yesterday at 8:49 PM
Hi! I'm Phillip Compeau, I teach in Carnegie Mellon's Griffin School of Computer Science, and I run CS Learn, which launched today.
We're releasing full semesters of lectures from current CMU computer science courses across a variety of topics, with one course premiering each day this week. More courses come in January.
Yes, MIT got here 25 years ago :). What I'm excited about is how we're doing it: a few of our students built tools that automate a lot of the production work, so a small team can get a course out quickly and still have people review every lecture.
We'll have lots more coming across a wide array of subjects. I'd love to know if there are topics you would like to see us cover!
today at 5:25 PM
I think this (and MITs precedent) makes an important claim about the value of higher ed: the value of an undergraduate education is not just access to curriculum, certification, or the admissions filter. Openness is an act of confidence that all the rest matters, too: the quality of instruction (and that is more than just lectures and distributed materials), the structure and guidance, the spirit of comity, immersion in academic values, having one's beliefs challenged, engaging in diverse ways of thinking.
For all of their flaws, I think universities are one of our greatest inventions and it's a mistake to view them purely in economic/transactional terms.
yesterday at 10:24 PM
CMU has some of the best courses on advanced topics, but the ones you have put out so far are the introductory courses. Would love to see courses on advanced topics such as database internals and distributed systems. Also please release materials such as slide decks and PDF course notes.
yesterday at 10:33 PM
Andy's database lectures are already all online: start here: https://youtu.be/APqWIjtzNGE?is=1QyZ_p0uYsSGYkN0
I plan to put my distributed systems lectures online the next time I teach it, but not sure when that will be after I rotate out of teaching 15-213. One of my colleagues who also teaches it might do it before then.
today at 4:25 AM
I don’t have a strong opinion on databases. I know they’re critical to modern computing and I use whatever works. I clicked on this link and started watching the first video. This dude makes me want to learn everything about databases.
today at 9:35 AM
I took an Advanced Databases course in my university (Edinburgh) solely because my thesis supervisor was the lecturer and he was an incredibly interesting guy. I hadn't taken the 101 course in the previous semester.
It was one of my favourite courses. The black-box concept of a database may be "boring", but DBMS have to solve almost all interesting compsci problems as part of their incredibly complicated stack. It's a great litmus test for architecture knowledge and the ability to context-switch across many different levels of abstraction/implementation detail.
today at 9:07 AM
Andy has that way about his course that makes you want to sit through the whole lecture
today at 11:50 AM
I studied database internals ~25 years ago, never specialised in them but I've always been interested. Has enough changed at the fundamental level that it's worth my time going through a modern course? A whole lot of what I remember was to do with optimisation wrt spinning metal.
yesterday at 10:46 PM
Thank you!
yesterday at 10:48 PM
actively sitting in the CUC forgetting that my professors are, in fact, people outside of class and do, in fact, also browse HN.
today at 6:33 PM
yesterday at 11:15 PM
The HN equivalent of seeing your homeroom teacher at the supermarket
yesterday at 11:23 PM
Had a teacher stop a class at one point and say “you guys think I’m just, like, the light in your refrigerator, that I just turn off when you walk out of here, don’t you?” - that one stuck with me.
today at 6:35 PM
today at 12:22 PM
I once saw my second grade teacher at Weight Watchers in the early 90s. My parents very wisely and simply told me "don't tell any of your friends, it would be very embarrassing to your teacher".
yesterday at 11:34 PM
The difference is that on HN the teacher can't see the student's shock and amazement at seeing their teacher in real life. I've been on the teacher's side of such encounters, and the flabbergasted expressions of some (though certainly not all) of the students were a sight to behold.
today at 3:40 AM
I know of two other public courses (have youtube playlist, lecture slides and assignments/homeworks) that are not DB or distributed systems, but seem to be pretty good:
1. 11-768: AI Agents [1]. I think this is ongoing.
2. CMU 10-799 Diffusion & Flow Matching [2].
[1] https://www.cmu-agents.com/#/
[2] https://kellyyutonghe.github.io/10799S26/
Edit: A third course that is going to be taught by George Hotz on compilers for machine learning [3]. I think he said lectures will not be publicly available. Initially all assignments were going to be public, but I think he changed his mind to only make one assignment public :(.
today at 6:03 AM
Agents course looks interesting but the recordings for more recent ones are not there.
today at 3:50 PM
It seems the Youtube playlist [1] has a few more lecture vids that are not currently listed on the course web page, however it's also missing the most recent ones. Not sure what's going on, but my guess is because the course is still ongoing the course leader is lagging behind uploading them (at least I know Stanford is usually like this).
[1] https://www.youtube.com/playlist?list=PLSN0qpDfUvTM
today at 6:28 PM
Is it me or is the audio "wobbly"? At first I thought the lecturer was really nervous, but it seems consistent throughout the video. The second hand anxiety I get from this effect makes it hard for me to focus on the content.
today at 2:00 PM
Visiting https://www.cs.cmu.edu/learn/ I just get the following popup.
"Your current browser version does not support nested CSS classes, making this site inaccessible. Please try accessing this website using an up-to-date browser."
Firefox 64bit somewhere around 113. Updated as far as the OS goes.
today at 2:47 PM
Firefox 113 was released on 9 May 2023 - over 3 years old! https://www.firefox.com/en-GB/firefox/113.0/releasenotes/
That's a LONG time in the world of web standards, bugs, privacy, security, etc
today at 2:12 PM
I do not update Firefox frequently, but looking at its version it still is some variant of 140.13, which I updated a few months ago.
With this Firefox version, the site works (on Linux).
113 might really be too old.
today at 2:37 PM
>Firefox 64bit somewhere around 113
Even firefox-esr is on 153 at this point. (Mainline is 159) Perhaps it's time to update like it is asking, but I assume you're just failing a version check so that is disappointing.
today at 2:24 PM
That's the latest of today: https://www.firefox.com/en-US/firefox/160.0.1 so maybe too far back?
today at 2:55 PM
Have you tried accessing the site using an up-to-date browser?
today at 4:02 PM
The parent did not ask for suggestions
today at 5:27 PM
And its parent did not ask for bug reports, yet they responded with a report for a non-issue caused by outdated software.
today at 1:16 AM
In fairness, Open Courseware predated widespread easy/inexpensive video though, once audio and video recordings became commonplace, they were never really integrated into OCW in a systematic way. (Of course, it originally also started off with a different--at least as publicly stated--purpose.)
Then, as you know, edX came along but the whole MOOC movement sort of fizzled over time though that's a complicated topic for another day.
In any case, nice addition!
today at 2:09 AM
MITx
today at 3:02 AM
Which became part of/associated with edX although I'm a bit fuzzy on how things evolved over time.
today at 1:21 PM
Anant Agarwal had 6.002x in 2012 to showcase exercises he built around the lecture video he uploaded from 2007(?) which evolved to MITx which evolved to edX once HarvardX and BerkeleyX got involved[1]. There was even an HN post announcing MITx[2]. (My memory is failing, but I believe it was used in the classroom the term prior, so 2011.)
Berkeley later hosted their own course content and later removed it because it was too burdensome to introduce accessibility features; lawsuits were threatened.
Harvard now has Online Courses. I think CS50 is still on Edx, although I think there were a few years it had its own website and grading system.
-----
6.002x was worlds ahead of the circuits/electronics classes I was taking in 2011--2012 at a state university and was both entertaining and informative. A few years later I worked remotely as a summer intern on 2.008 (manufacturing processes), they had pretty cool tooling (latex2edx, cat-soop) and I got an MIT email with Kerberos certificates, which I could then use to watch lecture videos on their media server and access course content that wasn't publicly available, including 3.091 lectures other than the two semesters available online plus assignments/videos from a whole ton of smaller, advanced classes.
I even made a ripoff of edX that created randomized variables in standard problem types so classmates could study for the FE/EIT exam, which helped me better understand heat transfer and electromagnetics and some other subjects.
[1]: https://www.csail.mit.edu/news/whats-next-edx
[2]: https://news.ycombinator.com/item?id=3665049
today at 5:39 AM
Eyes turned into dollar signs. All the early idealists left by about four years into edX. Shark tank of assholes remained, who continued to try to milk this to get (personally) rich or famous.
Some succeeded, but to the loss of humanity.
today at 6:44 PM
> I'd love to know if there are topics you would like to see us cover!
RSI.
today at 1:12 AM
Hi Phillip! CMU alum here - super exciting to see this on the HN homepage :)
How does this compare to some of the efforts led by David Kosbie & others around open access online CS education? I recall this being a big push back when I was there (2014-2018).
Is this essentially similar spirit, but covering actual SCS curriculum, rather than "pre-SCS"/highschool content?
today at 2:00 AM
Absolutely! And I've worked in open learning my entire career too. CSAcademy, which Kosbie and Mark Stehlik created, has been a fantastic effort, focused on intro CS and high school learners. Our aim is to have courseware aimed at a global audience across backgrounds; we are targeting early to mid career upskillers, but all are welcome.
yesterday at 11:31 PM
Hi Phillip, love to see this! I’m a working professional about to finish a Master’s in Cybersecurity Management. I’ve been debating the idea of a PhD on AI security (LLM, agents, NHI, etc). I’d love to see a platform where a working professional can obtain this at their own pace, using LLM for interaction, grading, and guidance, with a human overseeing the progression. I work in higher ed and I see your solution being highly effective at supplementing many of the gaps we currently have in academia. Looking forward to see how this evolves!
yesterday at 11:50 PM
seems pretty straight forward to take the content and ask an LLM to build the courseware?
yesterday at 11:58 PM
Maybe next year an LLM can supervise a PhD, but we're not there yet.
today at 12:23 AM
What would the pain-points be, in your view? Is it that the LLM is generally not “intelligent” enough, or is there some other limitations? As someone who enjoys using LLMs for self-education up to the bachelor level, I’m curious where doctorates might face issues.
today at 8:04 AM
From my experience using LLMs and supervising PhD students, the LLM is too focused on solving a task rather than educating or opening paths. It’s great if you have a specific question to ask, even at quite high level, it can explain you group theory or quantum chemistry quite well if you ask the right questions. The problem is that if students knew the right questions they would not be students. LLMs can get stuck explaining a topic without realising that it would be useful or interesting to branch out to something else. They can give you all the pros and cons of a given simulation technique without even trying to tell which of those actually matter in a specific case. They tell you if you ask, but you have to know what to ask.
LLMs are not great at open-ended discussions and don’t benefit of experience in the same way humans do. They are better to solve problems than to explore ideas. They keep getting better, though.
today at 1:15 AM
Your PhD advisor would be lending their wisdom and intuition to get you pointed down a promising (novel) research path. I don't think LLMs have been demonstrated to be at human level for that.
today at 12:43 AM
wasn’t there, a year or so ago, some sorta OpenAI plan where you’d pay $400 or something and get “PhD-level research“ capabilities? I wonder what happens to that :)
today at 2:36 PM
Really appreciate it. Thanks to everyone involved.
I would like to see more courses on intermediate to advanced CS topics.
More on scientific and numerical computing, functional programming, computer architecture, GPU programming, etc.
Kindly also post other course materials such as homeworks, handouts, psets, past test questions, etc.
today at 3:04 PM
ETH has fully open architecture courses taught by an ex CMU prof which he live streams https://m.youtube.com/@OnurMutluLectures The 'Computer Architecture' course is the advanced design course and if you go to the course page all assignments are open.
SciML book has recorded lectures and covers some numerical methods and is updated https://book.sciml.ai/ so does MIT's repository https://github.com/mitmath/numerical_hub
today at 7:17 AM
Sound awesome. Is CMU releasing the tools as well? Interested for my own courses. Another question: what hinders me is often copyright concerns regarding eg. images used: is this different in the US due to 'fair use'. It seems that in Germany the rules change as soon as I publish materials outside the university. Actually LLMs could help here cleaning up material as well.
today at 1:40 PM
CMU, that sounds familiar.
I think one of your courses has your students deploying and making changes to NodeBB?
In all the years you've had that in the course, we've only received a handful (single digits, really) of leaks in the form of errant PRs :) — so your warnings keeping your students away from the real repo worked!
Thanks for using open source to teach programming!
today at 2:25 PM
Not sure about the main campus but on the Qatar campus some of my friends have contributed to this as part of a course, foundations in software engineering I believe.
It's also to other open source projects too
today at 3:33 PM
Very cool!
today at 10:44 AM
Ha! Nice to see you here, Phillip, and with such a great gift! We used to work together on rosalind.info, I am Aleksey!
For the topics, I’d love a systems-perspective course on weak memory models:
* Which hardware optimizations we do at the level of physics in hardware.
* How those optimizations result in ISA-level memory model.
* How that memory model motivates higher level languages memory models as found in, eg Java and Rust.
* How compilers optimize code accounting for memory models.
* And what’s the state of formalizing this whole thing.
Separately, but relatedly, an implementation course for lock-free programming techniques would be nice!
today at 12:34 PM
Do some pop science something, like popular math YouTubers do. Pick some cool problems, preferably also some that are not solved or not solved well, and keep it simple enough, e.g., by introducing the basics needed to understand them.
today at 11:12 AM
I love that there will be an updated version online now for such courses now.
Yes MIT did it 25 years ago or w/e, but I'm sure _something_ has changed that is worth knowing about :D
today at 9:11 AM
@pcompeau, are we getting the assignments and recitation videos as well?
today at 2:50 PM
No, unless the class page persists after the semester (often doesn't), and assignment/recitation access doesn't require credentials which they usually do.
Some of these courses have a free online book with assignments and labs like the mlsystens course otherwise you have to invent your own assignments by working on whatever you want
today at 8:05 AM
I remember some years ago apavlo mentioned that lectures cannot be put online anymore because of some privacy (?) regulation. Has that been settled now and there are no more obstacles to putting videos online?
today at 7:11 AM
Hi, I'm an alum and love your bioinformatics courses and would really love to see like a full curriculum on the platform! The upcoming lineup so far looks really good too.
today at 6:21 AM
Kids these days are really lucky, all these self-learning materials. Back then I have to learn computer from TV shows that broadcasts 30 minutes every week, and they mainly teaches Microsoft Office (that's how I learned Front Page).
Just a suggestion: can you also Torrent these? These video is worthy of archiving and YouTube don't allow download, not to mention watch videos locally is far better than the online one. Thanks!
today at 8:08 AM
> Just a suggestion: can you also Torrent these? These video is worthy of archiving and YouTube don't allow download, not to mention watch videos locally is far better than the online one. Thanks!
There are things like yt-dlp for this kind of situation. Check your local laws for what is legal to download and what is not, if you care about this sort of things.
today at 7:02 AM
Thanks for this, and thank you for rosalind.info. it was a huge help when I pivoted into bioinformatics a decade ago.
yesterday at 8:51 PM
Hi! Love to see these released, thanks to all at CMU who help make it happen. Would y'all be willing to host this on a CMU CS Peertube instance as well versus only on Youtube?
yesterday at 8:59 PM
Great idea! We currently have a Panopto server that we're also posting videos to for internal use, and I will look into this as another option.
yesterday at 10:15 PM
You could push the videos to CloudFlare R2 and not pay egress fees, and provide an RSS feed for updates. An XML based index of the videos with transcriptions and video markers would be great.
Is your workflow open source? I have had great results building bespoke image editing pipelines using claude and ffmpeg.
It would be awesome to be able to ask an LLM questions and have it respond with lecture snippets.
yesterday at 9:01 PM
Thanks so much, I appreciate it. Happy to contribute financially if needed. Also happy to host the instance myself, at my cost, with CMU's permission.
today at 8:10 AM
> We currently have a Panopto server
It always made me laugh when we had one of those at a previous job. Did the people who created it really not know what the panopticon was? Could they not have taken a less dystopian name?
today at 8:34 AM
Hi Phillip, nice to see more university and their courses are opening to public learning. Good to see Carnegie Mellon's Computer Science here!
yesterday at 11:32 PM
It would be nice if people stopped posting this sort of stuff on YouTube. Or at the very least please cross-post it elsewhere.
YouTube has become very aggressive recently on trying to force you to sign-in, allegedly because its the only way it can "verify you're not a bot" – if you believe that, you believe anything. In reality its just Google being Google looking for ways to track you and add telemetry data.
Even without the "bot checks", I'm sick and tired of the ads and the intrusive pushes to watch other unrelated crap videos.
I just refuse to click on YouTube links anymore.
I've really no idea why people sitll support YouTube for serious videos. People should just leave the place to the Gen-Z content creators posing infront of something making sutpid faces and feigning surprise.
Rant over. :)
today at 7:15 AM
I LOVE YouTube. They have the widest most incredible set of content on every topic conceivable, with a great (if slightly responsive) recommendation algorithm. They host exabytes of video that anyone can upload to for free, and if you don't like ads its $15/month for al of it. I think people take all this for granted and don't even consider paying for it, but its one of the best most pro-consumer web platforms around.
today at 5:32 PM
I'm a big fan of Youtube as well, and for mostly the same reasons. But it's not all rainbows and unicorns. For example, their search sucks balls. You get all sorts of results in search that have NOTHING to do with the supplied query, and the filters don't even let you do basic things like sorting by upload date. :-(
today at 12:16 AM
Amazon is doing the same; now you have to log in to browse book reviews. And worse, even though I did put in my phone number to get the one-time code, it still asked for an email otc, and I don't even remember the password to that email because the bloody Yandex logged me out without my permission, and Firefox updated itself (again without my permission, causing me to lose my saved passwords data); so now I can't read Amazon book reviews. [attempts to retrieve my password ended up in Yandex wanting to collect so much more data on me that I felt nah, fuck it).
So your rant is justified.
today at 2:58 AM
StoryGraph is an independent alternative to Goodreads/amazon if you are looking for book reviews.
today at 4:47 AM
Thank you so much. I usually read book reviews as examples of customer research (in general but also for marketing books)-- and really disliked Amazon tightening the noose like that.
Wow I like StoryGraph. I am there as
janmsavage
Gonna add a few books to review and read others' reviews. This is super cool.
today at 6:04 AM
Imdb review reading ask to create an Amazon account...
today at 7:28 AM
ta, good idea!
today at 12:02 AM
Started having way less trouble this year, with ordinary win-10, Chrome, and uBlock Origin Lite. Probably even better with Firefox & uBlock Origin.
https://en.wikipedia.org/wiki/UBlock_Origin
today at 12:20 AM
Some of it's stuff you can't control, like other people on the same local ISP.
today at 3:00 AM
Probably much worse with Firefox, at least with its security settings enabled. Google doesn't need to hassle you to login when they have another way to create a persistent identifier for you. Chrome is chock full of security vulnerabilities that let Google (and the rest of the surveillance industry) fingerprint you. They don't consider your privacy as part of their security model, so WONTFIX.
yesterday at 11:42 PM
Peer tube ftw
today at 4:43 AM
youtube is the biggest tech monopoly that exists and somehow nobody fucking talks about it or takes it remotely seriously and i don't understand why.
today at 1:36 AM
Thanks for putting this out there. It is exciting to see more people having access.
I always loved core computer science and will take advantage of what I can. I also have a plethora of aspirational book readings that I have never gotten to in the 21 years since graduating Georgia Tech. Life got busy.
yesterday at 10:30 PM
Thank you for doing this!
In particular, I would love to see you cover CAD, esp. if you could work up a way to use FreeCAD as the subject.
yesterday at 11:25 PM
This. If someone has any recommendations for geometric kernel learning resources, I'd love to hear them.
yesterday at 10:47 PM
FreeCAD, & other Linux-centered apps, would be awesome to see in a curriculum.
yesterday at 11:25 PM
GP meant “how to write a CAD system”, which is a suitable topic for a CMU course.
Not “how to use FreeCAD”.
today at 1:45 AM
This. CMU teaches you how to build the system, not how to use the system.
today at 2:14 AM
Ok, my bad :)
today at 2:14 AM
Oops, my bad :)
today at 9:14 AM
Congratulations!! This is awesome. I'm glad my kids and millions around the world will have more options
today at 1:02 PM
Love your biological computing courses!
today at 10:59 AM
This is great! Are you also putting the syllabi, course notes, practice exams, exercises, and/or exercise solutions online? Those are by far the most valuable part of MIT OCW for me, especially the syllabi and exercises.
I can't learn machine learning by watching someone talk about machine learning; I have to actually try to do it. And I think this is true of everyone, which is why university courses universally have problem sets, even when they aren't part of the course grade. Watching someone explain something without doing it just gives rise to the illusion of fluency, not actual competence.
I did try to click through on your page to find out more, but just got diverted to what looks like a predatory vaporware startup page inviting me to "Mange your email preferences for CS Learn". Mange! Presumably that isn't intentional, but it is definitely funny.
Also, what license are you using for the material you post? MIT had to remove Walter Lewin's excellent lectures from OCW and their YouTube channel to stop him from sexually exploiting students, but, because they are CC licensed, people continue to benefit from the lectures even today. And YouTube is getting more and more predatory and unreliable, so it is essential for people to be able to share your videos on other platforms. What kind of licensing do you plan to use to protect their rights to archive and share this courseware?
yesterday at 10:53 PM
410 or GTFO IMO :p
I do wonder if/how you plan to release supporting materials. The labs made 213 awesome, and LLMs actually do a reasonable job of reproducing the lab itself (I did it for bomb lab for example). But it'd be cool to have a canonical one. I'd love to try malloc again "as an adult."
yesterday at 11:21 PM
Really would like to see https://boston.lti.cs.cmu.edu/classes/11-642/ made available
today at 4:56 AM
I think many people will be interested in the prospect of earning digital badges through the free courseware. Are these like edX or Coursera-style badges that go to Credly and sometimes award college credit, or do you have your own system in mind?
today at 8:15 AM
Are we promoting commercial rote learning with these badges and certifications? In future these badges and certifications will have no value. We should learn for improving our knowledge and not to show off. In this AI era our LinkedIn, GitHub, Credly profiles have lost their mojo. Education has gone backward after the introduction of rote learning. Does marks motivate you to learn more? It rarely does. I get satisfaction when I utilize my gained knowledge to solve real world problems not by earning digital badges.
today at 12:58 AM
Nice one! Love this initiative, are those tools available by any chance? Would love to have the same on my university
today at 3:07 AM
i'd love to see graduate level courses on YouTube as well. Are they included in this plan?
today at 7:57 AM
> We're releasing full semesters of lectures from current CMU computer science courses across a variety of topics, with one course premiering each day this week. More courses come in January.
That’s fantastic! I have a nephew who’ll love this. It’s great for kids around here who could only dream of getting to a top university in the US, even before Trump. Thank you for the work, and keep it up!
> Yes, MIT got here 25 years ago
It does not really matter. Better late than never, and diversity is good. MIT OCW should be more of a motivation to do better than a reason to give up :)
yesterday at 10:25 PM
Thank you! CMU grad here! I miss the Panopto platform, I prefer it over YouTube for lectures.
yesterday at 11:00 PM
I would write a mountain of a paragraph of what this means to me but my immediate response is…
Hell Yeah!!
yesterday at 11:54 PM
Makes me wish for more hours in a day to watch
today at 1:45 PM
I’d love to hear more about your production workflow! It sounds a lot like what I do at work. I imagine you cut very little of the actual lesson itself and most of the work is media management/polish but I’d be curious to hear if that’s the case
today at 7:54 AM
Amazing thank you so much
today at 12:24 AM
Great to see this (CMU alum!)
yesterday at 10:17 PM
Do you/could you release any information about the automation of production work? I'd be interested in what those tools look like and how they're improving your workflow.
yesterday at 10:19 PM
It's still very much a work in progress as we scale up for next semester. When things get settled I would love to publish this because it's been a very interesting problem to solve how to present all the different things that can arise in a university classroom. We benefit greatly from having an existing team of course videographers who were previously keeping recordings internal.
today at 1:32 PM
It is an interesting problem! I used to work in academic instructional design and I've gone corporate L+D and that's why my ears perked up.
yesterday at 10:51 PM
> mit got here first
Yes, Claude got the silver medal. Bronze still available. It’s yours ;)
yesterday at 11:29 PM
Ugh. Claude is so much lower level than an organized course. It’s not even a valid comparison because the professors CREATED theirs. Claude … didn’t really.
yesterday at 11:54 PM
will i get some kind of certificate?
today at 4:24 PM
Could always learn via the videos and then use what you've learned to take a certification test.
today at 4:17 AM
For watching youtube?
today at 8:27 AM
today at 1:31 PM
A bit off topic, but I’ll leave it here mostly for visibility: I’m annoyed at how only CS people and faculties are doing this to this extent.
Even for MIT stuff some classes are listed on OCW but not fully (eg: no video recordings).
yesterday at 10:46 PM
That's great the school is doing that?
What about the typical math prerequisites, will they also be available or will online learners be expected to know that stuff, before hand?
Will you offer certifications free or will the cost as well?
yesterday at 11:28 PM
Non-specialized math is well covered elsewhere.
I wouldn’t be surprised to see more advanced math courses appear. I have gotten a lot of mileage out of the already-published lecture notes from individual CMU professors. (Eg on the foundations needed for statistical learning.)
Math is not going to be an “easy entry level” topic because of the high level of preparation of incoming students at CMU.
today at 2:56 AM
I was hoping there would be a curriculum with the math integrated, without the need to take those subjects somewhere else, separately.
But I know most learning institutions tend not to deviate from the traditional "Modus Operandi" :(
yesterday at 10:29 PM
Have you ensured you don't make the mistake UC Berkeley did - not providing appropriate accessibility (captions, etc) as required by Federal Law?
Edit: Ouch. It seems I'm being downvoted for being the messenger. Whatever your views of the Berkeley lawsuit and resulting takedown, the Federal law hasn't changed, and any university not complying with the law can end up with the same fate.
today at 2:01 AM
Every one of our lectures has and will have state of the art captions. More broadly, accessibility is a major focus for us.
today at 1:57 AM
The University of California is a public university.
Carnegie Mellon University is not.
Title II only applies to public entities.
Hope this helps.
today at 4:04 AM
Any institution that receives federal grants (including student loans), private or public, has to conform.
That's Section 504 (overlaps with Title II). See https://www.ed.gov/laws-and-policy/civil-rights-laws/disabil...
> Section 504 is a federal law that prohibits any entity that receives federal financial assistance (such as grants or student loans) from discriminating against persons with disabilities.
> Title II of the Americans with Disabilities Act is a federal law that prohibits state and local governments (such as public school districts, public colleges and universities, and public libraries) from discriminating against persons with disabilities.
> In general, Section 504 and Title II nondiscrimination standards are the same, and in general, actions that violate Section 504 also violate Title II. However, where Title II requirements exceed Section 504 requirements, public school districts, colleges and universities, and libraries must also comply with the Title II requirements.
I worked part time for over a year for the disability division of my university. My boss's job was to assist professors in conforming when they provided online material (web sites, PDFs, etc), and also to ensure 3rd party providers (e.g. Blackboard) are in compliance. Nagging professors about this was part of his job.
He and his team worked with other universities and industry to come up with accessibility stuff for new technologies used in education.
This isn't limited to education either. If you're a private company receiving federal grants for research, you have to conform. Back in the late 00's, they had a whole accessibility themed PyCon where this point was being hammered home (i.e. people coding in Django, etc need to be aware of web accessibility if they ever want government money).
today at 6:02 AM
> Have you ensured you don't make the mistake UC Berkeley did - not providing appropriate accessibility (captions, etc) as required by Federal Law?
Wow, that's actually unbelievable, I can't believe I didn't know about that. Guess I shouldn't be surprised, various departments in government are absolutely chock full of K0-level reasoners incapable of observing second-order effects. "Like King Midas in reverse", so goes the quote.
And they wonder why universities are shuttering one after another at record speeds, as they get pincered by low enrollment and simultaneously cripplingly high overhead.
Perhaps doesn't apply to UCB and CMU, but it's impossible to know the chilling effect that situation had on other universities that otherwise would have published even more content for free online, for our public benefit.
This is why we can't have nice things.
yesterday at 10:27 PM
I watch CMU database Andy Pavlo's videos all the time. Prbly watched all of them . Big part of my career choice to work in databases.
today at 6:05 AM
These are so insanely good. And unlike a lot of CS video channels they do a great job of bringing on the truly most interesting projects in the field like Datafusion, DuckDB, DBSP and tying them to the curriculum
today at 9:05 AM
[flagged]
today at 3:41 AM
[dead]
today at 3:41 AM
[flagged]
yesterday at 11:07 PM
[flagged]
yesterday at 10:02 PM
Hi Phillip,
In the age of LLMs and Multi-modal models, do you think students across the world will benefit from this? If so, how do you think they will benefit.
For some context, when others released courses in youtube/coursera/udacity decades ago, many students, across the world, did not have any real access good teachers and qualified lecturers who could explain the concepts well. Those releases were boon to them.
Now, we learn by asking AI against anything, include youtube content and AI has been best teacher so far.
yesterday at 10:10 PM
On a long enough timeline, we are all dead. In the meantime, come check out Graham Neubig and Daniel Fried teach AI agents, and you will see that we have value still.
yesterday at 10:15 PM
> In the age of LLMs and Multi-modal models, do you think students across the world will benefit from this?
Even if your premise is 100% correct, releasing this will still help, because LLMs will ingest it or use it as a source and provide better answers.
today at 5:00 PM
LLMs are so sterile, I would never use them for primary learning. It would be joyless and require too much effort on part of student. They're no substitute for lectures and a good teacher. They're great for supplementing the learning material.
yesterday at 10:50 PM
These videos, themselves, serve as both grist and ground-truth for inference and agentic interaction.
yesterday at 10:05 PM
This is just absolute nonsense and not universal at all.
Just yday I tried testing each frontier llm on questions w.r.t deferred tax and it made up stuff.
Fundamentals should never be learned via llm’s.
yesterday at 10:17 PM
> I tried testing each frontier llm on questions w.r.t deferred tax and it made up stuff
This would be a stronger example if it were related to Comp Sci. It’s entirely possible that an LLM would perform better on CS fundamentals than details about taxes.
yesterday at 10:11 PM
Which model did you ask it to?
yesterday at 10:55 PM
+1. It's surprising that people aren't familiar with hallucinations, even today?
today at 9:01 AM
This is the same energy as the “Wikipedia is not a valid source” crowd had back in the 2000s. It just gets in the way of learning.
yesterday at 10:11 PM
You're going to get flagged for asking - it does seem to be in bad taste IMO - but I do think it is the elephant in the room.
"AI" will teach you anything at a undergrad level - customized to your style.
The real reason you'd go to CMU or MIT is for credentials, or, more honestly, for networking opportunities. That doesn't exist for watching YouTube videos any more than prompting an AI.
yesterday at 10:21 PM
What you're really saying is that we need to do more than just post some YouTube lectures. And I agree! More on that front is coming.
yesterday at 10:39 PM
Hard disagree on the reason to attend a school like CMU or MIT (disclaimer, I’m a CMU MatSci PhD)
Networking is a big value proposition, but the biggest benefit IMO is learning what questions are important and learning how to think independently about them. LLMs are great for learning but I don’t think they fully replace the skills you need to ask important questions. (This perspective is probably slanted towards research degrees but I think it’s at least somewhat transferable to engineering)
yesterday at 11:22 PM
Partially true. It's the rigor of the education, the thinking process taught, the caliber of people you work with / meet and especially the office hours with top faculties at CMU is really what at mold you. Not information.
yesterday at 10:26 PM
"AI" cannot teach a person what they don't know they don't know.
Human curation will always have a place in passing on knowledge for people who desire deep understanding of a topic, particularly young people and novices in a field.
yesterday at 10:38 PM
You can ask ChatGPT for a detailed listing of all the areas within a discipline and it will give a totally decent answer, better than you'll likely find anywhere else. You can ask it to list disciplines that might be useful, that are taught at university, or whatever.
You can even just ask it "help me know what areas/subjects/topics are useful that I wouldn't even know exist"
today at 1:31 AM
Me: Tell me about a data structure I may not know? Suppose I am a software engineer with 10 years experience
Claude: Here’s one that a lot of experienced engineers haven’t run into: the XOR filter, along with its newer variant, the binary fuse filter.
What it does. It answers the same question a Bloom filter does: “is this key possibly in the set, or definitely not?” The difference is that it uses noticeably less memory and answers queries faster. The trade-off is that it’s static. You build it once from a known set of keys, and you can’t add keys afterward.
...
Does that sound like something a developer with 10 years experience may not know?
today at 4:02 PM
I'm not a professional developer, only a scripter. I wouldn't know.
Suppose you're a college student with barely any concept of CS, let alone the capability to know what kinds of questions to kick-start your own education.
Nothing is new in my argument. AI is particularly useful for professionals that already know how to learn, and know how to sniff out BS and second guess the tool.
yesterday at 11:21 PM
[dead]
yesterday at 10:44 PM
And you probably wonder (if you were the wondering type) why so many are elated at the ongoing collapse of LLM empire.
yesterday at 10:31 PM
You have an amazing dataset, handed out for free by a top rated institution like CMU that can be used as the baseline for whatever learning flow you want to do with your AI.
You do understand that data like this is precisely the goldmine that makes AI good at answering general knowledge questions?
today at 6:19 AM
Why would I personally be invested in what information someone else's AI is ingesting? No normal person is developing their own AI that will benefit from this, and lack of content is not what's holding back llms. Your list is the opposite of interesting. Humans should care about the knowledge opportunity and not the ais.
yesterday at 10:26 PM
I'm with you, it's a reasonable question in today's age.
yesterday at 10:08 PM
> Now, we learn by asking AI against anything, include youtube content and AI has been best teacher so far.
I don't know who is we besides you, but was it the AI that told you you're being taught well?
yesterday at 10:39 PM
And who tells you that a youtube video has taught you well?
yesterday at 10:45 PM
My common sense tells me that a college professor is likely to have taught me better than an AI
yesterday at 11:01 PM
Common sense around a lot of topics can be misleading. Especially when human bias is taken into account. It's not at all clear to me that a properly trained LLM would not out perform a professor. Even we are talking about known great teaching professors, not average teachers, the LLM has several advantages. It can tailor curriculum to each individual. It can learn and adapt curriculum on the fly which a prof is unlikely to do. It can adjust testing and grading so that the student is maximally engaged. It never misses a day or has a bad day. It can answer questions thoroughly and with as much detail as the student wishes, taking as much time as the student desires. What are the disadvantages?
today at 1:02 AM
> It can adjust testing and grading so that the student is maximally engaged
This part is not reality.
In order to do this, the AI would need a way of measuring student engagement.
As soon as you have a measure like that, it is going to be gamed.
Somebody needs to set the standards. You can’t rely on an LLM for that because the LLM doesn’t make up a goal.
Even in a very basic non-LLM example like learning vocabulary, the acceptable rate of correct answers (a day or a week after studying) needs to be set externally.
yesterday at 11:10 PM
That it will make up stuff up out of thin air???
And what does properly trained mean, do we know how to "properly train" something with your characteristics?
today at 8:31 AM
My Stats I and II professor (don’t ask me how) spend the first 5 weeks explaining his take on best-practice US train travel, and sharing his experiences with us.
5 weeks!
They’re often just humans.
yesterday at 11:19 PM
General common sense tells everyone that not all teachers teach well, and that not all students can learn well from all teachers who may teach well because style preferences may collide. One may easily calibrate an AI to teach them in their preferred style; not so much a college professor.
today at 12:54 AM
It's more that you can continue prodding chatgpt/whatever. If you are a naturally skeptical person and try to poke holes in ideas by nature, you don't really get to do to your hearts content with a professor.
But, some people aren't skeptical. A professor says X and they just accept it and feel like they learned from a trusted source or authority or whatever.
yesterday at 10:55 PM
That's probably wrong, if you know how to use AI well.
yesterday at 11:10 PM
Sure, just tell it "don't be wrong"
today at 12:50 AM
You appear to have a naive level of trust in the competence of humans.
today at 1:10 AM
You appear to have an extremely naive level of trust in LLMs.
today at 1:30 AM
No, one can just keep prodding, in a way that couldn't be done with a professor. It requires no trust.
today at 1:35 AM
So, prodding an LLM makes it not lie?
today at 2:03 AM
No....just like that doesn't work on humans. But with prodding you can work out when it's wrong or at least when you should be skeptical. People usually call this "critical thinking".
If you go around assuming authorities on subjects are just right, you will be in for a difficult time. If you endlessly prod and show skepticism towards human experts/authorities, you'll piss a lot of people off. Most of the time they won't even engage.
If you are genuinely interested in learning things and have critical thinking skills, sota LLMs are as good as it gets.
today at 2:23 AM
No, just no.
You seem the kind of guy who distrusts professors and experts and used to trust YouTube idiots, and now turned to LLMs.
Professors will occasionally teach something wrong, but you'll also be studying from books, and if you really want you're free to eventually also use LLMs.
Beginning from them, though, is not too smart, and relying exclusively on them is just plain stupid.
LLMs are going to make stuff out of thin air, and can be more and more confident of it the more you "prod" them.
today at 4:53 AM
lol, these videos are on Youtube...
You seem to be struggling to read the thread. Nobody (except you) said trust, just use. Critical thinking is always required. And no one said rely exclusively on LLMs. And "prod" doesn't mean what you seem to think it means. You prod to find holes.
today at 4:47 PM
> lol, these videos are on Youtube
not from idiots
> And no one said rely exclusively on LLMs
It's what you seemed to be arguing; or, at least to use them as the primary education method
> And "prod" doesn't mean what you seem to think it means
Do explain
today at 8:33 AM
yesterday at 11:34 PM
Make no mistakes.
yesterday at 10:30 PM
In this day and age, why learn computer science? I learned to program before there was Google but with AI, someone who doesn't know the difference between an int and a boolean can make a digit thing. Where do you see the importance of big O notation and, say, eating?
yesterday at 11:13 PM
To those asking why this is still valuable in the age of AI?
If you learn the fundamentals properly, then AI will be a power tool for you. If you don’t, then it will replace you.
I don’t mean this in an aggressive “you better get used to it” sense, I realize there’s nuance - but this is the version that fits on a T-shirt.
today at 4:40 AM
AI is gonna replace you even if you "learn the fundamentals properly".
today at 5:46 AM
We are all going to be ai product builders at some point.
today at 9:43 AM
Why do you think that building AI products is a harder job for AI than other software jobs such that the other software jobs are going to be replaced by AI, but the job of building AI products is not replaced by AI?
today at 1:52 PM
There are some aspects of being human that AI doesn't understand as well as humans do, at least for now. Those are the things we can direct AI towards, make it build products we need/want. If you tell AI to come up with something on its own and keep building stuff, it might get stuck creating and solving weird math conjectures or language patterns. At some point it might be able to understand us better than ourselves but there's still that gap for now.
today at 6:15 AM
This trope would end soon. There's a tiny fraction of computer science students who will have "value" in the age of LLMs and they're the ones who actually would work on LLMs/models and stuff very close to those aspects of AI. They have value even today. Rest will be fair game, as they are today.
today at 7:23 AM
I'm very open-minded about entertaining drastic societal changes but I just don't think the facts support this prediction.
1. LLM performance has more or less peaked. Since Fable 5 released in June, every Claude model has been a ~1-2% improvement over the last one (if you go by benchmarks). Fable 5.2 marketing verbiage has shifted to how it's "lower cost per task" which imo marks a regime shift. I would bet money that this time next year, we're not going to have a model that's obviously and significantly better than the best ones we have today.
2. If you hold that this is about as good as LLMs are gonna get, we haven't found ourselves at a point where we can completely remove the software engineer from the seat. A software engineer is still required to create real commercial-quality software. The role hasn't been commodified either - a better software engineer still writes better software.
today at 9:39 AM
I’ll also anecdotally add that I’m busier at work than ever. Tasks can be completed quicker, (some much much quicker, some hardly at all) but there are so many more tasks than there used to be. Features/improvements/optimizations that used to be out of scope because they weren’t top priority, and just weren’t worth the time.
today at 1:15 PM
[dead]
today at 8:50 AM
I find it hard to believe we’ll end up in a permanent state where Terence Tao is crushed but I’m not. I think more scale and better RL will work wonders for LLM performance over the coming 12 months.
today at 11:10 AM
Terence Tao is a robot who solves math puzzles. We built an actual robot that does that now, hence he got crushed. It's not very different from being good at a sport or board game. The space in which they operate is soothingly well-bounded.
Can you imagine LLMs replacing a small company's CEO for a non-trivial period of time (even forgoing physical requirements, i.e. all I/O may happen over email)?
today at 2:11 PM
> Can you imagine LLMs replacing a small company's CEO for a non-trivial period of time
Perhaps not right now. But performance on long-horizon "fuzzy" tasks is growing rapidly (see eg Fable 5.5 performance on remote work benchmarks). Why would that trend suddenly stop now?
(A human CEO wouldn't be very effective over e-mail alone either, but that's a separate issue.)
today at 12:31 PM
>Can you imagine LLMs replacing a small company's CEO for a non-trivial period of time
The moment that LLMs can do a CEO's job better than any human is the moment that money ceases to have meaning.
today at 12:26 PM
Yes, we even have some benchmarks on that (Vending Bench).
today at 10:38 AM
I'm sure people also thought that AGI was right around the corner when Deep Blue beat Kasparov.
LLMs are good at math because of the verification loop which allows them to implement a guided search algorithm. They don't "invent" new math, they explore many different branches applying pre-existing results and see what works.
Which is not an approach you can generalize to other fields (such as writing text which doesn't read like AI slop, or code which is architecturally sound)
today at 5:58 PM
> I'm sure people also thought that AGI was right around the corner when Deep Blue beat Kasparov.
I sure didn’t.
> Which is not an approach you can generalize to other fields (such as writing text which doesn't read like AI slop, or code which is architecturally sound)
Isn’t that just a question of building more / better RL environments? Sure, math is easier than most domains because a proof can be verified for correctness with certainty. But I don’t see why you couldn’t build RL environments that reward models eg for architecturally sound code.
today at 12:28 PM
It would spectacularly coincidental that LLM performance just happens to peak ever so slightly above top human ability.
today at 1:07 PM
Why? It's trained on human input.
today at 5:09 PM
LLMs have not peaked. There is a linear steady growth of capabilities, with the tilt slightly accelerated when o1/deepseek-r1 came out 2 years ago.
You can see this at epoch.ai.
today at 9:47 AM
That has been said over and over for the past ~4 years, with this many resources, the field will keep finding ways around bottlenecks. Even as it stands now, these models are good enough from an economic perspective to automate a huge share of work, which is why everyone is focusing on efficiency now
today at 9:54 AM
Idk about this. I've used almost 10B tokens of Fable 5.2 in the past week (ultra plan, and the 3 free resets they gave) and the output quality is absolutely step change above even Fable 5 or Opus 5.5. It's probably heavily dependent on the domain you work in.
It's actually so jarring to go back to Opus 5.5 and seeing it do all the usual LLM mistakes again.
today at 8:01 AM
How much do you want to bet? And what would be the decision criteria? Any benchmark that you trust?
today at 10:55 AM
There's a tiny fraction of developers who have to worry about silly things like memory allocation operations because that has been automated and we mostly work on a higher level. But it's still important to learn about it nonetheless. The same will happen with AI. It was already true in 1995 that you didn't need a degree to do software, but learning computer science was still valuable.
today at 11:21 AM
Every lab is pushing for rsi.
today at 1:55 PM
That doesn’t mean they will or won’t succeed
today at 6:37 AM
> If you learn the fundamentals properly, then AI will be a power tool for you. If you don’t, then it will replace you.
Those aren't mutually elusive. You could still know the fundamentals and have AI be your power tool and get replaced.
The replacement part is an economics problem. Under basic capitalism: if AI is good enough to get the job done for most people at a cheaper price than you can support yourself, you're getting replaced, no matter how skilled you are.
There are a huge number of smart people currently employed writing software that are also no smart enough to work at the highest, most rarefied levels of it. And they (which is almost certainly you, the reader), have reason to fear.
today at 2:21 PM
learning is also fun!
yesterday at 11:55 PM
except, not really sure undergraduate comp sci is the fundamentals anymore. Lots of comp sci is now replaced by AI, and tricky to know what you need to know to be an effective software creator. Not that doing comp sci is bad and it will develop a deep understanding. There's also an actual track for doing actual computer science (research), but not sure it's the same track needed to create "engineers" (or whatever people want to call themselves now).
today at 1:39 AM
Comp sci was never necessary to become a swe. It's just one path.
today at 4:03 PM
When you review the output of the AI, how do you know whether it's fundamentally correct if you don't know the fundamentals?
today at 5:10 PM
Who cares? If the output is wrong just give it to the agent again
today at 5:37 PM
Yeah I hate to say it but I agree with this sentiment. No amount of mastery over TLBs or cache coherence is actually going to help you these days when you can just explain the problem to your agent
today at 3:07 AM
Why is comp sci not the fundamentals anymore? Maybe missing from the curriculum is how LLMs work.
today at 9:17 AM
I think I'd favor broader education. As I see it, you don't need all the knowledge of compsci but a subset, and a more general science / business education would probably serve you better. Given we now sit more in the middle of creating systems for particular domains.
today at 12:34 PM
CS should just go back to teaching assembly. High level languages are irrelevant at this point.
today at 1:08 AM
Man, I understand this sentiment so much.
today at 2:14 AM
[dead]
today at 1:24 AM
First, thanks to CMU for releasing these courses. That said, I feel this course might be a little late. AI has already gotten much better at math. This year alone, it has resolved the ETH, Unique Games Conjecture, and several other problems in CS.
Many top institutions are already using AI to accelerate their own research. I don't know why, but I can't shake the feeling that I'm already becoming obsolete, just waiting to become completely irrelevant.
Back then, I used to get really excited whenever new courses were released, especially before the AI 2022 era. These days, though, that excitement seems to have faded.
This is just my personal opinion, of course, and I could be wrong.
today at 12:55 PM
AI is a useful tool, but if we outsource all our thinking to it we've made a grave mistake.
"AI can do this, so there's no point in any human ever learning it" is an extremely flawed line of thinking.
today at 1:57 AM
probably, but your comment doesn't address what gp said, which is very much true
today at 1:08 AM
I think the problem is that in the age of AI, it's not very clear that a lot of these basics are going to be useful, and certainly not to the degree that they were. How many web programmers learn assembly?
I think what people aren't seeing is its more like "if you learn AI, then AI will be a power tool for you". Humans are no longer doing the research to speed up anything related to CS development. Telling the AI "oh you should size your chunks at the L2 cache size of the buffers and when you build your asic make sure it can fit 1024 skbufs so your tcp flow.." blah blah blah.
There might be a select few that work on the fundamentals and need to know how to improve things, but the rest of us those fundamentals will be as useful as assembly is to javascript kids of 2015.
That's all done. Now it's just you plug it in and it is going. Your job is to make sure the electricity and connectivity stay active. You monitor the system.
today at 12:43 AM
> If you learn the fundamentals properly, then AI will be a power tool for you. If you don’t, then it will replace you.
I don't get why people keep repeating this as if we have any idea. How do you know AI will be a power tool to those that know the fundamentals? How do you know it won't replace those people anyway? Why is knowing the fundamentals of computer science going to be the thing that keeps people from getting replaced by AI? You're just repeating something that sounds catchy, as if that somehow makes it more true.
today at 1:28 AM
Yea I agree. It's not about the fundamentals. The only thing that will prevent you from being replaced is your ability to provide value. Does knowing the fundamentals make you more valuable right now? Probably. But it's not because you know the fundamentals, alone. It's that, chances are, if you know the fundamentals you can probably provide more value than someone who doesn't. You put in the effort to learn the fundamentals so with that you can probably learn other things faster. You probably have a better work ethic. Etc.
But not necessarily.
today at 1:01 AM
I think it's very true. You can do so much more with a calculator if you understand maths, same with programming and AI.
today at 1:11 AM
And a computer can do way more than you with a calculator, because its a computer.
The same applies to CS fundamentals, and AI. You can be the best calculator jockey on the block and as mathy as you want, but a computer gonna out computer you.
AI is gonna outprogram you, l33tcode is dead.
today at 3:20 AM
CS fundamentals is not leetcode.
today at 5:41 AM
Pretty sure for a lot of these leetcode hard problems you end up having to implement red black trees and weird graph algorithms. What other CS fundamentals are you talking about?
Anyways, you get my point; the AI is more likely to know about and choose a better algorithm / data structure / scaling architecture than you.
Maybe you found a distrbuted project it sucks at today. But I think it'll kick your ass next year, or maybe the year after that. Either way, it is coming for us programmers and we'd best just try to learn to ride the wave. My impression is how well we can automate things or moreso, get the AI to automate things for us and claim responsibility for the success.
It's all managerial skills now.
today at 6:51 AM
For me fundamentals include more than algorithms, for example how a computer/LLM works, history, philosophy, HCI, networking, security just to mention a few things.
today at 4:37 PM
To expand, I’d throw software engineering fundamentals in there. How to build software products on a team, and maintain them. If you don’t have that then it doesn’t matter how much code you can generate, it’ll all be garbage.
Yes product managers and such can build programs using AI. They cannot build software if they don’t understand how to design systems and how to deliver.
That’s not to say that product managers and such are screwed, they can learn. But code jockeys are similarly screwed if they don’t learn the big picture stuff.
today at 12:36 PM
The only skill that is going to have any value is social ability, which is going to suck for the huge percentage of computer people that have autism.
today at 2:01 PM
It's because doing high quality work requires deep understanding. It's that expertise that lets the user prompt accurately enough to hit the right spot in the model's vector space. That model _has_ the expertise, but it manifests only by giving it the right inputs, which differ across tasks.
today at 2:26 PM
Well, the car certainly didn't end the veterinary field; but there are certainly less stables today than in the past. There are also less alchemists (but certainly more chemists.)
I'm sure these lessons will continue to be valuable, to living breathing people, even in the age of AI.
today at 4:37 AM
This! I don't have a formal education in CS but I taught myself how to make websites and apps and have been a SWE for over a decade.
I haven't touched a line of code in over a year but I still find myself reaching for fundamentals to guide LLMs. Just today I asked fable what the status of X was in my project and it shot out some nonsense, so I challenged it and said "did you git pull?" And it admitted it didn't and gave me an up to date answer. This is what frontier intelligence looks like folks.
And other things like knowing when to index or load balance or use canvas for perf... LLMs will always opt for the path of least resistance, you still need to guide it.
today at 6:25 AM
I saw a video a decade ago (Computerphile I think) where an old CS professor was saying that it's useful to learn how to do simple math operations in hexadecimal. Maybe that was useful in the 50 or 60's, but it definitely hasn't been useful for decades.
I fear that learning data structures and algorithms become an equivalent dated advice.
Years for now, Business Insider will do a documentary on "The last programmer keeping an decades old tradition alive" like other videos they've done. E.g.: https://www.youtube.com/watch?v=H3Den05Q_D0 . I hope it's me.
today at 4:44 PM
Did you even look at the courses?
today at 11:09 AM
Did no one actually watch or listen to the "Principles of Computing' lecture before posting it? One microphone near the middle, and when the professor walks away from it, her voice disappears behind echo, audio compression and aliasing artifacts. We completely miss most of a minute before she walks back.
Why isn't she wearing a wireless mic?
today at 1:22 PM
This is CS ... Fixing the audio quality is left as an exercise for the listener.
today at 3:32 PM
As an AV Tech, I've seen lots of speakers say they don't want a wireless mic and then proceed to walk all around the room talking. Then the video guy gets blamed for the poor sound.
today at 11:44 AM
Because it's a typical homebrew project with the quality you'd expect:
> few of our students built tools that automate a lot of the production work, so a small team can get a course out quickly
yesterday at 11:50 PM
CMU Alum here. Really glad to see Tianqi Chen's ML Systems being offered up. I had just started working through their book https://mlc.ai/modern-gpu-programming-for-mlsys/ so it should complement nicely.
Are requests for courses being accepted? :)
If yes, may I humbly request Larry Wasserman's Intermediate Statistics, Dave Andersen's Distributed Systems and Dave Eckhardt's Operating Systems (or whoever is teaching it now).
today at 2:04 AM
We are focused on computing courses (for now) but I will put in the good word to the two Daves ;)
today at 1:35 AM
Very excited to see that 15-259 Probability and Computing will be available soon. That was one of my favorite classes at CMU during my undergrad.
today at 6:39 PM
For those that know, you know... but secretly in the "serious" software engineering industry we have been considering CMU as light years ahead of any other comp sci program. This is really great news.
yesterday at 10:51 PM
I am a massive fan of open courseware, and that alone I believe has fullfilled the promise of the open web more than any other technology. Back in 2019, I was trying to help people find a web dev job, and the single most important thing that I was able to provide for people was accountability. Recently, I took to building https://wiserlearningapp.com for basically turning things like this into a structured learning path with a schedule where you get reminders and write your notes. If it helps anyone, let me know.
today at 12:01 AM
I made a tutoring tool for my son (year 10 / High school) https://mindshift.katoa.dev/ you can seed it with all kinds of things (like taking the transcript from one of these videos + the video) and develop a learning program that adapts to your learning progress. Haven't really made it available to anyone else, but I'm hoping to build up a whole "lab" of tools for teaching any thing. Currently I've primed his account with his syllabus and assessment schedule + NZ road code. It also notifies and keeps a schedule.
today at 5:56 PM
Hey this sounds interesting the way you're describing it but I read your description and I read your landing page and I have no idea how to use your tool or what it actually does still. My suggestion, if you're putting it out there, is to really simplify what you're trying to do or say you do.
yesterday at 11:12 PM
cool tool
yesterday at 11:54 PM
I see that courseware is coming next, that's really important. Lectures are nice, but without the projects and some kind of way to get feedback on it, lectures can fall flat a bit. Videos, lecture slides/notes, projects are needed.
Another thing that happens a lot with these is the CS1 gets posted, but not the graduate compilers course, which can be annoying.
today at 4:04 PM
Yes, as someone who has audited unknown number of courses (8? 10? I almost have a second Bachelor's degree from all the stuff I've audited, it's actually absurd), the courses that had the biggest impact on me were the ones where I actually got hands-on with some extra component. Reading/homeworks/projects, there had to be something where I applied what the lecture showed.
Any course where I just passively watched videos felt like failed endeavors for me, especially because the lack of engagement usually leads to quitting pretty early, like by the 4th or 5th lecture.
today at 6:04 PM
Why would someone want to learn / practice something that they will never use?
today at 1:49 AM
its great people do this, but sometimes i wish other fields would do this.
It feels like if you want to learn CS there is every resource in the world, but if you want to learn history, philosophy, etc the resources seem much more limited.
yesterday at 11:11 PM
What is the licensing?
I believe MIT's is mostly CC BY-NC-SA 4.0?
I never really saw a lot done with that, but I can imagine a near future where people/students start pointing coding agents at them.
(Whatever the licensing is, thank you for doing this.)
yesterday at 11:32 PM
Would be great if there was an open license on them - being "public on YouTube" is not the same as open, which limits the ability to redistribute them to learners in places with little or no Internet connectivity.
today at 10:34 AM
I'm currently following CMU's CS:APP lectures, and I'm really enjoying them. They have helped me understand low-level systems much better. It's great to see more CMU courses being made freely available. Looking forward to checking them out!
today at 5:08 AM
Any content on type theory, formal methods and related software verification topics?
yesterday at 11:11 PM
CMU alum and one-time adjunct professor there. This is wonderful news.!
yesterday at 10:31 PM
I haven’t watched any of these yet but seeing this takes me back to 2008/2009 when I was just learning to Java in high school and my teacher showed us Randy Pausch’s last lecture and the Alice project. That resonated with me for so much of my life. Also Alice was a really cool way to learn programming and probably still great to teach it today.
yesterday at 10:31 PM
Love it! I remember that the OS and the Compiler courses are pretty good.
today at 6:21 PM
amazing
today at 1:57 PM
Would be interesting to know how these courses changed in the last 5 years due to AI developments?
yesterday at 10:18 PM
On the contrary to speculations in the comments, i think this is great and shows integrity. TheY are offering their knowledge to the world. In the meantime we can still appreciate real human knowledge being shared. AI has it's place.
today at 6:08 PM
Just a reminder that CMU has done nothing about the racist professor they still have on staff. This person made statements like "I heard the chief monarch of a thieving raping genocidal empire is finally dying. May her pain be excruciating." "Die first, white man, then ask yourself why I'm dancing." and "Kirkmas" and "Kirk Mubarak" celebrating the assassination of a public figure.
They promised swift action, but have not taken any action in 2 months.
https://www.clevelandjewishnews.com/jns/carnegie-mellon-says...
today at 2:39 AM
Really appreciate the good work. It will silently help so many students. Anyone who keeps learning is a student.
yesterday at 10:39 PM
Awesome stuff. While others have pointed out that the material is introductory, any amount of free material helps enrich the public common knowledge.
yesterday at 10:52 PM
This is awesome, and am very stoked that you are including upper division lectures.
Will the computer engineering courses, 18- be including in the videos?
today at 9:53 AM
"2 unavailable videos are hidden"
Oh yeah, the title is not saying that the computer science courses are complete Eg. https://youtube.com/playlist?list=PLSH_5v_OOfP8
What's up CMU? Teasing and marketing? Or bug?
In any case, if CMU is really going OpenCourseWare path, that would be really good. To be confirmed based on their action.
today at 10:09 AM
Another post here said its getting released one per day
yesterday at 11:04 PM
I highly recommend you check out their database Youtube channel as well. It’s an awesome resource.
today at 11:06 AM
It may well be exacerbated by the problems that LLMs are causing.
today at 4:56 AM
I hope more core CS courses like 15-122, 15-150, 15-151, 15-251, 15-451 are also released.
Also, are assignments and homeworks going to be released or only the videos?
today at 12:40 AM
Is that related to the 10-202 course: Introduction to Modern AI, by Zico Kolter, that was also released on Youtube?
Was this a totally different initiative, or is this building on these experiments?
today at 2:03 AM
We have a lot of faculty who have already start sharing their teaching independently, myself included (Philomath, Bioinformatics Algorithms), and our sense is: the more the merrier; we promote faculty who share their teaching independently, including Zico, on the front page of our YouTube channel.
yesterday at 10:17 PM
I remember back in the late 2000s/early 2010s many many universities started putting much of their courses on youtube for free. But then they were made aware they legally liable in some jurisdictions for not having accessibility (youtube didn't auto-caption back then) and that liability pressured most universities into taking their courses down.
Glad to see things finally starting to get back to normal.
today at 2:22 AM
Andy Pavlou’s DB courses are amazing
yesterday at 9:18 PM
Associated news post: https://www.cmu.edu/news/stories/archives/2026/november/cmu-...
today at 11:43 AM
Is there a single university that doesn't fail the rather primitive basics in its published courses?
For example, the audio quality here is pretty bad, and watching a tiny person in a far away corner of the room doesn't help. Surprise, it was never really designed to be a video course, just another poor copy of the classroom experience.
today at 3:11 PM
Open University, UK.
today at 2:50 PM
Great - but .. man 2026 - a bit late to the game. MIT OpenCourseWare went online in 2001!
today at 2:16 PM
Maybe future versions of Sol will finally learn the fundamentals.
today at 11:05 AM
About a decade too late but uh thanks I guess. We'll feed AI agents off it and hopefully they'll be less stupid.
today at 4:38 AM
Didn't they do this already about 13-15 years ago?
today at 6:41 AM
If you could live inside any fictional universe , which one would you choose?
yesterday at 11:16 PM
CMU grad, here. Glad that you guys are doing that!
today at 12:14 PM
This is commendable but remember that CMU also unleashed CMMI straight out of the Mount Doom of Corporate-Industrial-Complex ;)
today at 4:16 AM
Well, the funny thing is that people don't want to pay for the degrees anyway. Why would they when there's no entry roles?
today at 6:48 AM
This is fantastic.
today at 3:36 AM
putting for free is a good thing , they want views to earn money
today at 2:03 AM
me <-- old enough to remember how coursera and edx were going to democratize learning with free high quality uni content -- which they did for a while (took some courses myself; good stuff), before both sites devolved into crap profit-seeking machines
today at 4:55 AM
and udacity
today at 3:38 AM
When i can't see it? Is it better for someone looking to first time learning it
today at 8:18 AM
What’s the incentive for making courses available for free? What’s the catch?
“Free” is a load-bearing word these days.
today at 11:45 AM
No longer profitable to sell seats, and likely no longer as lucrative to invest peoples time into learning.
Perhaps unfair, but they seem to imply "Free" as in worthless in the context of LLM stealing $9Tn in FOSS for isomorphic plagiarism code generators. =3
today at 2:53 AM
Good, my Agent needs to polish up on some of its fundamentals.
today at 9:36 AM
A bit late now eh?
today at 12:05 AM
[flagged]
today at 9:43 AM
[flagged]
today at 8:27 AM
[dead]
today at 5:06 AM
[flagged]
today at 9:59 AM
[dead]
today at 1:22 AM
[dead]
today at 10:23 AM
[dead]
today at 12:28 PM
[dead]
yesterday at 10:38 PM
its a shame a CS education is so undervalued these days you can only give it away /s
today at 3:29 AM
[dead]
today at 8:29 AM
I wouldn't get too excited:
https://en.wikipedia.org/wiki/Massive_open_online_course#Cri...
https://en.wikipedia.org/wiki/Learn_to_Code
today at 1:22 AM
Why would anyone bother with that? How long till there is no “App Store”? In less than 24 months you’ll have an iPhone with Siri and nothing else.
today at 1:30 AM
This probably bait, but no way. Chat is not the best interface for everything.
today at 5:31 AM
I don’t know. My little bots run more and more of my apps. At what point is it not easier for them to handle the data and do the work? Office documents, project management, contacts, keeping track of social media, email.
today at 1:48 PM
Yea but sometimes you just want to browse, seek, explore -- think browsing a shopping catalog, exploring Google Earth, chatting on reddit/HN, etc.
today at 2:13 PM
Everybody is gonna have their own project management and office document tool? Seems unlikely, unless you work alone
today at 12:38 AM
Only 2 decades late. Welcome to irrelevance
yesterday at 11:18 PM
Very fair of them to post this for free, given that that seems to be approximately the value of a computer science education in 2026
(/s(?))
today at 12:11 AM
You must be new to this whole thing. We’ve had computer science courses posted for free for a long time from top universities like MIT.
today at 1:02 AM
Carnegie Mellon as well, they have been posting some courses for sometime, I don't know what's new with this initiative.
today at 1:07 AM
You missed the joke. It was even tagged /s.
today at 6:33 AM
Nah
(/s(?))
today at 9:49 AM
I'm not new. I'm making a joke. Do try and keep up.
today at 10:10 AM
Yeah I get you think you were. The trick is to be confident about it.
(/s(?))
today at 2:38 PM
It's like guys who try to pick up women, but are "joking" when they get rejected. real loser energy.