AI Guides › Workbench

The Free Course Pick Table: Seven Harvard, MIT and Yale Courses, One Finished

By Nigel Guy · 6 min read

Most people treat a free-course list as a shopping basket. They bookmark seven, feel productive, and finish none, because nothing in the list forces a choice and nothing is at stake. Free is not the same as low value: the courses below are real university material. The cost is not money, it is the hours you will not spend on the other six.

The rule: bookmark seven, start one, and write down the date you will finish it before you open the first lecture.

The seven, side by side

I checked each of these against the official course or university page on 2026-10-04. Costs are what the pages say at time of writing; paid extras change, so check the current page.

Course What it is Cost at time of writing Best for Catch
Harvard CS50x Introduction to computer science, 11 weeks of material Free (£0) to take Complete beginners who want structure Certificate rules differ by route (see below)
Harvard CS50P Programming with Python, 10 weeks Free (£0) People who want Python specifically Optional paid edX certificate
Harvard CS50 AI Introduction to AI with Python, 7 weeks Free (£0) Learners who already code Needs CS50x or about a year of Python
MIT 6.0001 Intro to computer science and programming in Python (Fall 2016) Free (£0) Self-starters who like lecture plus problem sets Teaches Python 3.5; no certificate
MIT 18.06 Linear Algebra, Prof. Gilbert Strang (Spring 2010) Free (£0) Anyone heading toward data or machine learning Maths-heavy; prerequisites not stated on the page
Yale: The Science of Well-Being Laurie Santos, on Coursera, 6 modules Free to audit; certificate is a paid extra Busy people wanting a short, finishable course Free audit is not the certificate route
Open Yale Courses (for example Listening to Music) Recorded Yale College lectures Free (£0) Curious people who just want to learn No credit, degree or certificate at all

How to use each one

CS50x. Go to cs50.harvard.edu/x and work through the weeks. You submit problem sets and a final project after creating an account. The CS50 certificate page says a free certificate needs at least 70% on each problem set and the final project. The main CS50x page, by contrast, describes the free route as having no certificate and points to a verified certificate from edX, plus Harvard Extension or Summer School for transfer credit (tuition applies). Those two pages read differently, so check both before you rely on a certificate. If you only want the learning, none of this matters.

CS50P. Same site, same model. Start here if "programming" to you means "Python" and you do not want the broader computer-science tour. The page describes it as suitable with or without prior experience.

CS50 AI. Do not start here cold. The page asks for CS50x or at least a year of Python. Treat it as the course you earn by finishing one of the first two.

MIT 6.0001. On MIT OpenCourseWare you get lecture videos, notes, problem sets and programming assignments, aimed at people with little or no coding experience. No account is needed. The catch is age: it uses Python 3.5, so expect to adjust a few details on a current install.

MIT 18.06. Strang's linear algebra comes with video lectures, problem sets and exams with solutions. The solutions are the useful part: do the problems before you look.

Yale: The Science of Well-Being. Open the Coursera page and choose the free-access option rather than the certificate one. Coursera lists it as six modules, roughly a week at 10 hours, and also offers a "full course" option without a certificate. Coursera also says financial aid is available for the paid route. When it launched, press coverage quoted a certificate price in dollars; I could not confirm the current figure, so check the £ price at checkout.

Open Yale Courses. Recorded classroom lectures in video, audio and transcript, no registration, mostly under a Creative Commons non-commercial licence.

What is the honest limit on Yale?

Two different things sit under the Yale name, and neither gives you a Yale credential. Open Yale Courses states plainly that no course credit, degree or certificate is available through the site (two courses can be taken for Yale College credit through Yale Summer Online, which is a separate, paid route). The Coursera course lets you audit for free, but a shareable certificate needs the paid enrolment. The Open Yale page does not say whether individual lectures are up to date, so for fast-moving subjects treat them as a lecture series, not a current textbook.

How do you actually finish one?

Use the One-Course Commitment Card. Write four lines on paper before lecture one:

  1. The course: one name from the table.
  2. The weekly slot: two fixed sessions in your diary, with a start time.
  3. The finish date: count the weeks the course page lists, add a buffer week, and write the date.
  4. The proof: the one thing you will have at the end (a submitted final project, a solved problem set, three pages of notes).

If you miss a session, move it, do not delete it. If you miss three in a row, the card tells you the course was the wrong pick, not that you failed. Swap it and start the card again.

How to choose

What to skip

What won't this do?

It will not get you a degree, academic credit, or a job. MIT OpenCourseWare says it offers no degree, credit or certificate, and Open Yale says the same. It also will not keep you accountable: the card only works if you honour the diary slots.

Guardrails

Sources

All 751 AI guides · JulieMango plans from £17/mo