There is no shortage of free learning material. The problem for self-learners is the opposite: too much, uneven quality, and no obvious way to tell which video or course is worth your next ten hours. This guide gives you a repeatable way to find, test and organise learning content, whether it comes from a search engine, a video platform or an AI assistant.
The short answer is to start with what you want to be able to do, look in a few reliable places, give every resource a two-minute quality check, choose a small stack of three, and test each new resource with a short trial before you commit.
Why should you start with the outcome and not the platform?
Before searching, write down what you need to be able to do. "Understand basic statistics" is vague. "Explain and calculate a confidence interval for a class project" is searchable and testable. Use that sentence to filter, because anything that does not move you toward it is a distraction, however good it looks. Without it, you end up sampling well-made material that never adds up to a skill.
Where should you look?
University open courseware is a strong place to start, since many universities publish full course materials, including problem sets, syllabi, lecture notes and past exams. Non-profit and public-education platforms are another, and Khan Academy, for example, is widely described by users as free and good at explaining things. Video lectures are good for demonstration, especially in maths and science. One student said of maths and physics that "i dont find AIs as helpful as YouTube videos," a reminder that worked examples on video still have a place.
Textbooks and open textbooks are slower but more complete than most videos, and if you are preparing for a test, the official exam body's syllabus and sample papers are the gold standard. AI assistants can help you find terms, suggest sequences and explain things differently, but treat what they say as leads to verify, not answers to trust.
How do you judge quality in two minutes?
Give any resource two minutes and ask six questions. Who made it and why: a university, an experienced teacher, or a company selling something? Is it current? Dates matter for software, science and policy. Does it include practice, since content without exercises is reading and not learning? Does it show its reasoning, explaining why and not only what? Do others who know the subject recommend it, and do the reviews come from learners at your level? And does it match your goal or exam? A common warning is that AI tends to choose "what it thinks is important but not what exam actually focus on," and the same applies to generic courses.
Does free and open mean vetted?
No, and even respected open resources can have blind spots. University of Kansas researchers analysed 46 sets of open educational resources used in middle and high school social studies units on imperialism, across four platforms. They found that 85.7% of the primary sources were recorded by colonizers, with few from the people who were colonized, and that some platforms framed the topic as a "balance sheet" of costs and benefits (Phys.org summary, published in Social Studies Research and Practice). The lesson for any learner or teacher is to check whose voices a resource includes and whose it leaves out, and to compare it with a second source.
How careful should you be with AI-recommended sources?
Very. AI assistants can suggest resources that sound plausible but do not exist, or summarise a source inaccurately. Always click through to the real page, check the author and the date, and compare the outline with a second source before you rely on it.
What does a search look like for a real goal?
Take the goal from earlier, "Explain and calculate a confidence interval for a class project." For the core, look for an open university course or an open textbook whose syllabus lists confidence intervals, and check the week-by-week outline first, since it tells you how much you need to cover. For practice, use the problem sets that come with that course, plus one dataset you care about, such as sports scores, weather or your own habits. For reference, find a short glossary or a well-reviewed video series for when a term trips you up.
Then use an AI assistant, but in a supporting role. Ask it to explain the same idea two more ways and to quiz you, and do not ask it to choose your course for you. Notice that you spent about thirty minutes choosing, and the rest of the time will go on doing.
How do you read a syllabus quickly?
A syllabus tells you more than a course description. In two minutes, check whether it lists the topic you need, whether there are problem sets or projects and not only lectures, whether it says what you should know beforehand, and how many weeks it assumes and whether you can keep to that. If the prerequisites are missing from your background, either choose an earlier course or plan a short detour to fill the gap.
How do you build a small, ordered stack?
Rather than bookmarking 40 things, pick a stack of three: a core course or book that gives structure, a practice source such as problems, projects or a question bank, and a reference to look things up in. Order them by week. If you find something better, swap it in and do not add it on top.
How do you test a resource before committing?
Spend 20 minutes with a new resource, then try three questions on what it covered without looking. If you can answer them, keep it. If not, it may be too advanced, too shallow or a poor fit, and it is better to find out in twenty minutes than in twenty hours.
Is free the same as good, or forever?
Not always. Free material can disappear, change or move behind a login, so save the parts you need, such as notes, problem sets and your own summaries, in a format you own. Tools and offers change too. In August 2026, for instance, Google announced a free year of its Gemini Pro plan for college students, and offers like that are usually time-limited. Use them, but do not build your study plan on something that may stop.
Should you ask a person?
Yes. Do not overlook the cheapest expert available. A teacher, librarian or someone a few steps ahead of you can tell you in five minutes which resources are worth the time. Ask, "If you were starting again, what would you use?"
What is the collector's trap?
Saving links feels productive, but studying is what makes progress. A useful rule of thumb is never to have more than one unfinished resource per topic.
How Tutor AI helps
Tutor AI is built for self-learners who want help with the finding as well as the learning. It helps you locate content for what you want to learn and puts it in a sequence, then tutors you through it with questions and interactive practice. You can see whether it is working, because it tests you on the content, not just delivers it.
Frequently asked questions
How many resources do I need? Usually three: a core, a practice source and a reference. More than that tends to become collecting instead of learning.
Are YouTube courses good enough? Some are excellent, especially for worked examples. Apply the same quality check as for any resource, and combine video with practice you do yourself.
How do I know a course is current? Check the date, and for fast-moving subjects such as software, look for comments or updates that show it still works.
What if I cannot find anything at my level? Look for the previous course in the sequence, or ask a teacher or a forum. A short detour to fill a gap is better than struggling through a course that assumes it.
Related guides
- How to build a self-study plan
- NotebookLM vs ChatGPT vs an AI tutor
- Free AI tools for teachers
- Education News: latest research, a live feed of new education studies