Your CS Degree Is Teaching You the Wrong Things

AI is changing how we write software, but what does that mean for programming languages and computer science education? The crew debates Python vs. C, whether language aesthetics still matter, what CS degrees should actually teach, and whether low-level knowledge will become even more valuable in the age of AI.

• 0:00Speaker 00

  There's a thing that happens that's happening around here. I don't know if you guys have seen this, but like anywhere that has anywhere like a food truck or like in a food court or a place that has picture menus.

• 0:15Speaker 01

  Hmm.

• 0:15Speaker 00

  The food is AI generated now. And it's like not, it's, it's weird because it doesn't look, it doesn't look convincing. Like they didn't use, I think we have like, I don't know what the state of the art is, but we have AIs currently, I know we do have AIs that can generate convincing pictures of food that wouldn't look like AI art, but they're not using that.

• 0:39Speaker 02

  Frontier models. They're not using frontier models.

• 0:41Speaker 00

  They're not, they're using, I don't know what they're using, but they're using something.

• 0:44Speaker 01

  I know what it is.

• 0:45Speaker 00

  Okay, what is it?

• 0:46Speaker 01

  No, no, it's, it's just that they're literally, I'm pretty sure it's not that they're constructing individual pieces of the photos through like a good image and model. They're going to chat GPT free because they don't know that there's a difference.

• 1:01Speaker 00

  Okay, yeah, yeah.

• 1:02Speaker 01

  And they paste in their entire menu for the page. And they say, make a menu page for this. And it's like, cool. You have pancakes? Here's, here's pancakes with blueberries.

• 1:13Speaker 02

  The real world uses AI completely differently than.

• 1:16Speaker 00

  So it's gonna be, it's like, it's like free tier pancakes is what you got. You got, this is the free tier menu.

• 1:27Speaker 01

  Uh, anyways, I'm sorry.

• 1:28Speaker 02

  Speaking of food, check this out, guys. So Big Z hooked me up with these Krispy Kreme.

• 1:33Speaker 00

  Yes.

• 1:34Speaker 02

  Pokemon donut holes. Hold on, we're not done yet.

• 1:37Speaker 01

  Wait, wait.

• 1:38Speaker 02

  I'm pretty sure she's trying to give me the beaters.

• 1:39Speaker 01

  All together. You said Krispy Kreme, Pokemon donut holes.

• 1:44Speaker 02

  Yep, exactly.

• 1:45Speaker 01

  Trash, are you gonna save those? Are you gonna put those on your shelf and kind of keep them as souvenirs, you know?

• 1:50Speaker 00

  He's gonna eat them all before his kids get a chance.

• 1:52Speaker 02

  Case for these ones. I'm gonna share these with my kids and we're gonna destroy these. I'm gonna try them on, on while we're doing our thing.

• 1:58Speaker 00

  Immediately after we get off of this podcast, I'm going to get donuts.

• 2:01Speaker 02

  Crack open a Poké Ball? Like, is it stuffed? Do you think it's stuffed?

• 2:04Speaker 00

  We can send you some for next one.

• 2:05Speaker 01

  Just crack it up and take a couple bites out of it. Let's see, what, what is it like? Give us a little Poké rating here.

• 2:15Speaker 02

  So vanilla icing. Can't tell what the sprinkles are. Chocolate strip in the middle is a good little touch. Give it an eight out of 10.

• 2:23Speaker 00

  Nice.

• 2:24Speaker 01

  Eight out of 10.

• 2:25Speaker 00

  Yes.

• 2:26Speaker 02

  There you go. Thank you, Big Z. Shout out, shout out Randy at Krispy Kreme. That made it happen because apparently it was a whole deal trying to get me these.

• 2:35Speaker 02

  Thank you, Randy.

• 2:36Speaker 01

  Thank you, Randy. Thank you.

• 2:38Speaker 00

  Hey, leave a comment, everyone, saying, thanks, Randy.

• 2:41Speaker 02

  Randy from Krispy Kreme. Well, that's actually really good.

• 2:47Speaker 00

  Man, we're ordering some for everybody next week.

• 2:51Speaker 01

  I would like to say that normally what I end up doing, just generally speaking on the standup, I like to jump into my main scene and we usually, you know, grab something off the old Kanban board, right? And we kind of talk about it, but I feel like that's, that's not really, you know, the future. And so instead I think what I should do is open Cursor Agent and go like this. Hey, uh, could you check Linear and see if, uh, the standup has any blockers and, uh, list them out for me. There we go. Because I feel like that's like, that's kind of what the future is. You know what I mean? Like, do you really want to go and like type in a whole web address? Like, think about how long that could take.

• 3:36Speaker 00

  Especially when there's native apps for Linear that are really good. Hashtag add, but that is what I use.

• 3:42Speaker 01

  Add native apps are really good. Like, imagine having to like easily navigate and open up a page and read it right away when your model could do it for you.

• 3:51Speaker 02

  UI is dead. UI is dead. I'm telling you.

• 3:54Speaker 00

  UI is dead.

• 3:55Speaker 02

  UI is dead. Like, I'm not even kidding, dude. UI is dying, man. I'm telling you.

• 3:58Speaker 00

  Dude, UI died about 15 years ago when web designers took over.

• 4:04Speaker 00

  There's like no scroll bars and no visible, like you, the screen is just a big white page and no one has any idea what they can click on. Three, everything is just three little lines at somewhere that you're supposed to know. Is, are the, do the three lines mean I can drag this item? Sometimes. Do the three lines mean it's a menu? Sometimes.

• 4:20Speaker 02

  The hamburger. The hamburger. Oh.

• 4:24Speaker 01

  Yeah, we call that a hamburger.

• 4:24Speaker 00

  It doesn't look like a hamburger either. It's called a hamburger menu, but then, but if you turn it on its side, it's a grab bar. It's a thing that you can move up and down. Then it's not a hamburger menu anymore. 90 degree rotation, no longer hamburger.

• 4:36Speaker 01

  Great point, Trash. That's true. Hot dog or hamburger?

• 4:38Speaker 00

  Yeah, yeah.

• 4:39Speaker 01

  Really?

• 4:39Speaker 00

  Yeah.

• 4:40Speaker 01

  Completely different meanings there. Hey, is that HTTP? Get that out of here. That's not how we order coffee. We order coffee via SSH terminal dot shop. Yeah, you want a real experience? You want real coffee? You want awesome subscriptions so you never have to remember again? Oh, you want exclusive blends with exclusive coffee and exclusive content? Then check out Kron. You don't know what SSH is? Well, maybe the coffee's not for you.

• 5:16Speaker 00

  I like how you took Linear, which is like, I thought the whole point of Linear was that it was supposed to be like faster than Jira or something like this. And then you, you, you strapped, or so, like you strapped a huge like, uh, um, uh, delay on it.

• 5:31Speaker 01

  Yeah, that, that was the joke, Casey. Thanks for saying the joke out loud, which is that I use the world's most inefficient way to look at a list. Okay, there we go. Okay, thanks, Casey.

• 5:42Speaker 02

  How much did that cost you?

• 5:44Speaker 01

  Uh, how much did that cost you? Well, I'm using Grok 465 fast and that was only like 40,000 tokens.

• 5:49Speaker 00

  465 from inventing model, Dave.

• 5:51Speaker 02

  I was like, oh, it's a 465?

• 5:52Speaker 00

  465.

• 5:53Speaker 01

  Yeah, it's, it's a midpoint. Actually, TJCTJ is not a part of it. It's a like secret beta test where we get like halfway to four seven.

• 5:59Speaker 00

  That's not how decimals work. 465 would be 65.

• 6:06Speaker 02

  Got him. True, but it sounds cooler when you say 65. But also the name is obnoxious. Hi, fast? That's insane.

• 6:11Speaker 00

  You fell for the classic blunder of Semver.

• 6:15Speaker 01

  It really is. All right. Well, anyways.

• 6:19Speaker 01

  I don't really care. So let's just go on here. I want to talk about something. This is like a little bit of a blocker. Okay? So, hey, everybody, welcome to the standup. Today we have quite the show with you. We have one incredible blocker by one incredible programmer. And then second, we have Casey deeply regretting his life decisions. And then third, we have the main course, Python. Is it not good for education? Is Python not good for anything? Is C really how you actually understand computers? Or is that all a farce? Anyways, but before we get to that, we need to do a little bit of a blocker. Okay? I saw this tweet earlier and I don't know how I feel about it. I like Zach. Zach overflow, cool.

• 6:59Speaker 01

  Uh, you know, handle. He says, basically, anyone who says they dislike a programming language for the aesthetic syntax reasons, e.g. Rust, needs to just update their taste in languages. Is this true? Should you just always accept a language as it's looks or is the aesthetics of it important? Now I'm gonna throw something out there.

• 7:21Speaker 01

  I like how Lua looks.

• 7:23Speaker 00

  Wow.

• 7:24Speaker 01

  I like it.

• 7:25Speaker 02

  I thought you were gonna keep going there. I thought you were gonna program it.

• 7:27Speaker 00

  Yeah, like can you, I, I, was that gonna get dirty? Like, it sounded like you were going somewhere with that. It's like, I like how Lu, let me tell you about Lua. I got some things to say.

• 7:35Speaker 01

  I will tell you about Lua. I like when we use words to end things. Okay? I like end to say, hey, this is the end. If, end. I don't like squirrely braces, you know? Like, I don't think squirrely braces are nearly as, uh, beautiful as nice keywords. Okay? I like, I prefer keywords as openers and closers.

• 7:55Speaker 00

  How do you feel about the old school? We use the word written backwards to end the thing that it started with, prime.

• 8:03Speaker 01

  Oh, yeah. You're talking about the old case ESAC thing. Yeah, I know about that ESAC.

• 8:08Speaker 00

  Or if, if and phi. I don't know if you've ever seen that one. These are, these were common. These were, these were a thing.

• 8:14Speaker 01

  Vimscript, in fact, even has that one, if I'm not mistaken. Vimscript.

• 8:17Speaker 00

  No, Vim has if and then end if.

• 8:20Speaker 01

  No, no, the old, the old one. Isn't it FI? Or is it, is it?

• 8:24Speaker 02

  I hate that.

• 8:25Speaker 00

  If and then end if. Trust me, I've written.

• 8:26Speaker 01

  I know Bash does this.

• 8:27Speaker 00

  Tens or hundreds of thousands of lines of Vim.

• 8:29Speaker 02

  That's why it's cursed.

• 8:31Speaker 01

  Okay. By hand. I, I don't like that one. I'd rather see the word end.

• 8:35Speaker 00

  Okay.

• 8:35Speaker 01

  But I will take a backwards spelling if, if I must need to, uh, if I must. Okay. I would, I would take that. I prefer it over squirrely's. And here's my reason why. I hate that if statements can and cannot have squirrely's on them. Okay? I just want a defined syntax at all points and I just want an expectation that's always met.

• 8:54Speaker 00

  Well, that's a different complaint.

• 8:55Speaker 01

  I will give Odin a pass because Odin does have a do statement. So does if do and then it's allowed to be on one line. So I like that. I'll give that. But, uh, I think, I think aesthetics are important. I don't know. I think there's something that, I know we're leaving the day and age of caring about these things and most people will just not even know what programming looks like in probably the next five years. But I think there is something about the beauty of a language. Okay? I think there's something about it.

• 9:19Speaker 03

  Can we go back though quickly? Because if statements requiring you could have a language where you have to have end as the delimiter, but then still have single line ifs without an end. And that was like, that's so your complaint is more about-

• 9:33Speaker 01

  With an end on it. You have to have an end. I like, I just-

• 9:36Speaker 03

  But there's no reason that like C has to allow you to have an if statement that doesn't have curly braces afterwards. It's just that C chose that as an option.

• 9:45Speaker 01

  Yeah, and I think that's even worse.

• 9:47Speaker 03

  Okay. I was trying to figure out-

• 9:48Speaker 01

  That's what I'm saying. In fact, it prevents it from being an LL1 parser, right? Wasn't that one of the big design goals of the very first version of Java? And then they fumbled the bag because they were like, well, we need to have if statements without curly braces. And then the dangling else problem and then everything got awful.

• 10:02Speaker 02

  This is how lifters are born from this conversation.

• 10:05Speaker 03

  True.

• 10:06Speaker 01

  Okay.

• 10:06Speaker 00

  Do you like semi colons, Prime?

• 10:09Speaker 01

  Uh, no.

• 10:10Speaker 00

  Okay.

• 10:11Speaker 03

  Yes.

• 10:12Speaker 02

  I like semi colons. What?

• 10:14Speaker 03

  I like them.

• 10:15Speaker 01

  Why?

• 10:16Speaker 03

  It just makes it more obvious when an expression ends and where you expect it. You get better error messages from them. You don't have semi colons and you have a, you have a like trailing plus. It's unclear where you expect the error to actually show up in general. I think like it's easier to write parsers and like language tooling that says like, let x equals y plus one. If you, if you just do let x equals y plus, and then there's a semicolon at the end of the line, then you can know something's there or that it's missing. That's like obviously stupid example, but like terminators tend to help you get like error messages to the spot. Like this is one of OCamel's failings in my mind is that if you don't have, you, OCamel doesn't have terminators for a bunch of stuff.

• 10:59Speaker 03

  So you end up sometimes getting error messages that are like 100 lines away from the actual problem. Cause you're like missing an in like way before.

• 11:07Speaker 00

  Oh, okay. Let's, let's slow down here a second.

• 11:10Speaker 03

  Yep.

• 11:12Speaker 00

  Now, did I just hear you say that OCamel has a problem?

• 11:18Speaker 03

  Yeah.

• 11:18Speaker 00

  I know, I know you didn't say that because I'm pretty sure that OCamel is a perfect language that doesn't have any problems, Tej. So maybe you can rephrase that.

• 11:29Speaker 02

  No, that's, I think I could say that now because I started making my own language. So now, you know, it was like, OCamel was the best girl at my high school, right? But now we're at college, okay, Casey?

• 11:40Speaker 00

  Okay, alright, okay. Yeah. Got it.

• 11:43Speaker 02

  Oh my goodness.

• 11:44Speaker 00

  Understood.

• 11:44Speaker 01

  TJ, didn't you marry your high school sweetheart? Isn't that phrase just completely nonsense to you?

• 11:49Speaker 03

  She was still the best. That's what I'm saying. For me, it's true, but not for everybody. Chats never dated anybody, so they don't understand. They're still just admiring from afar, like in each of the different scenarios.

• 12:00Speaker 01

  Are they going to SF Hot or Not, whatever it's called, SF Hot List? That's-

• 12:03Speaker 02

  I do not want to talk about that.

• 12:05Speaker 03

  They just made that, Casey, I know you didn't see it. They made a hot list for SF and so far I think there's like four people on the list.

• 12:13Speaker 02

  I can't believe I'm not on it.

• 12:15Speaker 00

  For the city?

• 12:15Speaker 01

  San Francisco has finally found the end of innovation, which is they now just go, who's hot here?

• 12:23Speaker 00

  Like, is it because I need someone else to tell me? Like, what does that even mean?

• 12:26Speaker 02

  They got baited to move there and now they're just trying to find a mate now.

• 12:29Speaker 00

  Alright, yeah. I hope you like talking about AI and, and series A funding rounds.

• 12:37Speaker 01

  Alright, well, that was awesome. I actually, I was just actually curious if someone actually had a real take on, on uh, if, if aesthetics is important at all to syntax, but it seems like nobody thinks that. So we're just gonna move on.

• 12:49Speaker 03

  You didn't even give us a chance to actually talk about it. I was trying to understand your point.

• 12:52Speaker 02

  Yeah. Well, my, my actual opinion is, is like you kind of hinted towards this, hinted toward this earlier was in the age of AI, I feel like it's slowly just, it just doesn't matter as much anymore. Obviously like for this group and like other hardcore engineers like that really like, like to dive into that stuff. I think it still matters, but I think for like the general public, I don't think it's gonna be a thing. And I made a tweet like not too long ago where like DX doesn't matter because my, as long as my agent can work with it, fine, like have at it. And I feel it's an interesting time and it's kind of depressing to be honest.

• 13:25Speaker 01

  One small thing, which is that how I design my CLI arguments now no longer favors humans using it, right? I realize that like, okay, agents are super good at like just making a kajillion arguments. You can make, you know, you can, you design your interfaces a touch differently because you're just going, okay, an agent's gonna use this thing. And so it is, it is different. Like I would never ask a human to do the thing that I was doing, but I'm like, agent though? Agent loves this. They're gonna, it's gonna do so good.

• 13:51Speaker 03

  I was gonna say though, I kind of feel the opposite trash because since agents are gonna write it and they're trained, this is like my actual opinion and not just trolling.

• 14:02Speaker 03

  That because we train agents on like human corpus, right? Like we've trained it on all of the history of humanity. I think things that work better for people and communicate better people to people, agents will be able to do better. I, as since they're like human approximator functions, as Prime said in a recent video, which I like, it's like what they're trying to be, that it's successful or not. So I think like if in a language can express things like efficiently and effectively and elegantly, clearly, agents will do better with that thing. So whether that means the aesthetics of like where curly braces go, I don't really think that part matters so much. Uh, like I think the tweet was inspired by some people complaining about like Rust having

• 14:46Speaker 03

  a lot of colons and like brackets and stuff, but I think that that's actually like good aesthetics because it communicates a lot of information to me like clearly and concisely in the same way. Like there's a lot, every time someone does a screenshot of Rust with like 30 different things going on, it's like, yeah, it's from some crazy standard library function that someone spent a lot of time to make sure works for like every single person in the world in like a million different situations. Like most of the time you don't have to do that. I think OCamel beats Rust on that point. So there you go, Casey. You just have to really throw, we had to make it go back. But I'm saying like, I think uh, that it will still matter even when agents are writing it because tokens will still cost money and things matter. Like it's still, it's in the end, it's not that everything's free.

• 15:31Speaker 03

  Like we still have like supply and demand and like things going on and it matters if my DX takes 100 times as long because someone chose to write this in like the world's slowest language or whatever, instead of making it fast and the agent can complete in half the time, those things will still matter, I think at least.

• 15:48Speaker 00

  One would also assume that uh, to the earlier point you were making where it was like, okay, I've got this language and if I accidentally omit one symbol from it, the error message is from like, you know, 130 lines down. I feel like you don't really want your agent to have to consume those extra 130 lines to get the error and then it has to then spend reasoning time to like go figure out where the error was and then it has to go back and patch. It's like, if it could have just been like, here is literally where the error is, then that would be a trivial update that costs basically nothing, like no electricity at all. Uh, so it does, it also kind of seems like you do care how well, because you know, there's,

• 16:31Speaker 00

  there's just this, as long as you're still running these kinds of agent cycles on things, it seems like you, you would care how, how efficiently the language can, can find problems.

• 16:42Speaker 01

  But that's also true without, without agent, right? If you have to read a, like if you read a C++ template error that sometimes span for, for an actual hour of reading, it's like, I don't, I don't even know what's happening here.

• 16:54Speaker 00

  I, I'm assuming that it, I'm assuming that it would be translated similarly where it's like, yeah, you're just, you, yes, you don't have to do that now, but you're going to pay more money or more server time or whatever it is to have this done for you than if the language just didn't have that problem or whatever the AI is outputting was designed intelligently so that it minimizes the amount of work that the reasoning layer has to do to put the thing in order, whatever that thing is, right?

• 17:20Speaker 03

  So there you go, Prime. Is that what you were looking for?

• 17:23Speaker 02

  Boom, headshot.

• 17:24Speaker 01

  Yeah, something, something for you guys. Okay, so let's go on to the main course because I saw this little spat happen on the old internet, as they say, about a, a Google researcher, Lori Wired, for those that are unfamiliar with her game, she is a very incredible YouTuber, effectively says that CS degrees are kind of fumbling the bag because they're starting students on Python as opposed to C or C++ and thus it's driving more and more people going in to learn about computer science further and further away from the computer. And this is really kind of doing them uh, dirty. If, if we will say, uh, it is not allowing students to excel the best or come out of college really knowing how computers work.

• 18:06Speaker 01

  Instead, they know how say Python works, which is a further abstraction of the tree. And then this ignited an entire series of arguments of people saying like, uh, most C and CPP coders don't even know a lambda calculus or formal languages are the bedrock of computer science.

• 18:23Speaker 03

  Good point, actually.

• 18:23Speaker 01

  And people started arguing over what the heck is computer science? Do we even, does anybody even know what a true computer scientist is? There is no true computer scientist that walked this, this earth. Ryan Fleury, of course, was making fun of some things. There was someone that said something uh, kind of making fun of the whole computer science and C being a, a better abstraction, saying C is not actually representing what's happening on a computer. And then of course, Casey, you also had a beautiful tweet. Vaxry was just like, I don't even know, I, Vaxry seems like, he's like, I'm not even sure what CS is, but I think it's pretty much math proof and formal logic. And so none of this makes any sense. So C really doesn't help at all. And so it just kind of turned into everybody arguing about things. So I figured we could break this entire thing up into three steps.

• 19:06Speaker 01

  The first one has to be, what is the chief goal of a CS degree? Like what actually is, is the person walking out? Is it to become a good programmer? Is it for you to be able to just do the pigeonhole theorem just at any moment on the whiteboard? Or is it for you to practically understand how a computer operates? And I think those are three very different, you know, positions. And so I, I feel like Casey, you were the most excited and also got like a ragingly uh, or a decently tech viral um, tweet. Therefore you are required to answer this first.

• 19:41Speaker 00

  Well, I, I would like to point out that my, my tweet, the incredible content of my tweet-

• 19:48Speaker 03

  Legendary tweet, honestly.

• 19:49Speaker 01

  Legendary tweets, Casey, Maria, Tori.

• 19:52Speaker 00

  Filled with wisdom was, and I quote, 100% yes, thanks Lori Wired.

• 20:00Speaker 03

  Wow, Casey, great.

• 20:01Speaker 00

  That was it. That was all I said.

• 20:03Speaker 02

  Very exciting.

• 20:04Speaker 00

  So, so.

• 20:05Speaker 01

  He's getting Twitter, everybody. Casey's getting Twitter.

• 20:08Speaker 02

  I kind of knew everything you were gonna say because I actually listened to your podcast on pragmatic engineer. I think that's the one bit.

• 20:14Speaker 00

  Okay.

• 20:14Speaker 02

  And then I saw this clip and I was like, oh, this is like exactly where you're basically talking about that whole episode.

• 20:19Speaker 00

  Okay, yes. I mean, yeah, effectively. So I was just trying to say, what, this actually surprised me quite a bit, the reaction to this. I mean, I guess I shouldn't be at this point, but I kind of was because at least when I, when I saw that, because the reason I saw it was because that, I'm not sure, MTS something, the original post, I'm unfamiliar with.

• 20:43Speaker 01

  MTS?

• 20:44Speaker 00

  Yeah.

• 20:44Speaker 03

  Yeah, monitoring the situation.

• 20:46Speaker 00

  I'm unfamiliar with that source. I would never have seen, I would have not even have known that this occurred, except that it was in the trending, like it was in the Twitter trending things. And it said like, it didn't say Lori Wired, it said like, Google researcher, you know, says Python. I was like, oh, that's interesting. So I clicked on it. And I saw that the MTS, I don't know what this is. I'm like, oh, it's Lori Wired. Like, okay, I know who that is. I listened to the clip.

• 21:13Speaker 03

  You're not supposed to do that, Casey. Just as a quick heads up.

• 21:16Speaker 00

  Good point.

• 21:17Speaker 03

  Don't do that next time.

• 21:18Speaker 00

  You don't, don't, don't actually, this is comment. You just gotta comment as fast as possible.

• 21:22Speaker 03

  Thank you.

• 21:22Speaker 01

  By the way, Casey, if you look right here, I'm actually still getting it as trending. Google researcher questions Python shift in CS programs.

• 21:28Speaker 00

  And so, so, so I listened and I thought that, that Lori Wired's, like what she says is actually very diplomatic. Like, she did not phrase it in an incendiary way. She basically described quite literally like what the issues are, why she was concerned about it. And if you listened to it with any degree of like, understanding, you, you would understand that she, she never even says like, Python is bad. She said, she even says like, Python and Java are, are, I think she even says they're great things. Like, you know, she doesn't have a problem with them. She's just saying that you kind of need to know this other stuff. Like, we need to be able to teach people memory management.

• 22:10Speaker 00

  And it's, and she says like, it's easier to do that if you already know something like C or C++ because it's easier to extend downward. Like, she made the entire, yeah, exactly. Yes.

• 22:20Speaker 01

  This is what you're describing to the T.

• 22:22Speaker 03

  It's a classic, an all-timer, really, an all-timer tweet.

• 22:25Speaker 01

  It really is an all-timer.

• 22:26Speaker 00

  Cooley Bravo, yes.

• 22:29Speaker 03

  Thank you, Cooley Bravo, yeah.

• 22:30Speaker 00

  So I thought it was presented very diplomatically. And so I didn't, I didn't know there was, I didn't even know people were complaining about it yet. So when I said 100% yes, thanks Lori Wired, I was saying that because I was like, I was saying like, thank you for putting this so diplomatically and I think correctly and great, yes, thank you, right? Like, so I didn't think I needed to like have a tweet that was explaining anything about it because it's so well explained already in the original source material. But somehow, right, this seemed really, I guess, in the same way that like whenever there's a national tragedy,

• 23:09Speaker 00

  the first thing anyone does is figure out how they can talk about the thing they care about as being now very important because of this tragedy. Even if it's like the most tenuous connection you could possibly imagine, it feels a little bit like that. It's like, a very reasonable statement was said by someone in a high-profile context. And so now everyone's thing, whatever it is, they're like swerving it around to be that thing for some reason. So that's kind of my feeling about it.

• 23:37Speaker 03

  This is what it taught me about B2B SaaS on LinkedIn. That's every time there's, every time there's something.

• 23:43Speaker 00

  Yeah. It's, it's like, actually, computer science is about this and here is my whole thing about what I want to talk about all the time. It's like, okay, like, you know, so that's, I just, that's what I would say first about it. I mean, there's a ton of things we can talk about, which is my point. That's what it felt like to me. It felt like, it felt like this was actually a pretty hard, like, if you went and actually listened to this, I think it's actually a bit difficult to come away with like, without at least appreciating what was said and being like, yes, that, that, there is a problem there. Like, she didn't say the entire problem with computer science or all computer science is, like, none of that was said. So she's just saying, I think we have a problem. And I think she's right.

• 24:26Speaker 00

  So, and I think it's hard to argue that she's right because we definitely have an understanding gap there. And especially in the age of AI, when you can now sort of program in English, I fail to see how, like, you know, focusing a lot more on a high-level languages makes all that much sense. Because if you really want to be high-level, just use English now. Like, it doesn't seem, so I don't know. So anyway, go ahead with, with the, with whatever the other thing is. That, that would just be my framing, since you asked.

• 24:53Speaker 03

  So I kind of wish that maybe there exist schools where this is the case. But like, I wish we just had basically two separate degrees. We had basically like, applied CS and theoretical CS. Which is like what math has, basically, right? Like, if you go to a math, you can, or it's like, you have a math major and then you focus on whether it's applied or theoretical or something.

• 25:16Speaker 00

  I think pure is the word. Pure math or applied math.

• 25:19Speaker 03

  They don't like it when you call it theoretical. They're like, it's real to me, okay?

• 25:27Speaker 03

  Except if it's complex numbers, okay?

• 25:28Speaker 01

  That's the dumbest thing that's been said yet.

• 25:30Speaker 03

  Then it's imaginary, all right? So the, because I feel like there's a really good case to be made for, we need a lot of people to like do things on computers to solve people's problems, right? It's like, hey, we have to write some software, figure out how to like, use Excel better or something at a company like that. Scripting Excel should be a whole course at college. There you go. Boom. Hot take of the day. If you disagree, you've never worked at a real company. So there you go.

• 26:01Speaker 00

  Also, it should be like a seven-year degree because it's pretty frickin' touchy. And there's a lot of buggy, there's a lot of buggy Excel spreadsheets out there.

• 26:10Speaker 03

  Yes.

• 26:10Speaker 01

  Capstone, roller coaster tycoon inside of Excel. Like, until you can do something like that, you're not really an Excel.

• 26:17Speaker 00

  Yeah. Well, you have to, you have to go compete in those Excel Olympics things they have.

• 26:21Speaker 03

  Yes.

• 26:22Speaker 01

  TJ, unironically, like, he's material.

• 26:26Speaker 03

  I almost auditioned for that, Casey.

• 26:29Speaker 02

  For what?

• 26:30Speaker 03

  Like, the Excel Olympics.

• 26:33Speaker 00

  I, I thought you, okay, I was on that sentence, I was, I was hanging on, you were like, I almost, I thought you were gonna say one and it was like, auditioned. I almost thought about maybe going on the thing. It's like, okay.

• 26:44Speaker 03

  There was a time in my life where I was doing so much Excel that I was like, I watched some videos and I thought, I think I got this.

• 26:51Speaker 02

  For work? You're doing this for work?

• 26:53Speaker 03

  Yeah, yeah, yeah. Just a random job that I had in college. We just did a lot of Excel and then I liked it.

• 26:59Speaker 01

  But TJ was a pretty big accelerationist in, in college.

• 27:03Speaker 03

  Yeah, EX, EX. I was in EX, in Effective Excel.

• 27:10Speaker 00

  Can we give a shout out to, to Krzysztof here for their awesome like macro video that he does on Excel?

• 27:17Speaker 03

  We'll link that, we'll link that in the description for this video because it's a deep cut and it's a very, I think it is the second video Krzysztof ever uploaded to YouTube, if I recall. So that's a really, that is a really good one.

• 27:31Speaker 03

  But my, my main point that I was trying to make before I got distracted on Excel, a passion of mine.

• 27:35Speaker 01

  Can we just one more time?

• 27:36Speaker 03

  Of course, yeah.

• 27:37Speaker 01

  You are cousin, you're, you're Uncle Rico.

• 27:40Speaker 03

  You know, I actually, I could have, I could have been in the Excel Olympics.

• 27:42Speaker 03

  I could have put that spreadsheet in the bag, okay? We would have won states that year, all right?

• 27:49Speaker 03

  Prime, just because you watched Napoleon Dynamite this week, some of us haven't watched it in a while and still get, can pull that out, okay?

• 27:55Speaker 02

  I cannot pull those references. Oh, that's what Rico, okay, got it.

• 27:59Speaker 03

  Yeah.

• 28:00Speaker 01

  Honestly, if his boss would have put him in fourth quarter, they would have won states. That's true.

• 28:04Speaker 03

  Give me your sheets. You know what I mean?

• 28:06Speaker 01

  All right, sorry.

• 28:07Speaker 03

  So.

• 28:07Speaker 01

  Keep going, keep going.

• 28:09Speaker 03

  Applied CS, I think should be in, like, that actually makes sense. And the whole goal is, we're just only going to teach you what we think is the most likely stuff for you to get hired at someplace that does engineering things right after school, right? It's like, we focus on internships. We just like try, like, cool, you want to do front end? Cool, finding out how to, how you can learn like next JS, we're doing it really fast or something. I don't know.

• 28:34Speaker 01

  Isn't that just a bootcamp at this point?

• 28:36Speaker 02

  Yeah, I was gonna make that argument.

• 28:37Speaker 01

  Four years.

• 28:39Speaker 03

  Why not?

• 28:39Speaker 01

  You need like. Because.

• 28:42Speaker 03

  Okay, are we, we're making the case that six week boot camps really prepare people for their jobs? That's not.

• 28:48Speaker 01

  No, but I don't, I don't know if four years would also prepare you properly for the anyways. Okay, sorry, keep going. I'm sorry to interrupt.

• 28:56Speaker 03

  Be like in a, even in an applied mathematics, you still take some theoretical math courses. It's not like you own, I'm just saying you're like elective section, I think would be more focused on the things that people are literally hiring for right now. Hey, cool. This is, we are partnered with a bunch of companies. We do this like there's engineering degrees where you go and you get, you have to do internships for like either like the whole time or the last two years or whatever the required. They partner with local engineering companies to do this, whatever, right? Like in my mind, that makes sense as a pathway and is a viable complaint of what, you know, see or viable like place that CS degree should be for. But also, you know, I think it's also good that we have people studying the theory of

• 29:41Speaker 03

  CS and figuring out new stuff and trying to understand data structures better. And maybe they're the ones who figure out how to make map reduce 5% more efficient and get to delete 10 data centers off of the earth. I don't know. That seems great. You know what I mean? There's like, or they're discovering new theories of, of CS and stuff like that. But those don't seem like, strike me as like crazy overlapped in like a four year degree program. You, you may end up in some different places, but both of them should still start. Actually, okay, I'll, I'll, well, actually I'm interested in this take Casey. I think it's kind of nice to have a high level language as your very first CS course. So you can see things show up on the computer and get people excited. Then I think right after that, you should have to do C.

• 30:25Speaker 03

  So like whether it's Python or Java or something else, like my first course was Python and it made me fall in love with programming. So I'm interested to hear whatever else thinks about that.

• 30:32Speaker 02

  So that, that's well like University of Maryland's comp side program does. They do Java as like the 100 courses and they go to like, oh, camel, Rust or C after the fact. But I want to, you mentioned the bootcamp stuff and I think this is literally the exact same argument. But for like the college people, because bootcamps were like, oh, you're learning React, but you don't even know JavaScript kind of thing, right? Because React is like another framework. It's like, it's literally the exact same thing happening.

• 30:56Speaker 03

  I agree that that's bad.

• 30:57Speaker 02

  And this, but.

• 30:58Speaker 01

  You should never utter that phrase. If you find yourself saying that phrase, you should go, okay, I'm going to be quiet. Just don't say it. That's it.

• 31:05Speaker 02

  That's it. This is literally like what this whole thing is remind me of is that that whole kind of debacle that happened.

• 31:11Speaker 02

  But like to Casey pointed out is like, why would, why should we learn high level languages? If you can just code in English now with agents? And like, I think to me that hits me really hard. It's like, I almost want to say for people in college, you, you should just focus on like the low, super low level stuff, because that's what's going to actually make you stand out. In my opinion, it's like, if I learn Python, sure, I could go type some Python, make a prompt to make a Python program or something. But when things go wrong, that's when like people that understand at a low level how to optimize, like performance, all that stuff is like such a niche category. And that's like, if I go to any company, there's only like a handful of people that actually do like performance stuff like, so like, uh, prime, right? And our DP, your team, your, your past team and all your adjacent teams.

• 31:55Speaker 02

  You were like the performance people, but there was like, you were just a small little group out of a huge engineering firm, right? And I feel like if you could just focus on that stuff, it's going to be so much easier. Not so much easier, but I feel like you have a better shot of just standing out like in the candidate pool. Oh no. What do you think, Rob?

• 32:15Speaker 01

  So that's the problem is that I still feel like this argument over, we're still kind of struggling with what is the purpose of CS even with these arguments, which is even with applied. Well, is it really important to learn a low level language? And does low level language really make you understand what you're actually interacting with? Beyond just simply like, say, memory layout. I assume that using C, you actually see, oh, it is just a bunch of ones and zeros. And a struct is just like, oh, okay, it's 24 bytes. Like I can see, I can see the memories, you know, like it feels better. But does it actually?

• 32:48Speaker 01

  Is it like, does that teach you how a computer works? And I know the internet was so ready to misunderstand Lori on this one, but there is a bit of truth in it, meaning that however a computer work, which please do not ask me, ask Casey, I know there's a lot of indirection with memory and there's virtual stuff and stuff, and crap gets paged in and what I actually write versus what actually happens is, is magic to me at least. And so it's, it just seems like maybe, uh, you know, even though C is closer to what potentially is happening, it's not really what is happening. And so it can kind of give me like a false sense that I, I understand. But I guess the thing that's good about it is that you do understand, but at the end, does that make you more practical? In 99.9% of all situations for programming in today's world, I don't know if that actually,

• 33:32Speaker 01

  or let's just say 80% of all situations, that doesn't necessarily make you more practical or useful as somebody at a job, because often debugging is not what's wrong with the structure layout. It's more like, we have 19 services because we went to a conference and microservices is still the greatest thing in the universe. And now we have to debug a problem of requests getting like fanned out. And now we're making 400 internal requests for one single user request. Stop burning down our computers and that has nothing to do with performance. That just has everything to do with bad decision making, which I don't know if low level or high level helps you in any of those situations. And so I'm not sure if this is a good argument. And then as far as universities go, for me, universities, universities, at least at one point, used to be where rich people learn how to go to leisure well.

• 34:15Speaker 01

  Well, you know, you didn't come out of a university thinking, I'm going to now be super practical at my job.

• 34:19Speaker 03

  Yeah.

• 34:19Speaker 01

  And computer science always understood that. That's why if you go get a MEC E degree or a civvy degree today, you come out and you can do finite element analysis and fluid dynamics. And then you go to a job and you never do those things again. You're never on, you're never on the ground being like, bro, hold on, let me just whip out this double integral for a quick second. I see some greens theorem over here in my elect, my electronics, like no one's doing that. And so it's like, so much of it's done, you know, for you via the computer. At least that's what all my civvy and MEC and my EE friends do is that they have a lot of that all happen automatic, automatically for them. And so it's like, CS or C S E E, C E M E, all were there to teach you the theoretics. And then it's up for you to become morphed into a job ready person.

• 35:03Speaker 01

  And so I'm not sure if university really, does it really matter that university does or doesn't? I think the problem is, is that computer science is the only one that there's such a battle over what is important in computer science. Is it the theory or is it the practicality? No, I'm saying from a university perspective, because university.

• 35:23Speaker 03

  The every, every, every single, every math department has this all the time. Every math department has been doing this for way longer than computer science has even been a thing. They've been arguing about this.

• 35:34Speaker 01

  Yeah, but, but, but the problem is, is that computer science, they make all their money off of grants for the theoretics, but then they're supposed to get all the students ready for all the practical, like it's a really bizarre thing, because they're kind of expected that all students who leave a CS program are practically good. Like that's kind of the expectation, which I, I'm, I'm thinking it's like a broken expectation problem as opposed to anything else. Like I don't, huh?

• 35:57Speaker 02

  I don't think the practical part is true.

• 36:00Speaker 01

  I think, well, I think people expect them to be able to like assimilate and learn how to become useful on a job after they graduate. And that's sometimes not necessarily true.

• 36:09Speaker 02

  In my experience, when we hire like new grads, we already have the understanding that they're going to be literally worthless for like the first year.

• 36:15Speaker 01

  Really? But what's their interview?

• 36:17Speaker 02

  I mean, it's obviously, it's obviously like the data structure.

• 36:20Speaker 01

  New grad interviews. You know what I asked him? How to do an async request queue in TypeScript, okay? That doesn't sound like theoretical CS at all. I was expecting practicality from day one. Okay. I did like a hundred interviews, buddy.

• 36:33Speaker 02

  Yeah. Well, this is, this is like way before, this is like 2010 ish or whatever. And we were just asking just pure cop side data structure, algorithm questions. Nothing really practical, because we know they necessarily don't have, they don't necessarily have the experience. You can look at, they have like one internship. To be honest, most internships kind of suck from being honest. So it's like, I don't really pay much attention to it, but we know that they do have like the theory and the foundational skills. But then their first years where they actually pick up all the practical stuff, like how do you communicate with people? How do you, how like, how do you actually transfer your skills to the real working place? We kind of help them transition there. And then like at some point they can then start leaning in on like their foundational theory that they learned in school. So that's kind of how we approached it.

• 37:18Speaker 02

  But this was like 10 years ago, so I don't know what it's like now, to be honest, but.

• 37:22Speaker 03

  I mean, one of the problems here is there's just so many things wrapped up in this discussion that I'm not sure, like you could talk about any of them for quite some time. So I think the biggest thing I would point out is just that.

• 37:35Speaker 00

  I think that when you're talking about what you should learn in a college setting, or more specifically, what you should learn in any kind of educational setting that is supposed to be more serious and less immediate. So for example, if I just want to learn Python, I mean, I don't know about you guys, maybe this is atypical, but I don't think it is. Like, I'd never programmed Python before, and I wanted to do a little bit of that for some stuff I was doing on the Substack. It took like five minutes. Like, the amount of time it takes for me to learn to program Python was zero time, right? Right. So the idea that like, we need to think very hard about what language you're learning,

• 38:20Speaker 00

  to me, seems just kind of ridiculous, because if you're a decent programmer, then if someone asks you to program in a new language, it's like, I have never found that to be a relevant question. It just isn't. And maybe there are certain idiomatic things that I'm going to have to like, go read about, because they wouldn't occur to me, because I haven't been programming this language for 30 years, or whatever it is. So like, maybe I will want to go like, to a good reference of some people who, you know, that I would respect, and say, "Okay, so what are some idiomatic things that you tend to do in Python? Like, alright, I see, thank you." And that's it. So, to me, when we're talking about what is supposed to be there for even a practical

• 39:01Speaker 00

  education, not even the theoretical part, it's what are the things that remain true? Meaning, if I learned them 30 years ago, I didn't relearn them today. There is a very big, I think, misconception, promulgated by this "C is not a low-level language" statement, which I think is false and not interesting, but that's kind of irrelevant to the argument, that somehow, the things that you would learn in a course that was about low-level programming do not remain true, because hardware has changed. And that is just completely false. Like, it's literally completely false.

• 39:46Speaker 00

  The exact same things you had to consider in 1968, and 1972, and 1977, and 1987, and 2012, and 2026 are exactly the same. The only difference is the diagram you are reading about how these things are interconnected on the particular thing you have will be slightly different. That's the difference, right? But name a thing, whatever you want, SIMD, caches, loop caching, virtual memory, whatever it is, it was there. It was there since like, at least 1970, but usually a little earlier than that.

• 40:27Speaker 00

  And so, if you went and learned about how computers work, meaning you were like, "Here is a survey of how they work. Generally speaking, this is what the architectures look like. Here's the different kinds of components we have." That literally has changed almost not at all, you know? In like, 50 or 60 years. The thing that changed the most is really like, that everyone has the best of everything now in their desktop chip. Like, all the things that you would have been like, "Oh, I had to have a CDC 6000, or a system, you know, 36091, to get those things, like a loop cache, and now they're just on my PC by default that I have in my home, or my smartphone, or whatever, right?" That's the big difference, but that's really it.

• 41:12Speaker 00

  They're all the same concepts. And so, the reason that I think, just to tie it back to Lori Wired's like, original things she said, the thing she said is the important thing. It's that if you teach a course in something like C, I would say Rust is probably equivalent here, although I'm not really a Rust programmer, but you know, it's, you know, you pick a language like one of those languages. The important part is that it's very easy for someone to say, "And here's how we would access this part of the CPU. Here's how we would do a thing that was cache-aware. Here's how we would multi-thread this. Here's how we would do an atomic." Right?

• 41:54Speaker 00

  Because you're not forced to use some high-level, like, co-routine implementation thing. That's the only way you do that in a language, right? In some higher-level language, that's great. In C, yeah, the vanilla language C doesn't have these things in it, but that's not the point. The point is that if you already kind of know roughly how to learn something like C or Rust, you can then just write there, without teaching the student anything new, go, "Now, we are going to go use this crazy instruction on your CPU by using an intrinsic to go do this thing that looks at a SIMD vector, and takes the mask register, and packs the ones that are zeros into like..." And you can understand it with knowing nothing more than just the C you already learned,

• 42:39Speaker 00

  and the little bit that they're trying to teach you now. And so, at least what I got out of her statement, and I thought she said it pretty clearly, she even used the word like, "extending down to memory management," or things like that. The thing about C is it's very easy to extend down to whatever I need to teach you. It's not that the language itself has baked in the concept of vector processing, but I don't want it to, because if it did, I couldn't teach you it as easily, right? Because you don't have to learn how it actually works if the language is auto-vectorizing. If I teach you how to write shaders, you will never know about how things like vectorization

• 43:19Speaker 00

  work, because it's already baked into the threading model, the like, S-I-M-T model of programming, right? So to me, like, I think she was absolutely correct. I think the perspective that she placed it in was absolutely correct. And I think that's the critical piece that has to be understood, right? And if you do learn those things, I think they're useful for decades. Whereas if you just learn how to program Python, and the libraries that come with Python now, I honestly don't think 30 years from now, anyone's going to care what those things were.

• 43:52Speaker 03

  That was one thing I really liked. So I technically have like, more of a computer engineer, it's electrical and computer engineering degree, is like, I didn't go, I have a CS minor. And it was fun, because we just learned a bunch of stuff like, about computers, and like, physically how they worked, and it was always very helpful to me, because I was like, "Oh, you can look at the code, and you're like, 'Well, that's going to be slow.'"

• 44:15Speaker 00

  Yeah.

• 44:16Speaker 03

  Because there's no way to make the computer do that in any way that's not slow. You can just look at it, and be like, "That's not going to be good." Which has served me well for a long time, which I like. And it didn't matter what language we did it in.

• 44:30Speaker 01

  Honestly, that was a load-bearing description you just gave right there.

• 44:34Speaker 03

  Genuinely, that was well-formed.

• 44:37Speaker 01

  That was well-formed. In my honest opinion, that was good.

• 44:40Speaker 00

  Yeah, and so, and it also means that you can do it in other languages, right? It doesn't have to be C. It just has to be a language with that easy downward extension, which is, it's very cumbersome to do that in Python. Even in Java, it's cumbersome to do it. It doesn't mean can't, right? It's not trying to, you know, people get hung up on, "Can you or can't you do this thing in a language?" It's just, it's just how simple is it to just extend downwards into something I'm trying to teach you, right? And we could just teach it all in assembly language, but I think at this point, that just being able to read assembly language is enough. I don't know that we need to spend time to teach people to really write significantly-sized programs in assembly language. So, learning C and how to read the ASM that comes out of it seems like a very, or like

• 45:23Speaker 00

  I said, a Rust, or something like that, that's, has this capability, I think, is a good idea, and I think Lori was correct.

• 45:30Speaker 01

  Well, that was my next topic, was what makes C better? And Casey, I believe that you absolutely nailed that one, so that one's done. And so, the real question is, is Python a mistake? Or should that be, is Python itself the mistake? Or is it the fact that there is not a heavy emphasis on C? And I think you've kind of also answered that, which is that, it's the exact same thing as like, I mean, it's kind of like a bootcamp in some sense. You go to learn React. It's like, it's not a very extensible skill, unless if you go and you practice it a whole bunch. So you're not learning like, how to make a great UI, you're learning how to use React to make a UI.

• 46:05Speaker 00

  Or COBOL might be the historical analogy. It's like, it's not a bad idea to learn it at the time. It's just probably not that relevant 30 years from now, right? But it doesn't mean you shouldn't maybe, like, I think, to T.J.'s point, start with it, because it's like, "Hey, this is what people are using today. You'll see a lot of it. It's easy for you to get some stuff up on the screen that you'll be excited about, right? That all makes sense to me." So I don't think Python is a mistake, it's the lack of some other thing that's the mistake, if it's happening. Which I don't know, by the way, because I'm, I don't know, like Lori was talking about this move away from it, I don't know anything about that, so I'm not saying that's happening, I'm just saying, assuming that is happening, I agree with her that that's a problem.

• 46:48Speaker 01

  I will say about the COBOL thing, though, the use of COBOL cripples the mind its teaching should therefore be regarded as a criminal offense.

• 46:57Speaker 00

  Okay.

• 46:58Speaker 01

  Edgar Dijkstra, QED. So just in case you're wondering, you can't do that one.

• 47:01Speaker 00

  Yeah, I think you could find that from Dijkstra about, just about anything you wanted.

• 47:05Speaker 01

  Everything is just actually crippling the mind? Yes, he also said APL is for people who want to go play around. APL is a mistake carried through to perfection.

• 47:15Speaker 00

  Yeah.

• 47:17Speaker 03

  Yeah, I think so. We did like, I mean, like I said, the first CS class I took was Python. And that's like, before I took that class, I literally thought learning programming languages was like learning Spanish or like learning a foreign language. Like the way everybody talked about it, like in my life, right? Cause I like, nobody in my life is technical. My dad's an accountant and I always told him, Dad, I like math, but I'm not going to do your math when I grow up. That's boring. Um, right? And so like, I literally thought programming languages were cool. I have to go memorize like a bunch of random grammar things. And like, there's like a million new vocab words and there's like 35,000 exceptions to everything. And you can't do well on the test unless you study extensively vocab.

• 48:02Speaker 03

  And I didn't want to do that. I hated that. Um, so when, but I had to take a CS class as part of my like liberal arts thing, cause I was doing mechanical engineering. So one of our like gen ed classes we had to take was a CS one. And I took Python and I was like, holy cow, programming is just like logic and like pattern recognition. And like trying to communicate with the computer. This is so much fun. And then from that class, the next class we took was data structures. And it was in C or C++ I think it was in like C++ but we pretty much only used C features. You know what I mean? Like that kind of thing. And that was for me, I thought that was a really effective way of getting people into a track for CS where you can see things like appear on the screen quickly. You can interact with stuff.

• 48:44Speaker 03

  You can like, you know, draw a turtle that moves around on the screen and like connects dots. You're like, I just told it to turn left and it turned left. That's so cool. Like I think JavaScript is fine for some of this.

• 48:54Speaker 00

  Whoa, whoa, whoa, whoa, whoa, Tej. A lot of people are thinking you're talking about an actual turtle. Now you're going to have to logo pill them.

• 49:00Speaker 03

  It's the, sorry, it's just literally a dot. It's a turtle's library in Python. So it's, it's not as exciting as you might think. It's just, it's, it's, it's just dots.

• 49:08Speaker 00

  Logo pilled.

• 49:09Speaker 03

  Yeah, sorry.

• 49:11Speaker 02

  I don't have it in my head.

• 49:12Speaker 03

  Yeah. Well, Trash, that's because you're obviously not a real engineer. You don't know Python turtle library. So.

• 49:19Speaker 02

  But you know what? I did C++ in college and it almost made me quit coding. I hated it so much.

• 49:24Speaker 03

  Was it at the very beginning or in the middle, Trash?

• 49:26Speaker 02

  I think it was like a 300 course. I did VB before that. I did VB and then we went to C++. And I was like, yeah, what are these till days doing here? It was weird, dude.

• 49:38Speaker 00

  Like, uh, the, the, the part where maybe like I, I wouldn't, uh, like I, I'm not sure about what the, in the original context of the, the discussion. The, the plus plus part of C++ is a problem. Uh, and so like, you know, the, the, I think if you were teaching people C++ the, like the actual C++ the way that it is now intended to be programmed, I do think that would be a similarly transient thing to teach people, much like Python, like not something anyone will care about in 30 years. I know everyone's going to hate me for saying that, but like, so, you know, that I, I don't know about that part. And maybe, maybe Lori Wired was actually implying that teaching C++

• 50:18Speaker 00

  the plus plus part was important. And I don't know, I would part ways there, but you, you, I don't know.

• 50:24Speaker 03

  It doesn't seem like her, her, like her to suggest that, but yeah, right.

• 50:27Speaker 00

  Important part of what she, the point she was making though, I didn't think. Anyway, so.

• 50:31Speaker 03

  I think too, to your point, Casey, like as much as, especially like, you know, or let's say before maybe senior year, like my senior year, we did a lot of like very pragmatic stuff where we like built projects together and like had to figure out how to collaborate. Like, so we ended up using Git and figuring this out and much of other stuff, which was very helpful before going into the workplace. Like, let's say years one through three or something. Spending a bunch of time teaching you all about the Tildes, where they go in C++ will make programming not fun. It will not be fun for people. And I think it is actively like not helpful. You don't get a reusable skill on the other side and you hate it. At least if you're going to teach them a not reusable skill, make them like excited about something in programming.

• 51:16Speaker 03

  Like, cool, here's how to make a website. And it wasn't fun, but at least you can like submit a form when you're done or something. I don't know.

• 51:23Speaker 01

  Um, so I, I do have to admit something, uh, which is that see at MSU, Montana State University, not Michigan. Okay.

• 51:32Speaker 03

  Yeah. Uh, well, I don't like Michigan State either, so that's fine. Yeah.

• 51:35Speaker 01

  I know. That's why I was spitting. I spit on that thing.

• 51:38Speaker 00

  Prime, you know how many like CEOs of HTMX right now are hanging on every word you're about to say?

• 51:44Speaker 01

  Oh, yes, they will.

• 51:46Speaker 00

  Montana.

• 51:47Speaker 01

  I have been to his office. I actually know where it's at. I know, I know all about the EPS.

• 51:52Speaker 03

  Represent.

• 51:53Speaker 01

  The engineering physical science buildings in which also has the computer science laboratory in. All right. Anyways, I digressed a little bit, but, uh, during my time, when I started in 2005, Java was the thing they taught. And I, man, I hate to admit this. I was a stellar student. Uh, the programming language concepts, I had such a high grade, they had to throw out mine and then curved to the next one. I got like a hundred and 70% multiple times on, on many tests. I was very, very good at it.

• 52:25Speaker 00

  Whoa, whoa, whoa, whoa. A hundred and 70%? Oh, after the curve.

• 52:28Speaker 01

  After the, after being thrown off and the new curve shifting me up so far.

• 52:32Speaker 00

  Damn, son.

• 52:33Speaker 01

  I was very, very good at these things. Uh, computer engineering, actually going through and doing like addition from a CPU's perspective on integers by hand. I didn't hear what Trash said.

• 52:45Speaker 02

  I'm just kidding. I said, what happened?

• 52:48Speaker 01

  Oh, what happened? I'm about to tell you. That's the, that's what I'm, this is me getting to the shameful part, Trash. Okay. Let me tell the story in a way that is, you know, because you have to tell a story where people are like, oh, he's just humble bragging. And then I turn it around back on them. Now here's the thing.

• 53:02Speaker 03

  And he's just bragging, bragging.

• 53:04Speaker 01

  I only had one class on C. And it was kind of like a brief one where it's like, hey, here's how to do a linked list in C. Here's how to do a few things. And it wasn't really like that good. It was more like on the language of C. It was a 100 level class, like just like how to dip your toes into C. And the thing is, is that after I graduated college, I could talk about heap and stack memory and all these things. But it didn't dawn on me until like a couple of years after college that when I was programming C, that I was writing either to the stack or to the heap, because I never actually learned those things. So even though I aced every single thing, could do addition of a register by hand. I could multiplication. I used to know all that stuff, all the barrel fund stuff. I could do all the things.

• 53:49Speaker 01

  But I actually had no tying it together to the point where I just, I did not understand it because they never taught C. I only knew Java. And I do think that that was a great disservice that I wasn't pushed further into that category where you take the theoretics, like actually drawing the squares and then doing these squares in language in a very practical kind of sense. And I, I really do, I really do feel like I, I greatly missed out on that because all practical assignments, like doing AVL trees and everything were all done in Java. All my stuff was done in Java, except for networking, which was done in C. But again, you're just playing with like, you know, net.h, you're not actually, you're not actually doing anything other than crafting UDP packets. So it's, it's not quite the same thing.

• 54:31Speaker 01

  And so I do feel like I was certainly robbed and it is, uh, I actually still look back at that as like, what a waste that was. I was learning so many things that I didn't understand how good they were, but I could pass exams. Like I could a hundred percent all of them, but I literally learned nothing because I didn't have that, like the top, the, the real tie into everything. And so I look back at that as like a huge failure. And so I, I do actually agree that I wish more things were in a sense, reaching down, as you say, like that makes perfect sense replaying it against my past. I could lead code. Yeah. Uh, yeah. I could lead code like a boss. Like I was that classic person that could come out and balance a tree, but I have no idea about anything else.

• 55:10Speaker 00

  Well, and I think that gets to like Lori's like extending downward thing, which is the crucial part, right? It's like, you did learn some C, but you can't like it. They can't just say, oh, here's the C language. Okay. Now you know everything. It's like, no, it's like the C language is just what, what really we're saying is it's a way, it's a framework that makes it easier to teach these concepts that you actually need to know. And again, it's not, you can't teach them their languages. Like one of the things I did on my Substack, right? Is I'm like, hey, here's this Python program. It's extremely slow to add the, add numbers together and we can show why. Well, okay, we could, we could use Python to compile that Python to see, you could go reach for PyPy or something, but you like, those are all these steps that now you have to do instead of just, no, we can literally just write in this language

• 55:54Speaker 00

  without changing anything we're doing. We can just show exactly how to do the thing from here. It's a very short step, right? And so it's really just about, it's a very, it's just very convenient. It happens to hit a nice sweet spot. It's why it's been such an enduring language. There's so many things written in it. So it's easy to go look at the examples of people doing things like operating systems in it. Cause that's what they write operating system and blah, blah, blah, blah, blah. Right? And so, you know, those languages and the C lineage languages, like now successor languages like Rust that are trying to be, you know, C and then some new things without breaking the C-ness of them. Right? Um, those are, they're, they're just easier to do this education. I think, I think that's just the bottom line. And when you, if you made me teach the same thing in Java, it's just going to be harder for me to do that. Right?

• 56:39Speaker 00

  Uh, it's not that I couldn't, it's that it becomes just a lot harder to, to. To do it because it's the wrong base, it's the wrong base layer and you know, that's, I don't know how else to underscore that point. And I think that's what

• 56:54Speaker 01

  Well, it might be really hard to do it at all. Like I'm not, I, you know, at least when I was in my 1.6 days, I don't know if you can really disinviguate the stack and the heap.

• 57:01Speaker 00

  Yeah, like in

• 57:01Speaker 01

  You knew up an object. I'm not sure if you really understand what's going on there.

• 57:05Speaker 00

  Yeah, in like Java, I don't know how hard it would be to try to teach someone about like, let's, let's change the memory layout of these things, right? Uh, you know, and you can, can I disable the garbage collector in some Javas? You know, can I go in there and monkey with that stuff? I don't know. Uh, but you know, that would be the kinds of things that I start to have to think about if I'm going to go, you know, use something like Python or Java, like how am I going to teach those things? Whereas in C, it's just like, it, it's just made, um, you know, the people say like, it doesn't matter anymore because we're not putting out a PDP 11, you know, A, obviously don't know anything about a PDP 11, like they've probably never even read the manual. It, it had virtual memory. So I don't know what they're saying. Were they like a linear address space? Like, no, it wasn't right?

• 57:48Speaker 00

  Like it has a, it has an address. It, it has a tape page table, a, a, a, a, address translation table. Anyway, so like, there's weird, weird stuff like that, that people think, but it's like, no, it really is. And even if I want to teach you down to the micro op level, it's very easy to do that in C. And, and it's not that easy to do it in other language for a variety of reasons. Uh, and so languages like C,

• 58:13Speaker 00

  help.

• 58:14Speaker 03

  It also just kind of seems like what, you know, if our goal, even if the goal isn't to get people a job from college or whatever, right? It's just like, we want to teach them, why would we make it harder than it needs to be? Like if it's like, cool, our goal at the end of this semester is that you understand the difference between stack and the heap and how it is managed in real programs, like how it gets used. Cool. That's one of our syllabus goals. Then why are we picking a way that makes that harder to show than it has? Like, it just seems, that doesn't even make sense. You should just pick the one that's the best and like, with the least number of hops and the least extra things that you need to do. So you can focus on them actually learning, like the primary task, right? Like if we're like, cool, this, this semester is about memory management, then let's like,

• 58:58Speaker 03

  teach them that. Let's not spend a bunch of time showing them like, here's sealed classes in Java and why your homework didn't compile. Because we put this extra keyword here to make sure you couldn't do this and cheat and get the wrong, you know, like, awesome. We didn't have to do, we didn't have to do that, right? That's just like wasted time and effort.

• 59:16Speaker 00

  Well, and I, I guess, I mean, just to, uh, to try to address people's concerns who complain about this. I, I mean, like, one of the things that I saw mentioned while people were complaining about the Lori wired,

• 59:28Speaker 00

  interview, I don't know what that thing was, the presentation,

• 59:33Speaker 00

  was like, oh, but there's no difference between the stack and the heap, right? Which is, which is true in the sense of like, they are stored in RAM, right? And so, so I, I agree with the statement that if you're, if you're asking me about like, does some, is something taking up more space on the heap or the stack or something like that? Like, I mean, you know, it, there, there's, there's ways in which that makes sense to say, and I might say it sometimes if someone asked me a question about like, that where that was the correct answer. But it, to say that there's no difference between the stack and the heap in terms of educating you about what's going on is ridiculous.

• 1:00:15Speaker 00

  Because literally there are a number of things in hardware that are about the stack. And it's not that they're storing it in some magic thing that's not RAM, right? Or something like this. It's not that we're saying that. It's that there really are a bunch of things that are processor specific that have to do with dealing with the stack. And so you do need to understand that difference. So you need to be able to understand like what a push and a pop are, like that's a thing that is specific to like, what, what is the stack pointer? That's a thing that's like in hardware. What is the return address stack? That's in hardware, not stored in RAM, by the way. It's in a special thing inside the core, right?

• 1:00:59Speaker 00

  And so there's a number of things like that. And if you just fundamentally have no way of understanding what those two things are, that would be really bad. And furthermore, it would be really nice if when you write something in your language, whatever the teaching language is, that it's very direct. So if I have int x and I know that the int is four bytes, and that's going to translate into a push four or something like this, right? That's helpful, right? These, these are help, these are helpful things because there's a very short, path to me showing you what happens versus a very complicated, like very high level language, where the, it's like, I gotta trace you through 30 steps to show you how it got to the thing, right? And so again, I just, I think people are using,

• 1:01:46Speaker 00

  they're using irrelevant differences to argue that there's no value, but there, but there is value. Like the fact that it's stored in RAM is not,

• 1:01:57Speaker 01

  By the way, the stack is stored in the balls. Okay. It's not stored in RAM. That's ridiculous.

• 1:02:01Speaker 00

  Yeah. Yeah. It's in, it's a second D stack.

• 1:02:05Speaker 01

  And classic stack.

• 1:02:06Speaker 00

  Yeah.

• 1:02:08Speaker 01

  TJ is not happy.

• 1:02:10Speaker 03

  I was really trying to find a way to make a joke about how nobody goes to stack overflow anymore. And I was more disappointed in myself in the inability to connect those two things together. Like what? We don't need this stack anymore. We have chat GPT, but it just didn't connect. So I was more disappointed with myself than, than you two.

• 1:02:30Speaker 00

  Yeah.

• 1:02:31Speaker 01

  Okay.

• 1:02:31Speaker 03

  Yeah.

• 1:02:32Speaker 01

  Well, I feel like we've kind of touched all things, but Trash, I, I kind of want to get your last, last thoughts here as we close this bad boy down. Can you give us your last thoughts? And I want you to close down the episode.

• 1:02:42Speaker 02

  My last thoughts is,

• 1:02:47Speaker 02

  how should I say this? I think these arguments are very tiring and meaningless in most cases. I very much don't enjoy reading them at all. Uh, I do think it's helpful that we are bringing it to light. I do think there is some importance of, you know, telling people what they, what they should and shouldn't learn. I think ultimately it's going to depend on the person. So even if you were to teach someone C, it's going to be up to them if they want to take it even deeper. I don't think it's going to be up to the schools to kind of push them to want to learn these things. But in my humble opinion, I think if you want to make it in this day and age of AI, you should really try a lot harder

• 1:03:29Speaker 02

  than you're trying now, because it's going to be even harder to, uh, stand out amongst the rest. So for those that are unemployed watching this, I think, uh, yeah, keep up the good fight and try super duper hard because I, I can't imagine what it's like to be in your shoes right now. And with the age of AI, I know it's pretty terrifying. And I know myself, what I'm doing is I'm making sure that I'm still studying because, you know, one day I won't have the job I have now. And Lord knows what I need to know when I have to go back and interview people. I think it's going to be a completely different ballgame. So I'm actually going down the hole of just like, understanding everything at a little level, because I think that's the only way I'm going to stand out, to be honest.

• 1:04:10Speaker 02

  If you want me to code something, I mean, I just literally type into a text box and it will pop out. So, yeah. Those are my, those are my thoughts. Kind of depressing, but that's kind of just how I've been feeling lately. I think AI is honestly kind of depressing me, if I'm being quite honest. And I think the only way I can, uh, take those emotions back is to actually challenge myself outside of like the workspace and using AI and just trying to build things for me.

• 1:04:39Speaker 02

  So, yeah.

• 1:04:41Speaker 02

  That was kind of sad, huh?

• 1:04:43Speaker 01

  Yeah. No, dude, that was good.

• 1:04:45Speaker 00

  This was an emotional ending. That was an emotional ending to the, to the thing.

• 1:04:48Speaker 03

  Trash, I think we should unpack this a little bit more later too, because I'm interested to hear more of what you're thinking.

• 1:04:54Speaker 02

  Yeah, for sure. We can, we can. I know many other people feel the same way. So. Yep.

• 1:05:03Speaker 01

  Well, let's, would you like to also close it out though, Trash?

• 1:05:08Speaker 02

  Thank you for watching the stand up with the Primogym. I am Trash. Featuring legendary programmer, Casey Moratore and Tej. Right back at you, baby.

• 1:05:17Speaker 03

  Thanks. Hey, well, I put doom inside of a Neo Vim GUI today. Okay, Trash.

• 1:05:24Speaker 00

  Can we stop? Can we, can we get the G out of Neo Vim GUI? It just looks like text still. It just looks like

• 1:05:30Speaker 03

  Casey, did you see?

• 1:05:31Speaker 00

  I did.

• 1:05:32Speaker 01

  Casey, I don't think, I don't think you realize what you're saying.

• 1:05:34Speaker 03

  Did you see me put a circle? I put a circle in it.

• 1:05:36Speaker 00

  But it's just like, you're just putting so, uh, somebody else drawing graphics on top of your text. It's just,

• 1:05:43Speaker 03

  Just wait.

• 1:05:45Speaker 00

  Yeah, alright.

• 1:05:45Speaker 01

  Until next week, I'll do a demo.

• 1:05:48Speaker 00

  I don't know, man. We're delusions of grandeur with the Neo Vim people here. Come on.

• 1:05:52Speaker 03

  Casey, what can I do to convince you it's a real GUI?

• 1:05:56Speaker 00

  I don't know. Have like a lot of images in the graphs happening like next to the code or something like, give me something I can sink my teeth into, Tej.

• 1:06:04Speaker 01

  TJ, the thing that I told you you need to put into Neo Vim, that's what he wants to see.

• 1:06:09Speaker 00

  Yeah.

• 1:06:09Speaker 03

  The thing that you told me to put in Neo Vim, I wasn't listening probably. What did you tell me to put in?

• 1:06:13Speaker 01

  Yes, you were. We laughed about it this morning. I am very offended that you've already forgot.

• 1:06:17Speaker 00

  If you're not spending a million dollars in tokens, I don't think you're getting,

• 1:06:21Speaker 03

  Well, to impress you, Casey, I will.

• 1:06:23Speaker 00

  Spend a million dollars in tokens and get me a Neo Vim GUI that actually looks G instead of T in the front of UI.

• 1:06:29Speaker 03

  God, that's my, that's my assignment this week. I'll come back with something next week and I'll present it.

• 1:06:34Speaker 00

  Alright.

• 1:06:35Speaker 01

  Thank you, everybody.

• 1:06:37Speaker 03

  Bye.

• 1:06:37Speaker 01

  Bye bye.