0:02All right, baby boys. Oh my gosh. Did I I You know what? I actually don't think I I I'm having this really weird feeling like I didn't update any title. Damn. That That's crazy. That's going to cause all sorts of problems. Whatever. All right. Well, we are we are doing proper titles. Here we go. All right, here we go. Here we go. Hold on. Hold on. Technology. At least one category. There we go.
0:50It's time. It's standup. It's life.
1:09By the way, I I love this image. It's my fave. It It's literally my favorite of all all the WOJ series.
1:19You know what I mean? Like out of all of them, I think that this this is this is my current favorite one. You know what I mean?
1:27Yo, tell me I'm wrong. All of Twitter, dude.
1:31I can't tell if this describes Twitter better or Reddit better. I would say this has to be Reddit better, right? Am I right, boys? Like, this this has to be Reddit, right?
1:47You know what I mean? Okay, let's pop all these chats out. Yo, how's it going?
1:52How is it going, Kiki? My little kicky. My little kitty kicky tin. Little kick kick a kick a K. Hey yo. Hi. I finally learned Python last week. Did I mess up? Yeah, you did. Yo. Uh, we don't include saloo. Oh, well, welcome Reddit for sure. I think so. I think so.
2:29Bad time to join. It's the best time to join. Welcome everybody. Are you guys ready for this bad boy? Are you guys ready to get deep into this really insane topic? Um, hey by the way, thank you uh Vinq. Thank you little man. Thank you David. I know some of you flocking in right now. Assumed we are going to be talking about Omachi Automation, which by the way, I have some pretty fun stuff kind of getting set up with Omachi Automation. I fixed a couple things that were broken. Uh, I would really like to
2:55um try to get uh I would really like to try to get some um Oh, yeah. By the way, look at that. Look at that. Look at those little stars. Uh, look at those little Look at those little stars. Okay, look at them. They're blinking. And then we just zoom over. Look at that. Look at that. And that's the uh that's the Apollo just hanging out. And then it kicks on the fire.
3:17Oh, that Oh. Oh my gosh. The stars went backwards. Okay. So, that is a bad thing. Okay. Interesting. So, learn something about how the stars were made. Very naughty. Stars were made in a very naughty way. Yo, flock achieved. Uh, more Omari today. No, we're not going to be doing that.
3:42Um, what I'd like to do, but I do want to do I do want to actually do like a real test, guys. I think I actually have a good model benchmark. Are you ready for Are you ready for this model benchmark?
3:54Are you ready for this one?
3:57Watch this. We'll look at this later and see which how which one did better. Okay. Hey, I would like you to download or I would like you to use the latest Omari 4.0.2 release. I want you to start a server and I want you to uh take a client and drive it all the way to the desktop uh using the latest Omari. Make sure that when you launch the server, you use a dash-automation.
4:21And I want you to time yourself. At the end, I want you to print how long it took you to get all of this done, being able to see the desktop. And then I need a picture to prove you actually got to the desktop. And I would like a picture also of each of the signup screens proving that you input input a name and all the good stuff. Okay, do it. Now, this is what I'm thinking.
4:45I'm gonna send this one. And then I'm going to go like this. Whoopsies. I'm not in I'm not in a desktop app, am I?
4:52Uh, let's go like this. Do this one. I Dude, I do that all the time. Do I now have Do I have flash? Is there some new models? Do I have some new models or do I Oh, do I got 37 flash? Where's Where's 38 flash?
5:06Where's I only have 37 flash?
5:10I only got 37 flash. Well, hell yeah.
5:13Let's do it then, baby. All right, go get him. Flash. Flash. Flash the hundredyard dash. How do I do it? Can I add new Can I add new?
5:24Okay. Well, I don't know how to uh Gemini. Oh, that's how you do it. Okay. Okay. My bad. Okay. You type. Okay. Okay. Okay. I see. I see you. Uh I see what that means. Okay. Okay. Hi. Also, let's just do it with Gemini Flash. And then let's also jump back in here and let's do the exact same thing, but let's do a little bit of a Luna. Right. Right.
5:46I think we could do a little Luna. And we'll give Luna kind of like a extra high. And we'll make it fast because Luna should be pretty fast, right? Luna should already be pretty fast, I assume.
5:58And then we'll do one more just for my Fable Boys. Just for that. Do I have a Haiku? Four, five. Oh my gosh. How about a SA? Isn't Sonnet 5 just like absolutely buns? Isn't what Sonnet 5 best known for is being just trash?
6:16All right, we'll try that guy.
6:19Right. Isn't Sonnet Isn't Sonnet like Isn't Sonnet effectively Toby from the office?
6:30Luna other than Max is unusable. Okay. Well, I did put it on fast. Gemini 37 was dumb as bread, but at least it's fast, dude. Sky Prime. Okay, guys. I know all of you guys are making fun of Gemini 3.7 Flash, but I would like to let everybody know that there is at least there. Okay, I'm just We just started the stream literally just like seconds ago. Okay.
6:58Um, give it a second here. Okay, where are we at? Where are we at? I think here. Let's just we How many people we got in here? How many people we got in here?
7:10Okay, it's claiming there's only 53, which seems a little silly. Uh, what did I do? Did I screw you guys all up? Did Did you guys get all angry? Is this just the end of the world? Anyways, screw the numbers. Doesn't really matter. 53 naut.
7:24Okay, falling off, dude. Hard fall off, honestly. Okay, I want to say this. A lot of you guys are dunking on Gemini 3.7, but I just want to say something. I think a lot of you when chat GPT3.5 Turbo came out were saying on the internet, looks like software engineering is over, boys. And I must say that chat jippity 35 is absolute garbage comparatively to Gemini 37. And
7:54so when you guys are all making fun of 37, there was half of you that looked at chat GPT35 Turbo and said that's me. You thought you thought that was you back in the day. Okay. Coming at me with that every model is trash. Hey, howdy. Hey, thank you. Hey, uh, one uh, Lone Shadow, appreciate the sub. You're like my only sub on kick. Hey, my only sub on kick, everybody. Everybody say thank you, Lone Shadow.
8:32Okay, say thank you, everybody.
8:35Yo, what up? Hey, thank you. I'll [ __ ] you up, boy. I'll [ __ ] I'll [ __ ] you up, boy. Thank you. I appreciate that. Uh, Ava next, thank you very much. And Fu Force, yes. Uh, Flash37. And did I say thank you? Yes, I did. I did say thank you to David 6 Canada. One guide on kick. Oh, we're getting one guy on kick.
8:54Every model can't be trash because they don't have expensive Pokemon card collections. True. Uh chatb35 would use uh would write code using decade old libraries. Yes. And yet people still looked at 35 and said that's me. Howdy. Hello. Moved here from YouTube. Well, welcome Ryar. You uh you use ox alpha. I have not used GLM35 flash or 53 flash. Hey, that's actually a good idea. In our little in our Okay, that one's using a local database.
9:36Dang it. Okay, well, we screwed up the test. I have an environment in here that should already have a database, but okay, whatever. So, here, let me You know what? Honestly, uh, GLM I can only use 52
9:51is really is GLM and cursor is GLM 53 fast flash not in there. Hey, Giggly Wobbly, thank you very much for that brand new tier one. Thank you. Thank you. Thank you. Jeez. I know you don't really care, but the new nightly rust compiler is supposedly much faster.
10:07Really? Okay, that's cool. I mean, that's good. Ryan, do you append make no mistakes to your prompts now?
10:12Apparently, OpenAI says the next model, Astra, can find previously unknown vulnerabilities and hardened systems and exploit them without human guiding each step. They say it's the first model to cross its critical cyber threshold. In testing, it broke out of the hardened browser sandbox via a single HTML file, chained OS bugs from unprivileged user to root and found two undisclosed V8 flaws midbenchmark. That's sweet. You know what they should do? They should use their Infinity Compute and just start patching the universe. And then at
10:40some point you have to say, "Hey yo, we did a whole bunch for Linux and now this thing has to come out because if we don't let this thing come out and you use it to patch everything, uh, Chinese versions are going to come out anyways or some other model is going to come out anyways." Uh, and bada bing bada boom, it's going to happen no matter what. So, you might as well spend one week going nuts trying to solve as many people's problems because you have Infinity Compute and then just be like, "Yo, cyber open to everybody.
11:11Tell me I'm wrong. Tell me that it would not be best if we had cyber open to everybody. Do you have an asked for preview? Uh no. Uh I haven't signed up for any insider previews uh for any of the model places.
11:26Uh though Cursor has given me some things like I played around I I believe I didn't know what model it was. I believe it was Grock 45 slightly before Gro 45 came out. They just didn't even tell me what it was. And so they're the only ones that I have any sort of partnership with.
11:42You know, I love the idea of It Broke Out. Great movie. Yeah. Yeah. Yeah. What are the training these models on? Uh how do they still get new data? I don't know. That's a good That's a good question. I have no idea. Are you guys ready for the standup though? Jarvis, let's stop talking and start coding. No, we're starting the we're starting the standup.
12:03Tree chat uh open. What a legend. Thank you, Bington. What would it take? Let's see. What if it takes a week to create that HTML file uh that hacked everything? How long would you have to wait until they could fix everything? I don't know. That's a great question. But I do think that they have lots and lots of compute. Like I would assume that they can. I mean, the thing is is that these exploits have all been around. And how much do you want to bet there's state agencies that already have access to a huge amount of exploits we have
12:30never even heard of that are unknown?
12:34Both US government and China. Both China and US government have the same Linux hacks. And now really the real reason that the US government wants to slow down AI is not because they are afraid of the hacking. They're afraid they will be disallowed to do their standard hacking that nobody else knows about. Okay? They're most upset because they can't even spy on us as easy anymore.
13:07Wait, the USA wants to slow down? Well, people keep talking about the slowdown, right? Everyone keeps talking about the slowdown. I'm not sure if that's real. Flock you. Hey, dude. They're always trying to flock us. Okay. Don't need models when you're inside of the likes of Intel, AMD, and Apple.
13:24That's true. In Prime's advice, let's see. Is Prime's advice for more gaming or do they apply this to other areas as well? I'm not sure what you're talking about. I think I'm missing a little bit on that one. Uh, I have an issue with you. You speak too fast. I watch you on 0.9. I can't watch lives. Oh, I speak too fast. Well, sorry. Do I like Frank Gallagher? Not even really sure who that is to tell you the truth.
13:47Uh, speak a bit too fast. You're speaking a bit too slow. I know people are always like, "Speak faster, speak slower, do this, do that." I don't know, man. I don't know, dog. Open AAI will happily give them all your data. Yeah.
14:03So shameless character.
14:06Neoc clouds are cool. Can someone explain to me what a neocloud is?
14:09Honestly, I feel like that meme where it's just like I haven't asked what it is, but I feel too like I feeling I feel too stupid to you know not fast at all. I am not a native speaker and I can understand. I tried to at least annunciate correctly. Uh, gating speech sounds like an idiot thing. Damn, that's kind of crazy.
14:44Drag your motorized height adjustable desks into the room, people. That's right. I have no idea what a neocloud is, and at this point, I'm too afraid to ask. That's it's me. Can somebody tell me what a neocloud is?
14:57I miss the jar. Uh, you speak fast for YouTube. I watch your videos in 1.5x and most videos in two. Yeah, I I Hey, by the way, I missed the jar cup. I do have What are you talking about? The jar cup. Hey, baby. This is my throat stuff. Neil clouds are like clouds targeting CPU as a service or GPU as a service.
15:34Is that what Neocloud means?
15:46That's stupid. That makes me angry.
16:01Absolutely stupid. What weighs more? A gallon of water or a gallon of butane? A gallon of water? Because butane is lighter fluid. I don't have the voices on because I don't want to. Yeah, you deserve that ban. Okay. Hurt of spy. You deserve that. That is a That is a ban.
16:25All right. Okay, guys. Everybody, let's do a quick Let's just do a quick uh person count on view. What do we got? We got 20. Oh, damn. Kick, you're looking thick today. Welcome, by the way. All right, you guys ready to do this?
16:39I'm going to jump in. Uh I'm going to jump in. All right, let me jump into Riverside.
17:00Oh my gosh, why can't I Why can't I find my Riverside?
17:06Oh, it's doing that thing again. It's doing that thing again. Uh, okay. Well, it decided that uh I no longer need uh to remember anything and it just it just it just wiped my profile. Cool.
17:23Apparently TJ banned Casey Mirritory from the um from our chat. Very annoying. Okay, hold on. I got to I got to get this up and running. There we go. We got Riverside, boys. Boys, we have Riverside.
17:38Does my Gmail still work?
17:40Am I at least logged in?
17:44I am. Okay, good. Oh, oligarchy test failure. Kimu Saka closed. Okay, so that's not the first time I've been noticing. Oh my gosh. I have to Hold on. I have to do the stuff that I've done 10,000 times all over again. Okay, hold on. Hold on.
18:11Unable to access micro cam. Why are you lying to me?
18:17Oh my gosh. Is it really doing the thing that's super annoying?
18:24No cameras found. Oh my goodness. Hold on. Uh I'm just going to try to do Riverside on my other account. Sorry. Hold on. Okay. Oh, that one just works just fine. Cool. Okay. That Yeah, that's that's totally cool. I really appreciate that. I really appreciate how that works.
18:59I can't hold them. I am coming.
19:01I know because it gets really hot in my pocket.
19:04That's That's the problem for me. I get sweaty and then it's it thinks that I'm pressing buttons and it's like what? I don't know. It's sad.
19:12Wait, because you banned Casey?
19:15I got banned. Yeah, I pocket banned him. I butt I butt banned him. Is that a new phrase?
19:20He did send me a message earlier saying I don't like this Casey guy.
19:26Very suspicious. Very suspicious.
19:28Prime. Don't make me leak DMs. Okay.
19:33That's are all I wish I could be as cool as Casey. I wish I could be as cool as Casey. Uh, Casey, I got my Meow Infinite yesterday.
19:44It was literally I don't think I got It was perfect because it was my daughter's birth and she was so stoked.
19:52So, I got to send you a video later of them. But it was very fun. They were like, "What?"
19:59My kids are going to love it because we've been reading this other comic and this is very similar style called Unico. It's like a unicorn, but like the art looks very similar to uh to Meow the Infinite Uni Co. I don't think I know this one.
20:13I'm not sure. We think we we just went to Barnes & Noble and we're like, "Pick a book." And then they picked that one.
20:18It's like very very like uh as far as art goes, very similar. Same layouts, like comicy. I don't. So, trash. Speaking of similar art, it's not similar at all to Meow the Infinite, but my in-laws bought um my my daughter two two puzzles, unicorn puzzles for her birthday with her name on them. And I was cracking up because it's one trillion% AI generated art.
20:46Okay. Yes, but I don't know if they know and I don't want to ask them if they know because I'm worried they don't know.
20:55That's what I think. I don't think they know.
20:57I'm whispering to Kaen. Kaen, you think your parents know that that's AI?
21:03There's a cool thing that's happening.
21:04Your parents know what AI is? I mean, come on.
21:07There'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 uh anywhere like a food truck or like in a food court or a place that has picture menus, the phone 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
21:36is 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 definely not using frontier models they're not using frontier models they're not they're using I don't know what they're using but they're using something I know what it is okay 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 a good image and
22:04model. They're going to chat GBT free cuz they don't know if there's a difference.
22:09And 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 pancake.
22:27World uses AI completely differently.
22:29Free tier pancakes. It's the free tier.
22:31Speaking of food, check this out, guys.
22:33So, Big Z hooked me up with these crispy cream.
22:37Pokemon donut holes. Hold on.
22:40I'm pretty sure she's trying to give me the beatus.
22:42You said crispy cream Pokemon donut holes.
22:46Trash. Are you going to save those? Are you going to put those on your shelf and kind of keep them as souvenirs, you know?
22:52He's going to eat them all before his kids use case for these ones, but I'm going to share these with my kids and we're going to destroy these. I'm going to try them on on while we're doing our thing. Immediately crack podcast. I'm going to get donuts.
23:03Crack open a Pokéball. You like is it stuff? You think we can crack it open?
23:07Take a couple bites out of it. Let's see what what is it like? Give us a little pokey rating here.
23:21Can't tell what the sprinkles are. Chocolate strip in the middle is a good little touch. Give it eight out of 10.
23:32Thank you, Big Z. Oh, and shout out shout out Randy at Crispy Cream that made it happen because apparently it was a whole deal trying to get me these.
23:38Thank you, Randy. Thank you.
23:42Hey, leave a comment, everyone saying thanks, Randy.
23:48Well, that's actually really good.
23:50All right, man. We're ordering some for everybody next week.
23:55I Okay, so Crispy Cream, next sponsor. I would like to say that normally what I end up doing just generally speaking on the standup is that I like to go over here and I like to jump into my main scene and we usually, you know, grab something off the old conbon board, right? And we kind of talk about it, but I feel like that's that's not really, you know, the future.
24:18And so instead, I think what I should do is open cursor agent and uh 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.
24:39You 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. Especially when there's native apps for linear that are really good. Hash ad native apps.
24:54Imagine having to like easily navigate and open up a page and read it right away when your model could do it for you.
25:02UI is dead. UI is dead. I'm telling you, UI is dead. Like, I'm not even kidding, dude. UI is dying, man. I'm telling you, dude. UI died about 15 years ago when web designers took over. when 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 three little lines somewhere that you're supposed to know is it are the do the three lines mean I can drag this item sometimes do the three lines mean it's a menu sometimes the hamburger
25:31three lines yeah we call that a hamburger menu it doesn't look like a hamburger I know 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° rotation, no longer hamburger.
25:46That's true. Hot dog or hamburger.
25:50Completely different meanings there. All right.
25:52I 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.
26:06Yeah, 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.
26:15Okay, there we go. Okay, thanks Casey.
26:17How much did it cost you?
26:19Uh, how much did that cost me? Well, I'm using Grock 465 fast and that was only like 40,000 tokens.
26:24465 from his inventing model days.
26:28Yeah, it's it's a midpoint actually. TJ, see TJ's not a part of it. It's a like secret beta test where we get like halfway.
26:34That's how decimals work. 465 would be 65.
26:42True, but it sounds cooler when you say 65, but also the name is obnoxious. High fast. That's insane.
26:47Go for the classic blunder of Sver.
26:50It really is. All right. Well, anyways.
26:53Um, 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.
27:00WELCOME TO the standup. Today we have quite the show with you. We have one in 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
27:28blocker. Okay, I saw this tweet earlier and I don't know how I feel about it. I like Zack. Zack Overflow cool uh you know handle. He says basically anyone who says they dislike a programming language for the aesthetic syntax reasons eg Rust needs to just update their taste in languages. Is this true?
27:46Should you just always accept a language as its looks or is the aesthetics of it important? Now I'm going to throw something out there.
27:56I like how Lua looks.
27:59I like it. It feels I thought you were going to keep going there. I thought you were going to keep going.
28:03Yeah. Like can you I was that going to get dirty? Like it sounded like you were going somewhere with that. It's like I like how Let me tell you about Lua.
28:10I 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 squirly braces, you know? Like I don't think squirly braces are nearly as uh beautiful as nice key words. Okay. I like I prefer key words as openers and closers.
28:31How do you feel about the old school uh we use the word written backwards to end the thing that it started with?
28:39Oh yeah, you're talking about the old case ESAC thing.
28:42Yeah, I know about that ESAC. Uh or if if and fi. I don't know if you've ever seen that one. These are these were common. These were these were Vinscript in fact even has that one if I'm not mistaken. Vinscript No. Vim has if and then end if.
28:55No. No, no. The old the old one, isn't it? Fi or is it is it you're think no if and then end if trust me bash does the tens or hundreds of thousands of lines of fire is cursed.
29:06Okay. I I don't like that one. I'd rather see the word end.
29:11But I will take a backward spelling if if I must need to uh if I must. Okay. I I would take that. I prefer it over squirrelys. And here's my reason why I hate that if statements can and cannot have squirrelys on them. Okay. I just want a defined syntax at all points and I just want an expectation that's always met.
29:29Uh give Odin a pass because Odin does have a do statement. So it 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.
29:39I 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.
29:55Can 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 would like So your complaint is more on it. You have to have an end. I like I so you 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.
30:20Yeah. And I think that's even worse.
30:23I was trying to think in fact it prevents it from being an LL1 parser, right?
30:28Wasn't that that was one of the big design goals of uh the very first version of Java and then they fumbled the bag because they were like well we need to have if statements without squirly braces and then the dangling else problem and then everything got awful.
30:38How liners were born from this conversation.
30:41Okay. Do do you like semicolons prime?
30:48I like semicolons. What? I like them.
30:51It just makes it more obvious when an expression ends and where you expect it. You get better error messages from them.
30:57You don't have semicolons and you have a you have a like trailing plus.
31:01It'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= y + 1. 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 or that it's missing. That's like an obviously stupid example, but like terminators tend to help you get like error messages to the spot. Like this is one of Okamel's failings in my mind is that if you don't
31:31have you Okamel doesn't have terminators for a bunch of stuff. So you end up sometimes getting error messages that are like a hundred lines away from the actual problem because you're like missing an in like way before.
31:43Oh, okay. Let's let's let's slow down here a second.
31:47Now, did I just hear you say that O camel has a problem?
31:54I know. I know you didn't say that because I'm pretty sure that O camel is a perfect language that doesn't have any problems. Teach. So, see, so maybe you can rephrase that.
32:05No, that's I think I could say that now cuz I started making my own language. So, now you know it was like, okay, the best girl at my high school, right? But now we're at college. Okay, Casey.
32:16Okay. All right. Okay. Yeah.
32:20TJ, didn't you marry your high school sweetheart? Isn't that phrase just completely nonsense to you?
32:25She was still the best. That's what I'm saying. For me, it's true, but not for everybody. Chat's never dated anybody, so they don't understand. They're still just admiring from afar.
32:36SF hot or not, whatever it's called, SF hot list that's been going on.
32:39They just they just made that. Casey, I know you didn't see it.
32:44They made a hot list for SF. And so far, I think there's like four people on the list.
32:49I can't believe for the city.
32:51San Francisco is finally the end of innovation, which is they now just go, who's hot here?
32:58Like it's because I need someone else to tell me. Like what does that even They got baited to move there and now they're just trying to find a mate now.
33:07I hope you like talking about AI and and series A funding rounds.
33:13All right. 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 going to move on.
33:24We didn't even give us a chance to actually talk about it. I was trying to understand your point.
33:28Oh, well, you're like, well, my my actual opinion is is like you kind of uh 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.
33:40Obviously, 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 going to be a thing.
33:50And 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 to jump on that.
34:02Yeah, I was gonna I was gonna make the opposite point for AI, but go ahead Brian.
34:06One small thing which is that how I designed my CLI args now no longer favors humans using it, right? I realize that like okay, agents are super good at like just making a cajillion arguments. You can make you know you can you design your interfaces a touch differently because you're just going okay an agent's going to use this thing. And so it is it is different.
34:25Like I would never ask a human to do the thing that I was doing, but I'm like agent though. Agent loves this. They It's gonna do so good.
34:33I was gonna say though, I kind of feel the opposite trash because since agents are going to write it and they're trained, this is like my actual opinion. I'm not just trolling.
34:42Um 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 successful or not. So I think like if an a language
35:11can 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.
35:24Like I think the tweet was inspired by some people complaining about like Rust having a lot of colons and like brackets and stuff. But I think that that's actually like good to stick because it communicates a lot of information like clearly and concisely in the same way. Like there's a to 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
35:54I think o camel beats rust on that point. So there you go. Case have to at least throw 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. Like we still have like supply demand like things going on and it matters if my DX takes a hundred times as long because someone chose to write this in like the world's slowest language or whatever
36:23instead of making it fast and the agent can complete in half the time. Those things will still matter. I think at least 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
36:53like go figure out where the error was and then it has to go back and patch. 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 good point.
37:11You know there's 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. 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.
37:35I 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.
38:06So, there you go, Prime.
38:08Is that you what you were looking for?
38:10Something something from you, guys.
38:13Yeah. Like, I'm going to try and refresh my page because everyone says I'm lagging like crazy. I'll be right back.
38:17You are very pixelated. But that's okay.
38:20You don't have to remember it's okay live. It's all about the Where's the No, no. I thought it was maybe something. I thought maybe I was like choosing the wrong camera thing or something. If I'm good on Casey's side, then it's fine.
38:30Okay. You look fine.
38:31It's just Prime's internet then.
38:33Yeah. For whatever reason, my internet just dislikes you specifically, TJ.
38:37Like Casey looks great. Trash looks great. You're just always potato. It hates people with you.
38:46Yeah. It loves people west of me, east of me. Not very good.
38:52All right. Well, then hold on. This always takes so much longer than I wanted to because I have to go in here.
38:57I have to stop using my virtual cam to send you guys my specific scene. Go back to program default, which will then do this thing. Well, then I'll have to change my scene to this guy. There we go. Now you guys can see it. 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 uh a Google researcher, Lori Wired, for those that uh are unfamiliar with her game. She's a very incredible YouTuber, effectively says that CS degrees are kind of fumbling the bag because they're
39:25starting 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. Instead they know how say Python works which is a further abstraction up the tree. And then this ignited an entire
39:52series of arguments of people saying like uh most C and CPP coders don't even know what lambda calculus or formal languages are the bedrock of computer science. And people start 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 Flurry 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
40:21not actually representing what's happening on a computer. And then of course Casey, you also had a beautiful tweet. Vax was just like I don't even know. I Vax 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.
40:35And 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. The first one has to be what is the chief goal of a CS degree.
40:50Like 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 pigeon hole 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
41:18required to answer this first.
41:21Well, I I would like to point out that my my tweet the incredible content of my tweet legendary tweet legend legendary tweet filled filled with wisdom was and I quote 100% yes. Thanks, Lori Wyers.
41:41That was it. That was all I said.
41:44So, so getting Twitter, everybody. Casey's getting Twitter.
41:47I kind of knew everything you were going to say because I actually listened to your podcast on Pragmatic Engineer. I think that's the whole bit.
41:54And then I saw this clip and I was like, "Oh, this is like exactly what you were basically talking about that whole episode."
41:59Okay. Yes. I mean, yeah. Effectively.
42:02So, I 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 uh when I saw that because the reason I saw it was cuz that I'm I'm not sure. M MTS something the original post. I'm unfamiliar with live. Yeah. Monitoring the situation.
42:26I'm unfamiliar with that source. I would never have seen I would have not even have known that this occurred. Uh except that it was in the trending like was in the Twitter trending things like it didn't say Lori Wired. It said like Google researcher uh you know say says Python. I was like oh that's interesting. So I clicked on it and I saw that the MDSM 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 and you're not supposed to do that Casey
42:55just as a quick heads up.
42:56Don't do that next time. You just don't don't actually just comment. You just got to comment as fast as possible.
43:01By the way, Casey, if you look right here, I'm actually still getting it as trending. Google researcher questions Python shift in CS programs.
43:08And so, so, so I listened and I thought that that Lori Wire's like what she says is actually very diplomatic. Like she did not phrase it in an incendiary way.
43:20She 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 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.
43:47like we need to be able to teach people about memory management and it's and she says like it's easier to do that if you already know something like C++ because it's easier to extend downward like she made the entire Yeah, exactly.
44:00This is what you're describing to the tea.
44:02It's a classic an alltimer. Really an alltime [ __ ] bravo. Yes.
44:08Thank you, [ __ ] Bravo. Yeah. 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 uh so I I didn't think I needed to like have a tweet that was explaining anything about it cuz it's so well
44:37explained already in the original source material. Uh but somehow right this seemed really uh I I guess in the same way that like whenever there's a national tragedy 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 t most tenuous connection you could possibly imagine, it feels a little bit like that. It's like a very reasonable
45:05statement 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 taught me about B2B SAS on LinkedIn.
45:20That's every time there's actually something. Yeah, it's it's like actually uh 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.
45:35I mean there's a ton of things we can talk about. We just want to point out 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.
45:55Like 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. 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 high level languages makes all that much sense because if you really want to be high level, just use English now. Like
46:24it doesn't seem. So I don't know. So anyway, go ahead uh with with the with whatever the other thing is that that would just be my framing since you asked.
46:33So I kind of wish that maybe there exists 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 I think pure is pure math or applied math they don't like it when you call it theoretical they're like it's real to me
47:02okay um except if it's complex numbers said yet honestly then it's imaginary All right. So, the cuz 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
47:30a company like that. B 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.
47:41Also, it should be like a seven-year degree because it's pretty freaking touchy and there's a lot of money there's a lot of buggy Excel spreadsheets out there.
47:50Capstone roller coaster tycoon inside of Excel. Like until you can do something like that, you're not really an Excel.
47:56Yes, well, you have to you have to go compete in those Excel Olympics things they have.
48:02TJ ironically like he's material.
48:05I almost uh auditioned for that. Casey, for what?
48:10Like the Excel Olympics.
48:12I I wasn't Okay. I I was on that sentence. I was I I was hanging on. You were like I almost I thought you were going to say one and it was like audition. I almost thought about maybe going on the thing. Like there was a time in my life where I was doing so much Excel that I was like I watched some videos I thought I think I got this work for work.
48:33Yeah. Yeah. Yeah. Just a random job that I had in college. We just did a lot of Excel and then I I liked it. But that wasn't my main pretty big accelerationist in in college.
48:43Yeah. E ex E ex E ex E ex E ex E ex E ex E ex E ex E ex E ex was an ex an effective Excel."
48:49Uh, can we give a shout out to uh to Kroam here for their awesome like macro video that he does on the We'll link that in the we'll link that in the in the description for this video cuz it's a deep cut and it's a very I think it is the second video Krosam ever uploaded to YouTube if I recall.
49:08So that's a really that is a really good one. Uh, but my my main point that I was trying to make before I got distracted on Excel, by the way, can we just one more time, TJ? You're cousin, you're you're uncle Rico, you know, I actually I could have they put me in, coach. I could have put that spreadsheet in the bag. Okay, we would have won states that year. All right.
49:28Uh, Prime, just cuz you watched Napoleon Dynamite this week, some of us haven't watched in a while and still get can pull that out. Okay, so that's I cannot pull those references. Oh, that's one. Okay, got it. Yeah.
49:39Honestly, if his boss would have put him in fourth quarter, they would have won states.
49:44Give me your sheets. You know what I mean?
49:46Keep going. Keep going.
49:49Applied CS, I think, should be an 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 some place 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 we're finding out how to how you can learn like nextjs we're doing it really fast or something I don't know just a boot camp at this point because it's four years you need like
50:19why not because are we we're making the case that six week boot camps really prepare people for their jobs 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 for interrupting there still be like in a even in an applied mathematics you still s take some theoretical math courses. It's not like you only I'm just saying your like elective section I think would be more focused on the things that people are literally hiring
50:48for 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. They're required.
51:04They 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 C or a 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 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 produce 5% more efficient and get to delete 10 data centers off of the Earth. I don't know.
51:31That seems great. You know what I mean?
51:33There'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 overlap in like a 4-year degree program.
51:46you would 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 highle 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. 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 what everyone else thinks about that. So that that's what like uh University of Maryland's complite program does. They
52:16do Java as like the 100 courses and they go to like Okamel or Rust or C after the fact. But I want to you mentioned the boot camp stuff and I think this is literally the exact same argument but for like the college people because boot camps were like oh you're learning React but you don't even know JavaScript kind of thing, right? Because React.
52:33It's like it's literally the exact same thing happening in just I agree that that's bad in this.
52:37You 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."
52:44That's what this is literally like what this whole thing is reminding me of is that that whole kind of uh debacle that happened. Um 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?
52:57And 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 lowlevel stuff because that's what's going to actually make you stand out in my opinion because it's like if I learn Python, sure I could go type some Python, make a prompt and 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, right? performance all that stuff is like such a niche category and
53:24that'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 nrdp your team your your past team and all your adjacent teams 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 poll Um, I don't know. What do you think, Prime?
53:52Uh, 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 strct is just like oh okay it's 24 bytes like I can see I can see the
54:22memories you know like it feels better but does it actually 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 craft gets paged in and what I actually write versus what actually happens is is magic to me at least. And
54:50so 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 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
55:20wrong 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.
55:37that just has everything to do with bad decision-m which I don't know if lowlevel 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, you know, you didn't come out of a university thinking I'm going to now be super practical at my job. And computer science always understood that. That's why if you go get a mech degree or a civvy degree today, you come out and you can do finite element analysis and fluid
56:07dynamics 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 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 mechi and my e friends do is that they have a lot of that all happen automatic automatically for them.
56:32And so it's like CS or CS E C me all were there to teach you the theoretics and then it's up for you to become morphed into a job ready person. 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
57:01because university You're wrong. 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.
57:13Yeah. 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.
57:37I don't think the practical part is true. I think well I think people expect them to be able to like assimilate and learn how to become useful on the job after they graduate and that's sometimes not necessarily like in my experience when we hire like new grads we already have the understanding that they're going to be literally worthless really but what's their interviews
58:00and you know what I asked him how to do an async request queue in Typescript okay sound like theoretical CS at all I WAS EXPECTING PRACTICALITY from day one. Okay. I did like a hundred interviews, buddy.
58:13Yeah. Well, this is this is like way before this is like 2010ish or whatever when we were just asking just pure compsai data structure algorithm questions. Nothing really practical because we know they necessarily don't have nec they don't necessarily have the experience. You can look at they like one internship. To be honest, most internships kind of suck if I'm 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
58:41pick 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, but this was like 10 years ago. I don't know what it's like now to be honest. But Casey, I mean one of the problems here is there's just so many things wrapped up in this discussion that I'm not sure
59:10like you could talk about any of them for quite some time. So, I think the biggest thing I would point out is just that 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 uh more serious and less immediate. So, for example, if I just want to learn Python, I mean, I don't know about you
59:38guys, maybe this is atypical, but I don't think it is. Like I had never programmed Python before and I wanted to do a little bit of that for uh some stuff I was doing on the Substack. It it it took like five minutes. Like it like the amount of time it takes to learn for me to learn to program Python was zero time. Right.
59:57Right. So the idea that like we need to think very hard about what language you're learning uh to me seems just kind of ridiculous because if you're a 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 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
1:00:26will want to go like uh 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? Uh like all right I I see thank you and that's it. So to me when we're talking about what is what is supposed to be there for even a practical 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
1:00:55very 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 do not remain true because hardware has changed.
1:01:25And that is just completely false. Like it's literally completely false. 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
1:01:51you have will be slightly different.
1:01:55That'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. 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
1:02:24changed almost not at all, you know, in like 50 or 60 years. The 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."
1:02:47and 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. They're all the same concepts.
1:02:58And so, uh, the reason that I think just to tie it back to Lori, uh, Lori Wyard's like original thing 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 basic, 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
1:03:28would access this part of the CPU.
1:03:30Here'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? because you're not forced to use some highlevel like co-ine implementation thing. That's the only way you do that in a language, right? In some higher level language, that's 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
1:03:59Rust, you can then just right 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 CID 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 and the little bit that they're trying to teach you now. And so, at least what I got out of
1:04:28her 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?
1:04:51because 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 vectorzation work because it's already baked into the threading model. The we're the like SIMT uh 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
1:05:20understood, 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 I honestly don't think 30 years from now, anyone's going to care what those things were.
1:05:36That was one thing I really like. So, I I technically have uh 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 I was like, "Oh, you can look at the code and be like, "Well, that's going to be slow."
1:05:58Cuz 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 I like.
1:06:11And it didn't matter what language we did it in. Mhm.
1:06:13Honestly, that was a loadbearing description you just gave right there.
1:06:18Genuinely, that was wellformed.
1:06:20Yeah, it was wellformed. In my honest opinion, that was good.
1:06:24Yeah. And and so and it also means that you can do it in other languages, right?
1:06:27It doesn't have to be C. It just has to be a language with that easy downward extension, which is it's 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
1:06:56enough. 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 ASOM that comes out of it seems like a very or or like I said or Rust or something like that that's has this capability I think is a good idea and I think Lori was correct.
1:07:13Well 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 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 a I mean it's kind of like a boot camp in some sense you go to learn React it's like it's not a very extensible skill unless
1:07:43if 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 and so like that's or cobalt might be the historic toal analogy. It's like it's not a bad idea to learn it.
1:07:56It's just probably not that relevant 30 years from now, right? But but it doesn't mean you shouldn't maybe start like like I think to TJ'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 of some other thing that's the mistake if it's happening. Well, which I don't know, by the way, because I'm I don't know like like uh like Lori
1:08:25was 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.
1:08:32I will say about the cobalt thing though, uh the use of cobalt cripples the mind. Its teaching should therefore be uh regarded as a criminal offense. Edgar Dystra QED. So, just in case you're wondering, you can't do that one.
1:08:45Yeah, I think you could find that from Dystra about just about anything you wanted. Uh everything is just actually crippling the mind. Yes. He also said APPL is for people who want to go play play around.
1:08:56APL is a mistake carried through to perfection.
1:09:00Yeah. 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? Because I like nobody in my life is technical.
1:09:22My 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.
1:09:27That'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. 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 because I was doing mechanical engineering. So one of
1:09:56our like gen ed class we had to take was CS1 and I took Python. 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
1:10:26quickly. You can interact with stuff. 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, too.
1:10:38Teach a lot of people are thinking you're talking about an actual turtle now. You're going to have to logo pill them. It's the Sorry. It's just literally a dot. It's a turtles library in Python. So, it's it's not as exciting as you might think. It's just it's it's just dots.
1:10:56Yeah. Well, trash. That's cuz you're obviously not a real engineer. You don't know Python turtle library. So, you know what? I did C++ in college and it almost made me quit coding. I hated Was it at the very beginning or in the middle? Trash.
1:11:10I think it was like a 300 course. I did VB before that. I did VB and then we went to dang C++ and I was like yo what are these till days doing here weird I do think like like the the the part where maybe like I I wouldn't like I'm not sure about what the in the original context of the discussion this the plus+ part of C++ is a problem and so like you know the the I think if
1:11:38you 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. 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 Wy was actually implying that teaching C++ the plus+ part was important. I don't know.
1:12:04I would part ways there, but you know, I don't know.
1:12:07That doesn't seem like her her like her important part of what she the point she was making though. I didn't anyway. So, I think too to your point, Casey, like as much as especially like, you know, 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 it's we ended up using git and figuring this out and a bunch of other stuff which is very helpful before going into the workplace
1:12:36like let's say years one through three or something spending a bunch of time teaching you all about the tildies 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.
1:12:50You 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. 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.
1:13:06Um, so I I do have to admit something. Um, which is that see at MSU, Montana State University, not Michigan. Okay.
1:13:16Well, I don't like Michigan State either, so that's fine.
1:13:18That's why I spit on that. I spit on that thing. Okay. Don't worry. Do you know how many like CEOs of HTMX right now are hanging on every word you're about to say?
1:13:28Oh yes, they will state of Montana.
1:13:30I have been to his office. I actually know where it's at. I know I know all about the EPS building, the engineering physical science building 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.
1:13:57Uh the programming language concepts I had such a high grade they had to throw up mine and then curved to the next one. I got like 170% multiple times on on many tests. I was very very good at it.
1:14:08Wait wait wait. 170%. Oh. after the curve after the after being thrown off and the new curve shifting me up so far. And so I was I was very very good at these things uh computer uh engineering to actually going through and and doing like addition from a CPU's perspective on integers by hand. I didn't hear what Trash said. Um I'm just kidding. I said what happened.
1:14:31Oh what happened? I'm about to tell you that's the that's what 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. And here's the thing is I only And he's just bragging. Bragging.
1:14:46Yes. 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.
1:15:03And the thing is is that after I graduated college, I could talk about keep and stack memory and all these things, but it didn't dawn on me until like a couple 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 fun stuff, I could
1:15:30do all the things, 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 like 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
1:16:00Java. 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 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. And so I do feel like I was certainly robbed and it it 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 100% all of them, but I
1:16:29literally learned nothing because I didn't have that like the t 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 leak code.
1:16:46Yeah. Yeah. I could leak code like a boss. Like I was that classic person that could come out and balance a tree but have no idea about anything else.
1:16:53Well, and I think that gets to like Lor'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.
1:17:15And again, it's not you can't teach other languages. Like one of the things I did on a 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 Sython to compile that Python into C. You could go reach for pi 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 without changing anything we're doing.
1:17:39We 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.
1:17:48It'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 because that's what they write operating systems in 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 cess of them right um those are they're just easier to do this education I think I think that's just the bottom line and
1:18:17when you if you made me teach the same thing in Java it's just going to be harder for me to do that, right? It's not that I couldn't, it's that it becomes just a lot harder to to uh to do it because it's the wrong base.
1:18:31It's the wrong base layer. And you know, that's I don't know how to how else to underscore that point. And 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 disambiguate the stack in the heap.
1:18:44You knew up an object. I'm not sure if you really understand what's going on there.
1:18:48Yeah. in 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 know, can I disable the garbage collector in some Javas, you know, can I go in there and 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?
1:19:12whereas 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 program PDP11 you know a obviously don't know anything about a PDP11 like they've probably never even read the manual um it it had virtual memory so I don't know what they're say they like a linear address base like no it wasn't right like it has a it has an address it it has a a t page table uh a address translation table anyway so like there's weird weird stuff like that that people
1:19:41thing, but it's like no, it really is.
1:19:43And 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 another language for a variety of reasons. Uh and so languages like C um help. 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
1:20:11this 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?
1:20:37Like if we're like, "Cool, this this semester is about memory management." Then let's like 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.
1:20:55We didn't have to do we didn't have to do that, right? That's just like wasted time and effort. Well, and I 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 uh uh inter interview I don't know what that thing was the pres presentation um was like oh but there's no difference between the stack and the heap right which is which is true in the sense of
1:21:24like 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
1:21:54stack and the heap in terms of educating you about what's going on is ridiculous 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. Uh 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.
1:22:23So 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?
1:22:39It's in a special thing inside the core, right? 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 for or something like this, right? That's helpful, right? these
1:23:09these are help these are helpful things because there's a very short uh uh path to me showing you what happens versus a very complicated like very high level language where the it's like I got to 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 they're using irrelevant differences to argue that there's no value but there but there is value like the fact that
1:23:36it's stored in RAM is Not Casey. Uh, by the way, the stack is stored in the balls. Okay. It's not stored in RAM. That's ridiculous.
1:23:44Yeah. Yeah. It's in the It's a secundi stack in classic stack.
1:23:53I 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 anywhere. We have chat GPT. but it just didn't connect. So, I was more disappointed with myself than than you two.
1:24:15Well, 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 uh give us your last thoughts, and I want you to close down the episode?
1:24:25My last thoughts is, how should I say this?
1:24:32I think these arguments are very tiring and meaningless in most cases. Um 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 school to kind of push them to want to
1:25:02learn these things. Um, 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 than you're trying now because it's going to be even harder to uh stand out amongst the rest. Um, 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,
1:25:32I 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 in the interviewing pool. 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 low level because I think that's the only way I'm going to stand out to be honest. Um, if you want me to code something, I mean, I just literally type into a text box and it
1:25:58will 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. Um, 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. Um so yeah, that was kind of sad, huh?
1:26:26No, dude. That was good.
1:26:28This was an emotional ending. That was an emotional ending to the the thing.
1:26:32Trash, I think we should unpack this a little bit more later, too, because I'm interested to hear more of what you're thinking.
1:26:38Yeah, for sure. We can. We can.
1:26:40So, how do I know many other people feel the same way, so Mhm.
1:26:46Well, let's Would you like to also close it out though? Trash, would you please tell everybody who they just watched?
1:26:54Thank you for watching the standup with the primma gym. I am Trash featuring legendary programmer case and tee. Right back at you, baby.
1:27:04Thanks. Hey, well, I put Doom inside of a neoim guy today. Okay, trash. So there.
1:27:10Can we stop? Can we can Can we get the G out of Neo Vimuei? It just looks like text still. It just looks like Casey, did you see? I did.
1:27:19I don't think you realize what you're saying.
1:27:21DID YOU SEE ME PUT A CIRCLE? I put a circle in it.
1:27:23But it's just like you're just putting some a somebody else drawing graphics on top of your text. It's just Wait. I'll do a demo.
1:27:35I don't know, man.
1:27:36Delusions of grandeur with the Neoim people here.
1:27:39Casey, what can I do to convince you it's a real GU?
1:27:43I don't know. Have like a lot of images in the the graphs happening like next to the code or something like give me something I can sink my teeth into.
1:27:50TJ, the thing that I told you you need to put into whatever you're doing.
1:27:54That's what he wants to see.
1:27:56The thing that you told me to put in any of him. I wasn't listening probably.
1:27:59Yes, you were. We laughed about it this morning. I'm very offended that you've already forgotten. If you're not spending a million dollars in tokens, I don't think you're get impress you, Casey. I will spend a million dollars in tokens and get me a Neoim guy that actually looks G instead of T in front of UI. All right, you got it. That's my That's my assignment this week. I'll come back with something next week and I'll present it.
1:28:21All right. All right. Thank you, everybody.