← all guides
Media

We take no money from any platform, school, publisher, or creator. Nothing here is sponsored. We judge learning resources by public-interest value, openness, depth, accessibility, calm design, and cost.

Learning without the feed

The internet made learning abundant, then wrapped much of it in feeds, certificates, upsells, ads, streaks, and subscription funnels. The hard part is no longer finding something educational. It is choosing a resource whose incentives match the kind of learning you actually want.

The honest one-paragraph answer. Start with the public-good layer: libraries, open courseware, open textbooks, nonprofit explainers, public archives, and specialist references. Use commercial platforms when they solve a specific problem, but do not confuse polish, celebrity, or a certificate with depth. A good learning resource helps you leave with knowledge, notes, files, or skills. A weaker one keeps you watching, renewing, or chasing a badge.

Weigh what you care about

AxisWhat to look forWhy it matters
Learning valueClear curriculum, practice, references, examples, or expert authorshipA beautiful lesson is not enough if it does not change what you can understand or do
Free and openPublic domain, Creative Commons, open-source, free audit, or library accessOpenness lets people learn without asking permission or paying rent forever
DepthFull courses, primary sources, textbooks, research databases, or sustained projectsDeep resources can replace a class; shallow ones are better as introductions
AccessibleCaptions, transcripts, plain formats, mobile support, many languages, beginner pathwaysKnowledge should not require perfect bandwidth, money, language, or credentials
CalmNo infinite feed, autoplay, urgency sales, streak pressure, or algorithmic rabbit holesLearning needs attention; engagement design spends it for you
EconomicalFree, low-cost print, public library, nonprofit, or generous audit modelThe best choice is the one you can keep using without financial stress

The commons-first stack

Open Educational Resources, or OER, are learning materials that can be freely accessed and reused under open licenses. UNESCO defines OER as materials in the public domain or under an open license that permit no-cost access, reuse, adaptation, and redistribution. Creative Commons makes this practical by giving authors and institutions standard licenses for sharing educational work.

That matters because education is not only a product. It is infrastructure. OpenStax textbooks, the Open Textbook Library, LibreTexts, OER Commons, MIT OpenCourseWare, MITx Online, Open Yale Courses, OpenLearn, MDN Web Docs, Wikibooks, Wikiversity, Wikisource, Wikimedia Commons, Project Gutenberg, Standard Ebooks, Wikipedia, PubMed, arXiv, PBS LearningMedia, Libby through public libraries, and public archives are all parts of a learning commons: imperfect, uneven, sometimes dense, but not built primarily to keep you subscribed.

The right default is not "never pay." It is "use the commons first, then pay for a specific gap."

Pick the path before the platform

Before comparing platforms, name the learning path. This prevents the common failure mode where a polished course, app, or creator decides the shape of the goal for you.

PathStrong first sourceProof you are learning
understand a topicopen textbook, syllabus, encyclopedia, or public explainera one-page map of terms, debates, and sources
build a skillproject-based course, documentation, exercises, or community archivea working artifact, solved problem set, or practiced routine
research a questionlibrary database, archive, primary source, or specialist indexcited notes that distinguish evidence from interpretation
prepare for a credentialofficial exam guide, accredited provider, or recognized courseperformance on practice tasks that match the gatekeeper
teach or shareOER collection, public-domain media, reusable lesson, or open image bankmaterials you can legally adapt, credit, and preserve

Once the path is clear, the platform becomes a tool instead of a home. That shift makes it easier to use commercial resources without letting subscriptions, streaks, or certificates define the whole project.

Choose by the job

If you want foundations, start with structured public resources: Khan Academy, OpenStax, OpenLearn, MIT OpenCourseWare, MITx Online, Open Yale, PBS LearningMedia, CK-12, Code.org, Saylor Academy, GCFGlobal, Carnegie Mellon's Open Learning Initiative, Wikibooks, Wikiversity, or BBC Bitesize. They are less glamorous than many paid products, but they tend to teach the basics without turning every lesson into an upsell.

If you want language practice, use public-broadcaster and nonprofit resources before paying: BBC Learning English, British Council LearnEnglish, VOA Learning English, DW Learn German, Duolingo, Anki, and library language databases. Pair listening with production. An app streak is not the same as conversation.

If you want primary sources, use Project Gutenberg, Standard Ebooks, Wikisource, Wikimedia Commons, Internet Archive, Open Library, PubMed, arXiv, JSTOR through a library, the Library of Congress, National Archives educator resources, Smithsonian Learning Lab, Europeana for Educators, NASA learning resources, Google Arts & Culture, or a university/public archive. This is slower than a summary video, but it gives you contact with the actual text, paper, data, object, or record.

If you want art, history, music, and cultural context, start with museum and archive projects that show their sources: the Met Heilbrunn Timeline of Art History, Getty Education, Smithsonian Learning Lab, Europeana, the Library of Congress, National Archives educator resources, DPLA Primary Source Sets, HathiTrust, the Biodiversity Heritage Library, IMSLP, Musopen, and Google Arts & Culture. They are strongest when used as primary-source trails, not just as pretty browsing.

If you want research literature, use the right search layer for the question. Google Scholar and Semantic Scholar are useful broad entry points. DOAJ and CORE are better when you want open-access literature. PubMed, arXiv, ERIC, OpenAlex, and WorldCat are better when you need a specific disciplinary index, open research graph, or library trail. Treat citation counts, AI summaries, and "related papers" as maps, not verdicts.

If you want real-world data, use public data portals before reaching for chart screenshots. Data.gov, Census Academy, FRED, World Bank Open Data, OECD Data, Gapminder, Our World in Data, OpenAlex, and library data guides help you learn from actual datasets. The trick is to ask boring questions carefully: who collected this, what is missing, what changed definitions, and what does the chart hide?

If you want public-service literacy, prefer agencies and public institutions over influencer explainers. Investor.gov and CFPB resources are better starting points for money basics than trading apps. MedlinePlus and OpenWHO are calmer health-learning sources than search panic. FAO elearning Academy and UN CC:e-Learn are useful for food systems and climate literacy. Cornell's Legal Information Institute and Oyez are better for law and civics than hot takes about a case.

If you want practical skills, prefer project-based paths where you build something: freeCodeCamp, The Odin Project, MDN Web Docs, Harvard CS50, W3Schools, Code.org, Scratch, Instructables, The Carpentries, Exercism, Kaggle Learn, Microsoft Learn, documentation, practice problems, and communities with searchable archives. LeetCode can help for coding interviews, but it is interview training, not the same thing as learning to design and ship useful software.

If you want math and science practice, combine explanation with manipulation. 3Blue1Brown and Khan Academy can explain ideas; GeoGebra, Desmos, PhET, Wolfram MathWorld, Project Euler, HHMI BioInteractive, NOAA Climate.gov, USGS Education, Exploratorium Science Snacks, National Geographic Education, NASA Citizen Science, SciStarter, iNaturalist, and Brilliant help you test them against problems, data, maps, labs, living systems, and real research tasks. The better stack is usually one explainer, one reference, and one place to struggle with problems.

If you want a full textbook or serious book, check open catalogs before paying: OpenStax, Open Textbook Library, LibreTexts, CORE Econ, Noba, Wikibooks, Project Gutenberg, Standard Ebooks, Open Book Publishers, OpenEdition, and course-specific open materials. A textbook is not always friendlier than a video, but it is often more complete, more citable, and easier to revisit.

If you want teaching materials, use commons-oriented discovery before buying worksheets or slide decks. OER Commons, MERLOT, Openverse, Wikimedia Commons, Smithsonian Learning Lab, Europeana, PBS LearningMedia, CK-12, OpenSciEd, Illustrative Mathematics, ReadWriteThink, and library databases can provide reusable lessons, images, primary sources, interactives, and classroom resources. Check licenses; "free to view" and "free to reuse" are not the same promise.

If you want early reading or accessible books, look beyond the commercial ebook store. StoryWeaver, Global Digital Library, African Storybook, Bookshare, the National Library Service for the Blind and Print Disabled, Libby, public libraries, and public-domain book projects can remove the barrier that matters most: language, print disability, money, or availability.

If you want civics, news, and media literacy, use iCivics, Checkology, Common Sense Education, PBS NewsHour Classroom, KQED Teach, public media explainers, and local library resources. They work best when paired with current examples and discussion. A media-literacy lesson should make you better at checking claims, incentives, sources, and missing context.

If you want digital power literacy, learn the institutions and incentives behind the tools. Data & Society, AlgorithmWatch, Tech Policy Press, Deceptive Patterns, Tactical Tech's Glass Room, Privacy International, Ranking Digital Rights, Berkman Klein Center, Tow Center, All Tech Is Human, Global Network Initiative, Internet Safety Labs, and Integrity Institute are useful because they teach systems: platform governance, data brokers, AI accountability, manipulative design, trust and safety, journalism/platform dependence, and public-interest technology. This is the bridge between "protect my account" and "understand the market I am inside."

QuestionBetter learning source shapeWhat to do after reading
Why does this app want my data?digital-rights or privacy-research groupinspect one app permission or privacy setting
How do platforms shape attention?journalism, platform-governance, or trust-and-safety researchredesign one feed or news route
What are dark patterns?pattern library plus enforcement examplesaudit one cancellation, checkout, or privacy flow
How is AI governed?AI accountability institute, policy observatory, incident databasecheck one AI service's data-use and risk documentation
How can I work in public-interest tech?responsible-tech community, civic-tech field guide, public-interest pathwaychoose one concrete skill or organization to follow

If you want course discovery, use directories carefully. Open Culture, Class Central, MERLOT, and WorldCat can help you find free university courses, textbooks, library holdings, and certificate options across many providers. Their job is discovery, not depth. Once you find a course, judge the actual provider's cost, exportability, ads, certificate value, and cancellation terms.

If you want career training, treat commercial platforms as tools, not homes. FutureLearn, Alison, DataCamp, Udacity, Skillshare, LinkedIn Learning, Coursera, edX, and Udemy can be useful when they provide structure, projects, feedback, or a credential you need. They are weaker choices when they sell urgency, shallow completion, or subscription access to material you could learn from the commons.

If you want library access, check the public layer before buying. Libby, OverDrive, local library databases, interlibrary loan, Kanopy, JSTOR access through a library, HathiTrust public-domain access, DPLA, NLS where eligible, and physical branches can quietly beat another subscription when the goal is reading, listening, research, or a calm place to study.

If you want orientation, short explainers and educational channels can be useful. TED, CrashCourse, YouTube education, newsletters, and podcasts are often best as maps, not destinations. Use them to find the question, then move to deeper sources.

Build a learning loop

StepWhat to doWhy it works
OrientUse an encyclopedia, syllabus, glossary, or short explainerYou need the map before the rabbit holes
StudyChoose one deep source: course, textbook, archive, or referenceDepth prevents endless sampling
PracticeSolve problems, build a project, annotate, translate, teach backLearning needs retrieval and friction
KeepSave notes, citations, files, flashcards, or project artifactsDurable work survives the platform
ReviewReturn after a week and repair what was fuzzySpaced repetition beats binge completion

This loop is a quiet antidote to feed learning. The feed rewards orientation forever: another explainer, another thread, another video. The loop moves you from finding to doing to keeping. A resource that helps you complete the loop is usually worth more than one that merely feels educational while you remain inside it.

Use AI as a study partner, not the source of record

AI tools can make a hard resource easier to work with: ask for a quiz, a second explanation, worked examples, flashcards, or a plan for practicing. That is different from making the tool the authority. UNESCO's guidance on generative AI in education and research frames the issue around privacy, age-appropriate use, policy, and human-centered learning; for a learner, the practical rule is simpler: AI can support the loop, but the source trail and the final artifact should still be yours.

AI useBetter boundary
summarize a chapterread the source, then ask AI to quiz your own notes
explain a confusing paragraphcompare the answer to the paragraph and one outside reference
make flashcardsedit every card yourself before studying
draft a study plantie each step to a textbook, course, paper, or documentation page
answer factual questionsverify with a primary source before saving the claim

For factual, medical, legal, financial, historical, scientific, or technical claims, open the primary source, textbook, standard, paper, documentation, or public institution page. The learning artifact should be yours: notes, solved problems, citations, code, diagrams, or a teach-back, not only a pleasant generated summary.

Use a stop rule for course shopping

Learning-resource discovery can become its own hobby. Before adding another course, choose one current resource and define a small finish line: complete two chapters, build one project, solve ten problems, write a one-page summary, or teach the idea to someone else. If the resource cannot support that finish line, replace it. If it can, stop shopping until the work is done.

Drift patternStop rule
collecting coursesone active course per topic
watching explainers foreverone explainer, then one exercise or primary source
buying certificatescredential only when a real gatekeeper values it
saving endless linksweekly review: keep, use, or delete
asking AI for summariesverify with one source and make your own note

The point is not discipline theater. It is protecting the scarce thing learning actually needs: sustained attention on one hard-enough path.

Check the license, not only the price

"Free to view" is not the same as open. An OER text, Creative Commons image, public-domain book, or open-source course can often be reused, adapted, translated, printed, shared, or preserved. A free commercial course may disappear, block downloads, restrict classroom use, or sit behind a login that tracks learners. UNESCO's OER definition is useful because it includes both no-cost access and permission to reuse, adapt, and redistribute. That second half is what turns a lesson into infrastructure.

The marketing traps

  • Certificate theater - a certificate can help in some job markets, but it is not the same as skill. Ask whether the platform gives you practice, feedback, and durable work.
  • The infinite course shelf - buying ten courses can feel like learning while quietly replacing the hard part: doing the work.
  • Urgency pricing - perpetual "sales" and countdowns reward impulsive buying, not thoughtful study.
  • The YouTube rabbit hole - free videos can be wonderful, but the surrounding platform is optimized for the next click, not your finished thought.
  • Celebrity over curriculum - famous teachers can inspire, but a well-structured open textbook may teach more.
  • AI summaries as substitute understanding - summaries can help you preview a topic, but they can also hide uncertainty, citations, and the effort that makes knowledge stick.

A reasonable default

Build a small, calm learning stack:

  1. Use one public reference for orientation: Wikipedia, Britannica, or an encyclopedia in your field.
  2. Use one deep source for study: an open textbook, open course, archive, or library database.
  3. Use one practice loop: problem sets, projects, flashcards, notes, or teaching the idea back.
  4. Use commercial platforms only when they provide structure, feedback, community, or credentials you actually need.

This keeps the center of gravity in public, portable learning rather than subscription discovery.

What to pay for

Pay when the resource gives you something the commons does not: live feedback, accountability, mentorship, specialized equipment, professional credentialing, expert curation, or a teacher whose work you want to support directly. Paying for that can be a values-aligned choice.

Be slower to pay for recycled public material, shallow inspiration, or platforms that make canceling, exporting, or learning off-platform difficult.

The deeper standard

A conscious learning resource should pass a simple test: after an hour, are you more capable, or merely more likely to keep consuming?

The best resources give you durable understanding. They let you download, cite, revisit, adapt, practice, and leave. The worst ones make education feel abundant while turning your curiosity into another engagement metric.


Compare learning resources by your own weighting in the learning-resources explorer. For the wider attention economy, read digital literacy, healthy tech, and reading the news without being played.

Read next
Reading without feeding the monopoly

The book is the same almost everywhere. The difference is what happens around it: who gets paid, whether a local shop survives, whether a library can lend it, whether you actually …

Streaming music without starving the music

Streaming solved one problem beautifully: access. Almost every song is a tap away. It also created a harder problem: most listening now flows through platforms where the listener p…

Reading the news without being played

The thing most likely to mislead you in the news is not an outright lie. It is the structure you cannot see: who owns the outlet, how it makes money, whether it corrects itself, an…