The book organizes everything into nine principles.
Principle 1: Metalearning. Understand how the subject works before diving in. Map out your project: why, what, and how you're going to learn. Seek out experts who can point you in the right direction, in person if possible, emphasize what you need to learn, and cut out the rest. Dedicate about ten percent of your project timeline to planning, and occasionally revisit your learning approach.
Principle 2: Focus. The first obstacle is procrastination. Identify it, recognize that it doesn't last long, and commit to acting for five minutes — or use the Pomodoro technique, or study until the next win (keep trying until you get it right). When distracted, alternate between different activities and methods to avoid boredom. Also avoid low-quality focus — when you seem to be studying but aren't absorbing anything. Its main causes are fatigue, stress, trying to study for too long without breaks, or simply being in the wrong mental state for that activity at that moment.
Principle 3: Direct Practice. Learn in the context where you'll use the knowledge, or as close to it as possible. Direct practice improves transfer — the ability to apply in real life what you've learned. Connect learning to its context of use; use projects, immersion, simulations, or a more challenging approach, like putting yourself in a difficult situation to force progress.
Principle 4: Drill. Identify the hardest points that block your performance and isolate them — train specifically on those, even if it feels artificial and uncomfortable. It's the equivalent of the musician who repeats the five hardest seconds of a piece rather than playing it from start to finish. Once you've strengthened your weak points, return to practicing the whole.
Principle 5: Retrieval. Testing yourself through active recall — remembering material through mental effort alone, without notes — is more efficient than rereading. Tests, quizzes, and practice exams also work. Even without checking your answers, the exercise helps (but see Principle 6). Retrieval is especially useful a few days after learning the material. Difficulty here is desirable: the more effort you put in, the more fixed the knowledge becomes. Common methods include flashcards, free recall, writing what you've learned as questions to answer later, and practicing with the book closed.
Principle 6: Feedback. To know whether you're improving, seek feedback that guides and corrects — not just praise or criticism. Having a mentor or coach, if possible, is even better. The more specific and informative the feedback, the better. Try to get it quickly. Watch your ego: don't get discouraged by criticism or overexcited by praise — what matters most is knowing where to improve.
Principle 7: Retention. Forgetting is natural — the brain discards what it perceives as non-essential. Spaced repetition works because you revisit the content exactly when you're about to forget it, signaling to the brain that it matters. Also use review projects, or simply practice a little more than necessary.
Principle 8: Intuition. Invest in difficult problems. Instead of memorizing theories, try to arrive at them through your own effort. Use concrete examples, ask questions, explain concepts in your own words. This is the creative side of learning.
Principle 9: Experimentation. Try different resources and techniques, copy others, compare methods, create arbitrary constraints, combine different skills to create something new, and explore the extremes — get out of your comfort zone.
The book also includes two additional sections.
How to run an ultralearning project. A step-by-step guide for structuring a project from scratch: research the subject and the method, gather materials, build a realistic schedule, execute, review what worked and what didn't — and at the end, decide whether you want to keep that skill, deepen it, or simply let it go. That last step matters: not every learning project needs to become a permanent commitment.
Raising children as ultralearners. The central idea here is to cultivate autonomy from an early age — raising someone who knows how to learn on their own, rather than always depending on a teacher or adult for guidance. To do this, the author suggests starting early, tailoring learning to the child's interests, turning practice into play so it doesn't feel heavy, reinforcing progress positively, and above all giving space for self-direction — resisting the temptation to over-control the process.
How to Study in 2026
As mentioned, the book is full of useful ideas, but since it was published in 2019 it doesn't address the more recent use of artificial intelligence tools and learning apps.
But before we get into that, an important warning: using AI excessively can sabotage exactly what you're trying to develop.
The book is clear on this point — difficulty is often desirable. The more effort you put into retrieval, solving a problem, or producing a piece of writing, the more you learn. Asking AI to write for you when you want to learn to write, or to solve the exercise when you want to learn programming, is the opposite of what the book recommends.
In short, use AI after you've tried, not before. It's an excellent sparring partner, reviewer, and exercise generator — but it should be used as a coach, not a crutch.
Note: apps marked with an asterisk I haven't personally used.*
For metalearning, use AI to map the subject before you begin: "what are the essential concepts to learn X?", "what can I skip at the start?", "what's the ideal order?" In minutes you have a map that would previously have taken hours of research. Always review that map, though, and make sure it aligns with your goals.
For focus, the app Forest* gamifies the Pomodoro: you plant a virtual tree that dies if you leave the app during the session. Simple and surprisingly effective. Focus To-Do* combines a Pomodoro timer with a task list, also free. Personally, I just use a tomato timer on my phone or computer.
For direct practice, AI is a solid partner. Want to learn English? Speak English with it from day one. Want to learn to write? Write and ask for immediate feedback. Want to learn programming? Ask it to give you a real project to build, not abstract exercises.
For drill, after practicing anything, ask AI: "what were the weak points in that response?" or "give me exercises focused only on this specific point." It's a tutor available at any time.
For retrieval, Anki* is the most established app for spaced repetition with flashcards — free, with ready-made decks for almost any subject. Quizlet* is a more visual and collaborative alternative. But you can also simply ask AI to quiz you on what you just learned, exactly as we did in this article. Personally, this is the tool I've been using most. I ask AI to create tests for me — starting with multiple choice, then short essay questions, and so on.
For feedback, AI solves a long-standing problem: quality feedback used to require having a teacher or mentor available. Today, you paste a text, a piece of code, or an answer and ask for specific critique. The key is not to ask "what did you think?" — but rather "what is weak, imprecise, or confusing?"
For retention, Anki* already covers a lot. But you can go further: ask AI to generate a set of flashcards from any content you paste — a PDF, a summary, a book chapter. Personally, I like the idea of scheduling "tests" at set intervals. For example, take a test one day after studying. If it goes well, schedule another for five days later. If it goes well again, for twenty days. If it goes poorly, study a bit more and repeat in a week. Remember that the format of the test should be close to the real situation in which you'll use that knowledge — it's the principle of direct practice applied to review.
For intuition and experimentation, AI works as a sparring panel: try to solve the problem first, present your reasoning, and ask for an evaluation. This forces the cognitive effort that builds real intuition — very different from simply asking for the answer.
Help spread the word about this blog! If you enjoyed this article, check out other posts and share it with as many people as you can. Thanks for reading!
