Coursetexts

Project summary

Coursetexts is a library of advanced course material. We’re publishing course notes for Harvard’s graduate-level classes as those courses are being taught. All our courses are open sourced with the professor’s explicit consent.

We grew up teaching ourselves online. When school didn’t satisfy us, we turned to the Internet: computer science papers, physics textbooks, and essays galore. But our reading often hit a wall; not from lack of curiosity, but lack of access. Paywalls or out-of-date material limited what we could learn. We have high hopes for Coursetexts. We want it to curate the highest quality education on the planet so anyone can be a true expert in their desired field. And, we want it to serve as a guide to frontier research as its being published.

We’re building the library we wish we had — and we’re excited for the next generation of learners to do more than we could have imagined. You can see our existing courses at coursetexts.org.

What are this project's goals? How will you achieve them?

Overall, we want to open source as many new, advanced courses as possible for as low of a unit cost per course. Our concrete goal is 50 by August 2025. Right now, we have 20 up and it costs ~$400/course between AI compute to clean up the videos, and to pay contractors to edit and review the courses. We believe we can drive this down via automation.

How will this funding be used?

Money will go to both engineering work to automate the most expensive processes, build copyright handling pipelines with AI to attribute and replace copyrighted material, and pay contractors for their time to review course info manually.

Who is on your team? What's your track record on similar projects?

Selena; MIT grad who founded Momentum AI, a 501c3 that helped 1000+ underresourced high schoolers to ship projects in ML research and web development.

Aayush; MIT graduate who previously co-started and currently advises MIT SOUL on open-sourcing the MIT economics curriculum. Brings experience in building organizations and AI processing pipelines, with direct knowledge from work with MIT OpenCourseWare. Has done AI research, and led multiple open source organizations.

Our ops team (Yassine, Josh, Ezra) helps with outreach, quality control, and development. We are advised by Lawrence Lessig, Peter Suber, and Michael Nielsen, each of whom has at least 10 years of experience in open source education.

What are the most likely causes and outcomes if this project fails?

If this fails, we will pause open sourcing new courses, but will be able to leave all the existing courses and videos up for the public indefinitely.

How much money have you raised in the last 12 months, and from where?

We raised 10K from The Institute and Michael Nielsen. We spent that open sourcing our first 20 courses, available on coursetexts.org.

Approving this proposal, and funding it to half of the total ask. I'm not personally in the target demographic (it's been a while since I've studied off course notes), I'm excited to fund this becuase:

Michael Nielson and Lawrence Lessig are two folks I've long admired, and their support is a strong signal
While I was not previously familiar with them, the team behind coursenotes seems quite strong, with multiple interesting projects to their name
The design of the site itself is nice, and I appreciate the clarity of vision listed on their "why" page

Best of luck to the team!

(Aside, a minor UX nit: I think this is better as a single select rather than multiselect, rn almost no coursenotes fit the intersection of any two options)