Casey Destroys Optimization Myths

Don’t let bad code get merged without reviewing (hopefully not by merge cop!). Checkout out Code Rabbit at https://trm.sh/coderabbit We dive deep into technical performance debates, specifically the nuances of floating-point math—is multiplying by a reciprocal actually faster than division on modern CPUs? We also break down the latest from Microsoft and NVIDIA, including the "RTX Spark" and the vision of "unmetered intelligence." Plus, don't miss "Trash Facts" about professional bowling and the latest "Snack Picks." [Sponsors] Linear: https://trm.sh/linear Sentry: https://trm.sh/sentry

• 0:00Speaker 00

  If you guys do not know, chat, chat, everybody in here, Casey, Trash. The startup is going through a bit of a change, alright. There's been some investment interest. And we are now raising a Series A like any proper multi-billion dollar unicorn startup.

• 0:17Speaker 01

  Cool.

• 0:17Speaker 00

  And so therefore-

• 0:18Speaker 02

  Decacorn.

• 0:19Speaker 00

  Decacorn, really, we're a decacorn. And so, I don't know what that means yet, but someone told me it's like, that's the best kind of unicorn. I think it means it has multiple horns. I don't know what it means.

• 0:28Speaker 01

  Ten horns. Ten horns, I'm assuming.

• 0:31Speaker 00

  Okay, nice. So it's kinda like, uh, the beast in the bible, the devil, right? Very cool.

• 0:36Speaker 02

  He doesn't have ten, does he?

• 0:42Speaker 00

  Uh, anyways, I'm sorry.

• 0:43Speaker 02

  Fact checkers? Fact checkers?

• 0:44Speaker 00

  Fact checkers? And so with that-

• 0:46Speaker 01

  I mean, I thought TJ would fact check that one.

• 0:48Speaker 00

  I actually thought TJ was gonna fact check me, but he fact checked me incorrect.

• 0:52Speaker 02

  It didn't seem ten, but it could be, it could be.

• 0:54Speaker 01

  Bunch of Christians and no one's read Revelation. What the hell's going on here, guys?

• 0:57Speaker 00

  Oh, huh.

• 0:58Speaker 02

  Well, that is kinda what it answers, you know?

• 1:02Speaker 01

  Anyway.

• 1:03Speaker 00

  Alright, so he has ten horns. Called it. Okay, primary reference, I got it. Okay, I was correct on that one. Boom.

• 1:08Speaker 02

  Okay, good.

• 1:09Speaker 00

  Referenced. Alright, so with that in mind-

• 1:12Speaker 02

  We're adjusting AI for our theology, alright.

• 1:16Speaker 00

  Let it happen, everybody, okay? My gosh, alright. This is not starting off the way I wanted it to. So anyways, nor does any stand-up. Alright, so with all of that said, that means as part of our amazing content, I have now organized us an amazing linear board, which will help us walk through a more organized stand-up, because if there's one thing I know about in raising a Series A is they wanna see agile, okay? I'm pretty sure that's correct still in today's day and age, so we're getting really agilized. No, it's actually gonna be, it's gonna make us hopefully have a more organized and amazing stand-up, so that the world's most attended stand-up, which is our current stand-up right now, right now, just in case you're wondering. You know, you may not know this, Casey.

• 1:59Speaker 00

  There's, there is, uh, there's 2,500 people on the stand-up right now. This is the world's most attended stand-up.

• 2:05Speaker 01

  It's like, you couldn't even, like, we, previously, no one knew if that many people could even stand at the same time. And we've just shattered that. We shattered that.

• 2:14Speaker 03

  Oh, yeah, definitely. What was the record before, like, 100 people, maybe?

• 2:16Speaker 01

  Yeah.

• 2:19Speaker 01

  Now we're at 2,500. It's world, it's world-shattering. It's totally, it's a game changer.

• 2:24Speaker 03

  Yeah. I want to add, though, we didn't get an invite or the topics until this morning, like.

• 2:28Speaker 01

  No, yeah, so this, this stand-up's gonna be amazing, cuz, cuz Prime's like, uh, we're gonna talk about this new, well, you know, I won't spoiler it, but let's put it, put it this way. It probably was a topic that would have been nice to research beforehand if we're gonna talk about it. That's, that's all I'm gonna say. I don't want to spoil anything.

• 2:48Speaker 00

  Well, you know, that, that makes this authentic, Casey. You could have reacted to Trash's workout gym knowledge.

• 2:55Speaker 01

  Okay.

• 2:55Speaker 02

  Right.

• 2:55Speaker 00

  Rather than facts.

• 2:56Speaker 02

  The expert.

• 2:57Speaker 00

  You get trash facts.

• 2:58Speaker 03

  I am not the expert.

• 2:59Speaker 00

  Trash is the expert.

• 3:01Speaker 02

  Trash fact.

• 3:02Speaker 01

  Dude, if there was a trash fact every episode, that would be, that would be fire.

• 3:06Speaker 00

  It's, it's literally going, it's going on the linear board.

• 3:09Speaker 03

  I literally have endless lore.

• 3:11Speaker 01

  Um, another thing that I would like to, to, uh, ask at this point is, if we are serious about this series A, which I think we should be, uh, is it time for us to start putting together a pitch deck?

• 3:25Speaker 00

  Oh, wow. I still don't know what one of those are, but I think we need to do that. I think you're, I think you're.

• 3:30Speaker 01

  We need some slides.

• 3:32Speaker 03

  My startup, dailybjj.ai, is going nuclear.

• 3:35Speaker 04

  Congratulations on product market fit. Sorry I called you the Difler.

• 3:40Speaker 03

  I knew you weren't the Difler. I always thought it was the intern with the weird glasses.

• 3:49Speaker 03

  Should've seen his code. But anyways, if you want to get on the ground floor of this thing, I'm taking angel investments.

• 4:16Speaker 04

  So you finally cracked the case, Merge Cop. You put the white space in the diff. You made the synthetic traffic. You made me approve that PR. You're the diffler and I always knew it.

• 4:33Speaker 04

  You've merged your last PR, diffler. I already clicked "merge" on my PR 35 seconds ago. Huh? My merge is blocked. Curse you, CodeRabbit! You saved the city from a code-review-related crime. Thanks, commish. Well, technically, it was CodeRabbit that saved the day. Not only do they have advanced AI features that can detect security vulnerabilities, like what the diffler was trying to merge, they also have ways to enforce styling, linting, and a variety of other tools, and so you can stop wasting time reviewing code that humans don't need to review.

• 5:11Speaker 02

  You can try it yourself at coderabbit.ai.

• 5:15Speaker 04

  So, Merge Cop, I would say it's a little misleading to say you did it all by yourself. Huh. Would you like some cake? Oh, sure, thanks.

• 5:28Speaker 01

  We need slides that say like, "Here's our total addressable market," which is every human on Earth. That's the market is everybody. People in, like, uncontacted tribes in, like, indigenous places will be using the stand-up somehow. They'll be standing up.

• 5:47Speaker 00

  Because, Casey, you know who our market is, right? People who are affected by AI.

• 5:52Speaker 01

  Which is everybody.

• 5:53Speaker 00

  That's TAM.

• 5:54Speaker 01

  Which is everybody.

• 5:55Speaker 00

  Earth is the TAM.

• 5:56Speaker 01

  There you go. TAM infinite.

• 5:58Speaker 03

  For next week's episodes.

• 6:00Speaker 01

  Aliens on another planet, like, they're gonna be getting the signals of the stand-up beamed into space. They're gonna, they're part of the market. They're being advertised. When they show up to contact Earth, they will buy your product if it is advertised on the stand-up, ladies and gentlemen. What else do you need to know?

• 6:15Speaker 00

  They'll be jonesing for whatever's on the stand-up.

• 6:17Speaker 01

  That's right. Trash, trash snack sponsor, the aliens will be like, "Ooh, I wonder what that tastes like."

• 6:23Speaker 02

  Great point. Great point. I'd love to see next week, maybe we each bring a pitch deck for what we believe the stand-up pitch deck would.

• 6:30Speaker 00

  Oh, let's do this. I don't even know, dude, it's gonna be so good. Let's each have a pitch deck.

• 6:36Speaker 02

  We'll each present ours next week then. Yeah.

• 6:38Speaker 03

  What would your pitch for the stand-up getting a sponsor or like for your own personal sponsor?

• 6:43Speaker 00

  No, this is a pitch for the stand-up.

• 6:46Speaker 01

  Why the stand-up is worth, worth probably a trillion dollars long-term, a trillion dollars long-term, I'd say, just to be conservative. I don't wanna, I don't wanna sound too aggressive, but a trillion, the pitch deck for why it's worth a trillion dollars long-term.

• 6:59Speaker 00

  Yes, I like it.

• 7:00Speaker 02

  We are a deck of corn.

• 7:02Speaker 00

  Alright. Alright, so with that in mind, I figured we'd do our first task here, which is trash fact. Trash, can you hit us with the trash fact really quickly?

• 7:12Speaker 03

  Yeah, I gotta think about this now.

• 7:14Speaker 00

  Well, you said you have endless supply of them, so I-

• 7:17Speaker 03

  It's true. I tried to become a pro bowler and my average was like 203.

• 7:21Speaker 00

  What? What? How is it every time something I never could have imagined?

• 7:27Speaker 02

  Trash, that is a good fact.

• 7:28Speaker 03

  I bowled for like a year, for like 10 years and I went super hard and I had to stop because I rock climbed a lot. So I was like, you know, rock climbing and bowling too much.

• 7:38Speaker 00

  Competing.

• 7:39Speaker 03

  But I used to do like-

• 7:40Speaker 02

  There's only one in your heart.

• 7:41Speaker 03

  I used to do like cash leagues after work and would go play for money. It was fun.

• 7:46Speaker 00

  You were a bowling hustler?

• 7:48Speaker 03

  I wasn't a hustler, like I competed like in leagues or whatever, but I had a really good average, so I thought I could really like make it, make it pretty far.

• 7:56Speaker 00

  What would be considered like a professional average?

• 7:59Speaker 02

  That's good.

• 8:00Speaker 03

  I think that's a professional average. I think like 190 plus maybe is pretty good.

• 8:05Speaker 00

  Really?

• 8:05Speaker 03

  I think my highest score is like 279, so I was like one strike away from a 300.

• 8:10Speaker 00

  Oh.

• 8:11Speaker 03

  You know?

• 8:13Speaker 01

  True story, I grew, I grew up in Massachusetts and so I, like, the kind of bowling that everyone talks about when they say bowling is like giant, giant ass ball with like the three holes in it, right?

• 8:26Speaker 03

  Yes.

• 8:27Speaker 01

  That you stick your fingers in, okay?

• 8:28Speaker 02

  Yeah.

• 8:29Speaker 01

  Never seen that. We didn't have that. We had candlepin bowling, because we're near Rhode Island. And so, it's this little tiny ass ball, it's like about that big.

• 8:38Speaker 03

  Duck bowling.

• 8:39Speaker 01

  It's not duck pin, duck pin's a different kind of bowling. It's, it's basically ten pin, uh, it, well, it's not even ten pin pins, it's long, thin, fairly heavy pins, and a tiny ass ball. You're, like, there's no way you have a 200 average unless like you're like the best bowler in the world. It's like really hard. Uh, because the ball is super tiny, and it's pretty hard to get a strike. Like, a good game is, you know, I don't know, three strikes or something like that, whereas if in real bowling, you know, or ten pin bowling, you'd be like, oh, three strikes, okay, like, you're probably not, that was, you're an amateur or something, right?

• 9:14Speaker 00

  Okay, I've never heard of this game. I've literally never heard of some alternative bowling game.

• 9:19Speaker 01

  Then duck pin bowling is even more narrow folk, I think that's like Vancouver and Rhode Island are like, and like, that's it. I don't even know who does duck pin. Not many people.

• 9:28Speaker 03

  Oh, where did bowling originate then? Did it not make its way up there?

• 9:32Speaker 01

  I don't know, I don't know, but like, in my town, there was only one bowling alley there, and there was, uh, and it was candlepin, and there weren't, I don't even know where the next bowling alley would have been, and I'm not sure what it would have been.

• 9:43Speaker 03

  Are you like a small town? Are you from like a small town?

• 9:46Speaker 01

  Very small, yeah. I think at the time, there were 2,400 houses in this town.

• 9:48Speaker 02

  That's why the pins are smaller, trash.

• 9:50Speaker 00

  Oh, that's small. Okay, that's actually small. You're from a real small place. Okay, I didn't realize that.

• 9:55Speaker 01

  Yeah, yeah, very, I mean, it's not, there are tinier towns, but yeah, 2,400 houses, I think, in it, when I was growing up.

• 10:02Speaker 03

  What was the population?

• 10:03Speaker 01

  Well, when I was, oh, and, and when I lived there, we, I lived to see the first actual stoplight put in. Like a, like a thing that could control. Before that, we had only, we just had this hanging light that would just blink at the, at the one main intersection.

• 10:19Speaker 03

  Are you from Radiator Springs?

• 10:23Speaker 03

  TJ? Radiator Springs, they just have the blinking yellow light? That's for you, buddy.

• 10:30Speaker 01

  Longer.

• 10:33Speaker 02

  Wow.

• 10:33Speaker 01

  But anyway, moving on, so that was an amazing trash fact and now I'm just like, what don't you do, Trash? Like, is there a thing you don't do? You like, you do jujitsu, bowling, climbing?

• 10:46Speaker 03

  I will, I will share more as the episodes go on, but I actually like, I have like a weird personality. Like as soon as I get interested in something, I take it as far as I can go, get bored and then I find the next thing. And I take as far, so I just have like accumulated many random skills.

• 11:00Speaker 01

  This is great.

• 11:00Speaker 00

  You know, the thing is Trash is I've also done that, except for I've done it on much more useless talents such as yo-yoing. I used to be a professional grade yo-yoer at one point. Uh, yep. And so one time I went to a conference where it was called the server challenge and I forget IBM, I believe, where you had to plug in 18 ethernet cables and uh, plug in 12 hard drives. And I was the fastest person at a 3,000 person conference and got my wife a MacBook Air for winning it. And it turns out like, they're like, hey, there's a useless talent award. And I'm like, oh, that's me. I'm that guy. You get, I was plugging in six ethernet cables at a, at a time and they're just like, this is, this is a revolutionary status going on. So it was pretty great. All right, but hey.

• 11:43Speaker 03

  Oh yeah, we'll continue.

• 11:44Speaker 00

  We can just continue on because Trash, I don't know if you know this, but it's, it's also your snack picks. Okay. You dominate the intros around here.

• 11:50Speaker 03

  Uh, well, what I'm eating right now is the Pimpcorn spicy nachos and I got my favorite red yum earth bag with me. This is my current duo. I already went through four, four of these bags this week, which is terrible.

• 12:03Speaker 03

  I, I have no excuse. So I've been working out. So I worked, I work, I work out in the morning and the afternoon.

• 12:08Speaker 01

  Just to do that?

• 12:09Speaker 02

  To try and make up for the yum earth.

• 12:10Speaker 03

  To like to make up for like my sins that I'm doing. I feel like so bad. And I make sure like I rinse my mouth out with water and all that stuff.

• 12:19Speaker 01

  And then I shove a Twix bar, I shove a Twix bar in there to just clean the breath.

• 12:24Speaker 03

  Well, that's what my dentist said. He said, don't let it sit in your teeth. Just rinse it with water and that's a good thing. Right?

• 12:29Speaker 02

  You talked to your dentist about it, you're saying?

• 12:31Speaker 03

  Well, fun fact, almost all the teeth in my mouth are like basically fake.

• 12:36Speaker 00

  I could never have guessed.

• 12:38Speaker 02

  We actually, we talked about that already. That was a different fun fact.

• 12:44Speaker 03

  I'm all messed up dude.

• 12:45Speaker 00

  Alright.

• 12:46Speaker 02

  Yeah.

• 12:46Speaker 00

  We should, honestly, Trash, that's fantastic. TJ, do you have any, you usually jump in with a little bit of snack?

• 12:52Speaker 02

  Um, I don't really have anything. I'm drinking a little bit of cinnamon, cinnamon spice tea because my nose hurts really bad because yesterday I was playing basketball at our friend's kid's birthday party and then one of the kids just ran directly into my nose and I had a enormous nosebleed for like the rest of the evening. The kid did not attempt to steal the ball. They just like lowered their head and ran straight into my nostrils.

• 13:22Speaker 03

  Like a, like a sheep?

• 13:24Speaker 02

  Yeah, it was.

• 13:26Speaker 00

  What was the plan? What was like, were they trying to do some weird form of a box out that we haven't seen? Some new, some new revolutionary?

• 13:32Speaker 02

  Me versus like probably seven or eight children and then I was like, I would, uh, pretty, pretty little, like between

• 13:42Speaker 00

  TJ's short if you're wondering.

• 13:43Speaker 02

  Most of them.

• 13:43Speaker 01

  Yeah, really short, I guess.

• 13:45Speaker 02

  Yeah. Well, you, you get low to do good basketball moves.

• 13:49Speaker 01

  Okay, okay. Yeah, that's a good point.

• 13:52Speaker 02

  So I'm, I'm spinning around and I'm dribbling like a maniac and then one of the kids just like wasn't, it just was excited. I don't know. And then ran right into my face and it hurt pretty bad.

• 14:03Speaker 01

  Well, isn't this, I mean, this could be a valid tactic in basketball, right? Like you just sacrifice, you just, you know, you take the technical.

• 14:09Speaker 02

  One for one.

• 14:10Speaker 01

  And you just, yeah.

• 14:12Speaker 02

  I'd get so mad. They tried it in karate, kid. Yeah.

• 14:14Speaker 01

  Yeah, yeah, there you go. And then you just, you know, it's since you're not only the star player of the other team, you're the only player of the other team.

• 14:22Speaker 02

  Exactly.

• 14:22Speaker 01

  Taking you out would be very valuable for their win, win chances.

• 14:26Speaker 02

  Yep.

• 14:26Speaker 00

  Win, win by default. That's a good move, honestly.

• 14:29Speaker 02

  Yeah. So I didn't, I didn't have any snacks right now. Sorry. Sorry to disappoint.

• 14:34Speaker 00

  That's okay. Alright, well, let's, let's, uh, let's, let's move on. By the way, I've been using this also for my stream. It's shocking how much more organized you are if you list everything out. Uh, it's, it's like, honest, it's, it's, it's, it's crazy. So Casey.

• 14:47Speaker 03

  What'd you do as a software engineer?

• 14:49Speaker 00

  You're talking about Casey? Uh, I kept it all in my head because Jiro was the worst experience of my universe. And so I was just like, I would rather forget what I am doing. than ever use that prop like I'd rather have to come to a meeting and explain to my boss why I didn't do something than use that okay

• 15:06Speaker 02

  that makes sense

• 15:07Speaker 03

  that's a valid excuse

• 15:08Speaker 00

  that's a valid excuse and you know you're at Netflix you know exactly the experience um alright so Casey I saw this tweet and I figured you could maybe give me like a little I think I understand the purpose here and what the problem is of it but I thought maybe you could help me a little bit on this one this is my blocker by the way for the day which is I saw this where someone said hey don't use division use reciprocal math instead can you can you can you can you maybe give us like us dumb people why this is either good or bad idea

• 15:41Speaker 01

  oh boy

• 15:42Speaker 03

  I have no idea what that means

• 15:45Speaker 00

  so right here he's doing array index I divided by pi versus array index I multiplied by the reciprocal so one divided by pi so he does the division thing once and then hits it with a bunch of multiple occasions

• 16:01Speaker 01

  yeah um so there's a lot of things that you would want to talk about here and that's uh I actually saw that tweet that you're talking about and unfortunately

• 16:11Speaker 00

  did you respond when delete your account

• 16:13Speaker 01

  no I mean see here's the here's the problem right so delete your account I love delete your account and I love what Ryan's doing there but what mostly what he's doing there is he's just like um and he even described this on one of his own streams he's just saying like look

• 16:30Speaker 01

  basically like you guys are all the people who like come up and and beg for change on the street or something like you're just totally making the world difficult to live in by like trying to get all like just just hounding everyone who goes into a particular area to try and get them to give you something right which is your attention whatever they're they're like these AI shell accounts or whatever people trying to get engagement they're trying to get follows whatever it is right and it just really degrades the experience of being in that place that's basically what he said on like one of his streams I'm probably phrasing it poorly um I I think he he's I think he said the phrase if you go to a third world country like France was the was literally I think the phrase he used

• 17:15Speaker 00

  I love Ryan

• 17:16Speaker 01

  in that stream so I don't know don't don't blame me don't come after me for this I'm just repeating what I believe it was um But anyway, the point is just that that was sort of the point of delete your account. And I, and I'm, he's right. Like, there's just all of this kind of, you know, really bad behavior that's going on there. And the, uh, this tweet, I don't really like being harsh on people who are trying to talk about performance, because I'm glad that we somehow went, you know, uh, over the past, you know, 10 years or so, to people actually starting to think that it matters that you're thinking about how the CPU is going to do things. So I don't want to like be mean to anyone who's at least trying to say like, hey, let's think about what this is doing.

• 18:00Speaker 01

  The problem is that it requires some actual knowledge to like actually get these things right. And a lot of times, if you just ask the AI, or if you just read something on a blog, and you then repeat that, it's not accurate, right? Because performance is a very specific thing. It's, it's kind of like repairing a car or something. You can't just know like one thing about like, oh yeah, you know, the carbureator or something, right? It's like, you have to actually know how the whole thing works, right?

• 18:28Speaker 00

  Carburetor.

• 18:29Speaker 02

  That's what a car is short for, right?

• 18:30Speaker 01

  Exactly. Like, I'm using, I'm using that as, as an example. Like, like, I'd just be like, oh, the spark plug or something, right? You know, it's like, I don't even really know much beyond like, there's a thing that ignites the gas, right? Okay, that's it. I couldn't tell you like, is it broken? Is it working properly? Does it need to be tuned? What's going on, right?

• 18:45Speaker 02

  Spark plugs are for electric cars, right? Because you get it from the batteries.

• 18:48Speaker 01

  I mean, that's, there you go. That must be. I mean, clearly. So yeah, this particular tweet is not, is not good though, because it misleads people, I think, on a lot of points. I'm not sure if the person actually tested this. They've got like seconds listed there, which is a little weird because I don't, those numbers, um, I mean, there's, there's a lot of things that could have happened there. I don't know. So I don't know the circumstances of it, but I just don't want people to be misled.

• 19:19Speaker 01

  When you're looking at a loop like this, both of those loops that are in that tweet, because I remember this. Of course, I can't read what's on Prime screen because Riverside won't let me, but I remember it fairly well. You're basically going, you're going through arrays, and it's basically a read, modify, write, right? It's loading a value out of an array. The value's a floating point number. Loading a floating point value out of an array. It's modifying that floating point number and writing it back to the array, right? And the operation that's being done on the array, the only difference between the loops is one of them is doing a divide. So it's loading, dividing, and rewriting. And the other one is loading, multiplying, and rewriting. Correct.

• 20:04Speaker 01

  Now, obviously, in math, if you do a divide of a number, if you say, "I'm going to divide by x", there's no difference between that and saying, "I'm going to multiply by 1 over x", right?

• 20:20Speaker 02

  True.

• 20:21Speaker 01

  That's true in math, meaning in like in grade school, you write it down. If someone said, "Oh, you know, a divided by..." What, sorry?

• 20:30Speaker 02

  Okay, all right, all right.

• 20:31Speaker 00

  TJ was a math major, so, okay, he's flexing that he knows these things, okay?

• 20:36Speaker 03

  Yeah, math major, get out of here.

• 20:38Speaker 00

  Yeah, Tej.

• 20:40Speaker 01

  This is, this is like in, in like junior high, right? Or earlier, probably, right?

• 20:44Speaker 03

  Yeah, yeah, totally junior high, yep.

• 20:46Speaker 00

  Yeah, totally.

• 20:48Speaker 01

  Come on, guys, like, seriously. I am not good at math. I am not good at math. If I know this, you know this. Come on.

• 20:56Speaker 02

  Yes, Casey, this is true. This is true and straightforward, yes.

• 20:59Speaker 01

  This is just something people did in grade school, right? And so, the idea, this was certainly true in the old days for computing, right? Is, look, if I have floating point, right? If I have the ability to use floating point numbers, integers, obviously, if we're using integers, then you have to do a lot of much more crazy stuff if you wanted to divide instead, I mean, multiply instead of divide. Go into that, because that's a whole nother ballgame. But if, assuming you're in floating point, then that means that you could literally just... Just store the number 1 over x, whatever you're going to divide by, and then multiply by it instead. So if you're going to do a bunch of multiplies, this is potentially an attractive option if dividing is slow, right? So if we look at divide and we go, the dividing's too slow,

• 21:44Speaker 01

  we could just do it once and then multiply if the multiply is faster. Now, it's worth noting that these are not necessarily the same operation. Because in floating point, right? An actual divide, this divided by that, right? May produce a different answer because you don't know, like, when you do an inversion like that, you're only capturing part of the answer. Like, the answer might have gone on for quite some time and you're only getting a smaller part of it. So you're not doing infinite precision math. And so you also have to be aware of the fact that this will change the answer. So you have to, you have to know, right? If that mattered to you or not when you're doing this. So if it's just some random thing you're doing for some UI element or something like that, then it probably will never matter.

• 22:29Speaker 01

  If this is scientific computing, it could be a very big difference. And so this is the first thing that I want to point out is that this is not a thing you can just tell somebody, always invert your numbers and then multiply, because that may be very bad advice in certain circumstances. That's thing one that wasn't mentioned in the tweet that you have to understand.

• 22:48Speaker 02

  The butterfly effect.

• 22:50Speaker 01

  Yeah, well, it's, yeah, basically like, you know,

• 22:52Speaker 00

  Jeff Goldblum with the water going down his hand effect.

• 22:55Speaker 01

  It might even just be the, the 1,000 mile butterfly. Like, it's this huge butterfly. Because it could be, literally, it could just be, just literally, the answer you get right here could be wrong enough that you care, right? Depending on the circumstances and depending on what those input numbers were and what their scales were and all this stuff, right? Who knows?

• 23:14Speaker 02

  Can you just, I, there is a decent chance a few people in the audience don't understand why that happens on a computer.

• 23:23Speaker 01

  Well, so to be fair, I'm the wrong one to explain it because I don't work on scientific computing, right? And people who work on scientific computing are much better at analyzing things like, you know, what's called ULP, or basically like the last, what happens to the final bits in floating-point numbers and things like that. But in general, when you're working with floating-point on a computer, what's actually happening there is that you're taking the size of the storage that you, you know, made for those numbers. So let's say that you're doing 32-bit floating-point numbers. In scientific computing, it might be 64-bit numbers, whatever it is. If you just have 32 bits, you only have a certain amount of space to store what you

• 24:06Speaker 01

  would prefer was an infinite series. Like, let's say we want to compute with pi, right? Everyone knows from grade school, pi is this infinitely long series. Like, there is no storage that can store pi. It just keeps going. So we have to approximate that, right? By putting it into some format. And all the numbers that we work with in floating-point computing, they have this kind of format. Just something simple.

• 24:30Speaker 00

  Can I interrupt you for a quick second?

• 24:31Speaker 01

  Please.

• 24:32Speaker 00

  Trash, when he says "pi," he's not talking about a snack, just in case you're wondering.

• 24:36Speaker 02

  I don't even like pie, actually. I actually hate pie.

• 24:38Speaker 01

  Me neither. Trash, yes. Another reason Trash is the star of the stand-up. No pie zone.

• 24:43Speaker 00

  Yeah, it's fruit.

• 24:44Speaker 02

  Trash, yeah. What's your favorite pie-adjacent kind of thing? Like, is it cake or brownies? Or like, what is it then? Love brownies.

• 24:50Speaker 00

  Pi is not adjacent to cake, and you'll take that back. We can, you know, we just got to move on because that was very offensive.

• 24:55Speaker 01

  Okay, that's it. Next episode.

• 24:57Speaker 02

  I'm putting it in the backlog. How adjacent is "take the pie"?

• 25:04Speaker 01

  Okay. So anyway, if you, if you look at what happens to these things during all your computations, every computation that you do is going to have some infinitely precise answer that's the answer you actually wanted and that you will want, you would have wanted to carry through all of your subsequent operations. But at every step of the way, the computer has to like round it, basically, to whatever will fit back into that 32-bit encoding. And that 32-bit encoding is basically, you know, it's a sign bit that says whether the thing's positive or negative. It's some number of bits for the exponent, right? Which says-

• 25:40Speaker 00

  The mantissa. I think is the sign.

• 25:42Speaker 01

  The mantissa is the next thing.

• 25:43Speaker 00

  Oh, the next thing. That's right. That's right.

• 25:45Speaker 01

  Uh, the exponent basically says what the scale is, like how big roughly is this number? And then the mantissa are the bits that actually say like what the number is. So the reason it's called floating point is because it's broken up into these pieces. Floating point, the point is the decimal point, if you will. The floating part is that we store an exponent. So we say where the decimal point is, and then the mantissa is just the actual digits. So we're basically floating the point to where it needs to go using the exponent, right? And the reason we do this is because we're trying to compute with extremely large and extremely small and everything in between. And so we can't just take 32 bits, which has a very small range if we just use an integer, right?

• 26:29Speaker 01

  If we fixed point, if we just said the decimal point is here, it would kind of suck. That's what floating point is for. It's why we use it for everything that isn't just nice, you know, typical whole number kinds of stuff that we're doing, right? You know, usually when we, when we use, uh, just ints or things like that, just integer numbers.

• 26:47Speaker 02

  The common, the common one that people laugh at, right, is it's what 0.2 plus 0.1 in JavaScript? And you get the 0.299999, right? So it's like, this is, this is why that happens though. So just, I feel like there's probably a decent amount of people who are like, just haha JavaScript stupid. I mean, you're right, but like, that's not, that's like

• 27:06Speaker 00

  Right for the wrong reasons.

• 27:07Speaker 02

  The JavaScript problem. Exactly.

• 27:09Speaker 00

  That works in C, it works in Java, it works in all of them.

• 27:12Speaker 01

  Yeah. Yeah. JavaScript is weird in that it kind of only has double precision arithmetic period. Like, it's just like, that's just what we do everywhere, I think is kind of how it works.

• 27:21Speaker 00

  Unless we use bitwise operations, then it converts to a signed 32-bit number.

• 27:26Speaker 01

  No, just magically at that time.

• 27:27Speaker 00

  Yes, for that one moment, you can actually do a, you can actually do a overflow at 32 bits. Only if you use, uh, bitwise stuff.

• 27:35Speaker 01

  Again, this is always why JavaScript compilers are so hard, right? Like, because they have to go like, oh, I don't, you don't want to do double precision arithmetic for like your loops and like counter and things like that. So it has to analyze and go, okay, I can turn this into an integer for this code, right? It's got to do that work. Normal languages, it's already knows, right? It's just been told like this, it doesn't have to think about that, right? But anyway, so. That's the floating point math thing. And that's sort of what this snippet was trying to say is just invert the floating point number so you only pay for the divide once, then multiply by it because that'll give you the same answer. Again, the tweet didn't say, no, it's not quite the same answer, and it might matter depending on what you're doing and depending on what the inputs are.

• 28:16Speaker 01

  But anyway, the other problem I have with this is it misrepresents the performance as well for a number of reasons. So the first reason it misrepresents the performance is because it has some weird statements about how long a division is. So it says something like 20 to 40 cycles for a divide or something like that, I think.

• 28:37Speaker 00

  Yes, 20 to 40.

• 28:39Speaker 01

  And I just have no idea where that's coming from. Like, I've no idea where they got that number from. Like, I really wish it had said where it got that number from, right?

• 28:47Speaker 00

  Came from his heart, all right?

• 28:48Speaker 01

  It came from his...

• 28:49Speaker 02

  Yeah. Yeah. You could feel it.

• 28:51Speaker 00

  You could feel those cycles.

• 28:53Speaker 02

  The CPU whisperer.

• 28:54Speaker 01

  Yeah. So a floating point divide, you know, if you look, so if you look at a modern chip, like, let's say a Zen 4 or Zen 5 chip, right? Something that you might be running today, that might be in a computer you'd buy today. Typically, their floating point divide units are extraordinary, very fast. I want to say that they can complete these things in, I don't know, five, six cycles, something like five cycles, maybe. I'm not sure what it is. We should probably look it up. But 40 is nowhere to be found. I mean, that's, it's not quite off by 10, but it's something like 10. Like, it's very, very wrong.

• 29:35Speaker 00

  I think it's from Chad Gippity, because if I'm not mistaken, I actually asked this question not too long ago, and I got the answer of like 8 to 30 or something like that, cycles for a divide depending on the CPU and architecture.

• 29:47Speaker 01

  I mean, the part that ChatGPT got right is the depending on the architecture part, because it does depend.

• 29:54Speaker 00

  Okay, okay.

• 29:55Speaker 01

  But like, let's, here, I'll just look it up on my machine right now. Like, we'll just see what roughly it is. So if I go in and I say, I want to do a, you know, 32-bit, like, we're going to do a div PS here or a div SS, let's say, and we look at what's going to happen on a Zen 5 chip, the total latency, right, is less than 10 cycles for that operation. Latency, right?

• 30:20Speaker 01

  So that means that if you-

• 30:22Speaker 02

  Multiply then?

• 30:23Speaker 01

  Sorry?

• 30:24Speaker 02

  What about for multiply?

• 30:26Speaker 01

  Multiply will be 4.

• 30:28Speaker 02

  Okay.

• 30:29Speaker 01

  I want to say, right? Um, so 3, 3 on Zen5.

• 30:33Speaker 00

  How many FPS though, Casey? We just need to know how many FPS is 10 cycles.

• 30:37Speaker 01

  Right.

• 30:37Speaker 02

  If it were a bouncing ball.

• 30:39Speaker 00

  Casey, how fast that ball?

• 30:41Speaker 01

  So 3 on that, and I think 4 is on my CPU, which is Zen4. I think it should be 4, right? Uh, is that correct? Zen4. Also 3. So, alright, this is better than I would have expected. I'm used to 4, these get 3. Right? Now, so, in general, that is the number that you would use if you were trying to analyze how long it takes you to get the answer back. Right? These loops don't care how long it takes to get the answer back, because nothing is waiting for that. Right? On the iteration of the loop, as you go through, right?

• 31:21Speaker 01

  You're going to do a load, the floating point op, either divide or multiply, and then the store. Right? When you do that, the next iteration of the loop doesn't have to wait for that to complete. So those are just in the scheduling queue, going. They can be dispatched. The next iteration of the loop is already decoded, and probably also in the scheduler, given this is a hot loop, presumably we're just running a thou-, you know-

• 31:50Speaker 00

  It's good looking. It's good looking, for sure.

• 31:53Speaker 01

  It's just cooking. So these are all going to pile up in the scheduler. The scheduler's going to dispatch them every cycle as it can. And so what we care about is not how long it takes to get the answer, because no one cares. What we care about is how long will it take us to issue the next one? So if we issue them on this cycle, how long does it take us to issue the next one? Right? And so that's what we typically call, that's a throughput number. So we typically call that, right? And if you look at the throughput numbers for multiply, right? Typically, you can issue 2 floating point multiplies per cycle, which means it takes half a cycle, effectively, to multiply. Half a cycle. Okay?

• 32:35Speaker 01

  That is what you, that is the number you should be listing or thinking about when you're looking at this loop. For divides, the throughput, again, is remarkably good on modern CPUs. So it's like in the 2 to 3 range. So is multiply faster than divide? Sure, right? Half a cycle versus 3 cycles when we're analyzing this loop, let's say, or something along those lines. But then you have to ask the question, do you care? This is a load, a op, and a store. Unless you're entirely out of the L1 cache, which is going to service, you know, at about,

• 33:17Speaker 01

  you know, like, let's say, one, you know, it's probably like 2 per cycle, 3 per cycle it could get you answers back from the L1 cache, right? If you're talking about the L2 cache, then, I mean, I don't know, the number of clocks it's probably going to take, 14 cycle latency, something like this. If we're looking at how long it takes you to actually get the memory back, it's not clear to me that unless the, unless the CPU was predicting extremely well what you were going to do and always had things, like, you'd have to go look at the actual bandwidth to see if you were actually going to get enough back such that you could operate on them at that speed. Maybe you can. It's, it's not in my brain right now as to whether you would actually be able to do

• 34:02Speaker 01

  that at a speed that would matter at the, like, let's say, 3 cycles throughput number that you're going to get on the divide. So it really just, it was very misleading, the whole thing, and as I, you know, you heard me say all that stuff, I would want to go look at this loop and, and check it first, and I'd want to think about it a little bit before I would post anything like that. And it was just, that clearly wasn't done, because none of those numbers make sense that are placed there. Is it true that multiply is faster than divide? Yes. Does it even matter in this case? No. I don't think it probably does, but maybe. And furthermore, usually, in most cases, right? When you're looking at actual loops, more is going on.

• 34:47Speaker 01

  And so it's also gonna mislead people. So point number three, it's also gonna mislead people in this final way, which is to think that they always have to do this. It's that this is the problem with these kind of like out of context performance things. People read it, they think, oh, I just need to invert all my numbers. And then my code goes fast. And then they go around inverting all of their numbers and turning them into multiplies. And the problem with that is, in a lot of cases, unless you were, unless you had serial dependency chains that really could have been unblocked by this change, you're just wasting your time. The divide would have been just as fast. It would have overlapped with other things you're doing in the loop. You were, the thing was waiting on a, on an uncached read or something. So the whole loop was, was waiting 80 cycles anyway.

• 35:30Speaker 01

  So none of this mattered, right? So the problem again is that like, performance is a thing. Performance is a process. It's an understanding. You have to understand all those things that I just said. They're not actually that hard to understand. You just, you can't, you just need to learn them. You need to spend a few weeks learning this stuff. And then you can think about, do I need to invert this number? Do I not need to invert it at that point? What you don't want to do is try to package it into a tweet like that, when you haven't really done the work of understanding why it matters or anything like that, because other people will take away the wrong lessons. So that was a very long-winded way of trying to explain everything I didn't like about that tweet. Again, no offense, sorry to the person who posted it. It's just, I don't think that kind of thing is helpful. It's sort of like saying, always call mem set, or never call mem set,

• 36:13Speaker 01

  or these other weird performance things that people pass around, which out of context, don't help anybody. Because you're just gonna make incorrect decisions if you think that the key to performance is always do X or never do X. That's not how it works. Spark plugs.

• 36:29Speaker 02

  Great job, Casey. Thank you.

• 36:32Speaker 01

  I did my best.

• 36:32Speaker 00

  Thank you. I figured you'd be able to help me on that one, because I had the rough idea that it would introduce error, but I had no idea to the extent that divide and multiply may not make any sort of actual real difference in your program. Plus, all the crappy things I do anyways, probably destroy my program's performance long before we get to this loop.

• 36:49Speaker 02

  Yeah, prime, we've got other issues before we get to whether we're multiplying or dividing, I think, yeah.

• 36:55Speaker 00

  I just really want to take my Lua program seriously, TJ, okay?

• 36:58Speaker 01

  It's usually, the one thing I can put in someone's head that's generally not wrong is, don't confuse integer multiply with floating point multiply. Floating point multiplies are much faster than integer multiplies typically. And so some of the time when you see people saying, oh my God, like divide is so bad. Sometimes they're thinking about integer divide. And there's a lot of reasons, there are a lot of reasons why integer divide is bad. A lot of reasons.

• 37:26Speaker 00

  I'm gonna put that on the blocker list for next week. You can, we can, we can talk about that because that's actually, that seems very interesting.

• 37:32Speaker 01

  Okay.

• 37:34Speaker 03

  Is it on?

• 37:35Speaker 01

  The pause, the pregnant pause.

• 37:37Speaker 03

  Yeah, I was like.

• 37:39Speaker 00

  Sorry, I was, I was just putting it on. I thought Casey would finish anything. Now, hold on, can I, can I just jump in for a second? Can I jump in here? I'm gonna say that this instruction thing is done. Trash, I had, TJ said you had a blocker that you need to discuss. I don't know what it is.

• 37:51Speaker 03

  Blocker, Trash, that we needed to get out to the world.

• 37:54Speaker 00

  What is this blocker that you had to talk about?

• 37:57Speaker 03

  Oh, dude, I'm locked out of my GoDaddy account that I set up like a decade ago. And I got hit with like a renewal, whatever. And I tried to get in and apparently it's using like my college email, which got deactivated. God knows when. So then I hit up, I hit up like their chat bot, not chat bot, but there was like apparently a supposedly a real person.

• 38:18Speaker 01

  Yeah.

• 38:19Speaker 03

  And he just kept asking me, show me a screenshot of the error. And I was like, dude, I'm locked out of my account. I just need to change my email. He's like, oh, show me a screenshot of the account. And I was just, I got to the point where I just got so fed up because I was trying to do it for like an hour. I went on Twitter and then I just like tagged him. Now I'm like making some progress here because I want to get a refund for my like domain renewal. Because before I was like a programmer, I did photography. So it was a photography domain.

• 38:48Speaker 01

  Of course you did.

• 38:49Speaker 00

  Of course you did.

• 38:50Speaker 02

  Save that for next week's fun fact, bro. You just spoiled it, dude.

• 38:53Speaker 03

  Oh, fine, whatever. But anyways, I'm just trying to get rid of that, that domain and get my money back. It was like 200 bucks or something, which is insane.

• 39:01Speaker 01

  It's Pokemon cards.

• 39:03Speaker 00

  I can't believe you could have something renew a decade later successfully. Like, how, did you keep hold of some sort of long running credit card? Like, what do you got on there?

• 39:11Speaker 03

  I don't know. I just got a random email. I was like, this domain's gonna renew. And I was like, huh, I haven't seen that.

• 39:15Speaker 01

  I was expecting like Trash to be like, well, before I was a photographer, I was in fintech. And so I had this.

• 39:24Speaker 02

  I started a credit card company.

• 39:27Speaker 01

  Before I was an archaeologist, I used to do deep sea diving. The time where you hold your breath.

• 39:33Speaker 03

  Yeah, I got some other things. I got some other things up my sleeve. But yeah, that's my blocker. Hopefully, GoDaddy, if you're watching, I want my refund and I need my email changed ASAP, please.

• 39:45Speaker 00

  Nice. I like it. Alright, well, we've now finally hit the part where I can introduce the main topic, everybody. Today's stand-up. Welcome to the stand-up. Today, we're going to be talking about Nvidia and Microsoft, talking about what they're releasing. I got some hot takes on Microsoft. I didn't tell Casey long enough beforehand, so he probably has zero hot takes on what's happening, but guess what? We're still gonna talk about it in today's stand-up. Uh, one thing I'm truly disappointed by with today's stand-up introduction is that TJ has an in-progress task called Interrupt Prime's Intro, and he did not, in fact, interrupt my intro. I am now done. TJ, what have you done?

• 40:25Speaker 02

  See, in the business, that's what we like to call power play. And I just made you do the whole intro, waiting for me to interrupt the intro. And I didn't even have to do anything.

• 40:40Speaker 03

  Living in a diner, free.

• 40:41Speaker 02

  In progress and I did nothing.

• 40:46Speaker 00

  Anyways, that actually sounds like your daily life, doing nothing. Okay, so here we go. Got him. Alright, so I think the first thing we have to talk about is this Microsoft Surface. Have, has everybody seen the Microsoft Surface?

• 41:01Speaker 01

  Yeah. Oh, you mean in the 15 minutes between when you sent out the invitation and us recording this? Did we see the mic? Is that what you're talking about?

• 41:08Speaker 00

  Yes, yes, yeah, yeah, that one. That specific one.

• 41:11Speaker 03

  That's the one that has the, the DGX in there, or the RTX in there, right?

• 41:15Speaker 00

  Yeah, the RTX Spark. And so they're saying it has one petaflop, which by the way, tossing out one petaflop? I, I don't know what that actually means.

• 41:23Speaker 03

  Is that a real word?

• 41:25Speaker 00

  Yes, petaflop, yes. I petaflop too if it looks good enough. But here's the deal, is that...

• 41:29Speaker 02

  I think it's for how badly the product's gonna fail. It's gonna be a petaflop.

• 41:35Speaker 00

  It's a petaflop. Uh, so...

• 41:37Speaker 01

  It's FP4 Tensor, Tensor Ops, I'm, I'm pretty sure.

• 41:41Speaker 00

  Okay, you just said a bunch of words at me?

• 41:46Speaker 00

  Like, I should understand that? I just think it just means... It's... What is that? That's 1 above Terra, right? So that's like 12, e to the 12 amount of floating-point operations? Per second?

• 41:58Speaker 01

  Yeah, so, okay.

• 42:01Speaker 00

  What'd you say?

• 42:03Speaker 01

  I said FP4... So, so 4-bit floating-point, right? FP4 tensor ops. So, meaning these have to go through the tensor core, like, largest, largest dimension matmul.

• 42:19Speaker 00

  Okay.

• 42:20Speaker 01

  I'm probably just speaking gibberish at this point. So, we had this big, long discussion, coincidentally, about floating-point just moments ago, and we were talking about that whole exponent mantissa, whatever thing. So, in the quest to quote higher and higher numbers, GPU vendors have been making progressively smaller and smaller floating-point types. So, we have 32-bit floating-point. They made 16-bit floating-point. That was, that's been around for a while, actually. Different kinds of...

• 42:52Speaker 00

  What do they call a 16-bit floating-point? Because I know, they, they call them doubles? Floats for 32 back in the day, and doubles for 64. What's, what's a 16?

• 43:02Speaker 01

  They have various names because there's different encodings. So, there's like bfloat16, and I don't know. Like, I, I don't use this stuff, so I have no idea what the, what the... There were gra, there were 16-bit floating-point numbers for graphics, which were used for things like, oh, you know, we can, this frame buffer that we're storing results in. We'd like to be able to store lighting results in something better than 8-bit color, right? So, we're gonna go up to 16-bit color, where we store that as a 16-bit floating-point number, because we don't want to go all the way up to 32-bit floats, because that's a lot more memory bandwidth, and blah, blah, blah, blah, blah, right? You can imagine all the things that go downstream with this.

• 43:38Speaker 00

  Yeah.

• 43:38Speaker 01

  Stuff like that. So, 16-bit floating-point is nothing new, although the particular type of floating-point you use can change. Like, how many bits you dedicate to each thing, right? Because remember, we talked about exponent mantissa? There's like, how many bits do you dedicate to scale? And how many bits you dedicate to the actual digits is up to you in something like this, because you can move, move that trade-off around, depending on what you were, whether you wanted more range, or whether you wanted more digits in your number, right? Um, so, uh, yeah, so we went down to 16, then there was eight. So 8-bit floating point numbers, again, very, very small amount of information, but you know, still doing it. Right. And now they're, they're actually at 4. So they actually do 4-bit floating point, which I don't even, since I don't do AI.

• 44:23Speaker 00

  I think they have 1.5 bit. Isn't that like the, the AI super quantized models, 1.5 bit, uh, numbers representing?

• 44:30Speaker 01

  I, I don't know. I don't, I don't look at AI stuff, so I'm not sure. All I can tell you is I believe this is an FP4 number. That's all I know.

• 44:37Speaker 00

  Okay.

• 44:38Speaker 01

  And that could be wrong, because like I said, I only had like a very brief amount of time to even see what the heck this announcement is we're supposed to be talking about. So everyone out there, take literally everything I'm about to say with a pretty big grain of salt. I mean, always a good idea anyway, anytime anyone's talking these days, take it with a big grain of salt, but, but an even bigger grain than you ordinarily would have shaken in, out of the shake, salt shaker into your mouth. So the way that they quote these numbers now, I believe, because they're marketing numbers, is they take the smallest possible thing they can do on the chip. So if your chip does FP4, you're quoting FP4 numbers, right?

• 45:12Speaker 00

  Correct.

• 45:13Speaker 01

  And you're, and you may ask why. The reason is because all GPUs, they are wide OPS. They're like SIMD OPS, single instruction, multiple data. The, the lane width is typically, depending on how you want to consider it, either 32 or 128, depending on how you look at it, um, in terms of the, uh, the, the sort of the, the size, uh, the, the bytes that you're going to operate on at a particular time. So when you talk about how many operations you can do, they're not talking about how many like instruction operations, like I told this thing to multiply. They're saying, when I tell it to multiply, it's going to take this giant byte pattern

• 45:56Speaker 01

  with a ton of stuff packed into it and do this big multiply, right?

• 46:02Speaker 01

  So they take that and they say, okay, how many of these can we pack in there? Take the smallest thing, which is SP4, right? Take the most efficient thing we have, which is when we multiply these two things together, what instruction can do the most of these multiplies in the shortest amount of time, right? And then that is the number that they quote you. And just for good measure, because typically what you're doing is F Mads, floating point, multiply, add, right? It's a, it's a multiply and an add in one step. They, just for good measure, multiply the number by two, because they're like, well, hey, you're doing two operations. So technically, it's twice as much, right?

• 46:46Speaker 01

  So you have to do all of this to figure out the actual instruction rate, right? But yeah, so that's what's happening. Sometimes they quote dense, sometimes they quote sparse. I don't know whether this is a dense or a sparse number. I didn't have time to go look at the math to see. You can tell pretty easily by just looking at the CUDA core count of the, of the thing and the clock rate. So that's, it's not hard to figure it out. But, and the way that that works is these kinds of cores have this ability to effectively pack data, sort of, pack and unpack data as part of the operation. So that if you have something where you know there's a lot of zeros in it, like it's a sparse, like it's doesn't have, not all the numbers are always filled in, they can exploit

• 47:30Speaker 01

  that sparseness and get higher throughput by just the chip actually seeing the sparseness and only issuing the multiplies it actually needs. So it can basically fit like more, you know, it can pack the zeros out. So one of those big 128 byte lanes can hold more than 128 values because the zeros have been eliminated or something, right? There's all these kinds of things they do in there and I don't study this. So I don't study the specifics of that and I don't know which number this would be, whether it's a dense or sparse. But that's everything, anything where they quote these numbers, all those shenanigans are on the table.

• 48:08Speaker 00

  Okay, this makes a lot more sense because that number, you know, in my head, I'm thinking, okay, this is like, you know, uh, what's called architecture-wide floating-point operation, so they're doing like double, like a pet a flop of doubles. I was like, this seems like that's really fat. Like, in my head, that's like supercomputer fast, so I must be misunderstanding what we're talking about at this point. Because then I went on the internets, you know, as I come and do from time to time, and I started asking like, I went to the AIs, of course, to ask about the AIs, like, with the new Spark thing, like, how, how fast are these models actually being able to be ran? Because it looks like it can be ran at like, super fast, right? You could have like, local models running at infinity speed. And a lot of the forums slash what the AI had to say was saying, you're gonna be running at somewhere between like,

• 48:50Speaker 00

  eight to 16 tokens a second, which would be on a 32, on a 32B model. So they're not, the speed and all that, I don't, I don't know actually what these speeds mean, because that just seems like the ultra-fast number I've ever seen in my lifetime, but really, it may not be nearly as fast as, as, at least my brain would assume that means.

• 49:10Speaker 01

  Well, again, I mean, so if you look at the actual chip, like, like what's in this thing, right? Um, it's basically like the same number of CUDA cores as like a 5070, right?

• 49:23Speaker 00

  Nice. Um, and so- It's like it could run Fortnite really good, so I assume it was a pretty good, you know, graphics unit.

• 49:31Speaker 01

  Well, I mean, maybe. The other problem, though, is it's, so the memory is gonna be different, right? Because, uh, this thing, this thing is LPDDR, uh, so low power, um, the stuff that's in laptops. And so if you compare it to something like a 5070, a 5070 is using GDDR, which is the like, the high clock rate, high power consumption, high clock rate.

• 49:56Speaker 00

  The gains.

• 49:56Speaker 01

  Right? Um, and so the trade-off, the actual thing that you have to think about here is- A 5070 has like three times the memory bandwidth or something of this thing. So this thing is actually a third the speed of getting you the actual memory. So I don't know how good it would be at Fortnite, maybe, fine, I'm not sure. But it's definitely worse than a 5070 at Fortnite. So you're kind of doing a trade. You're trading the fact that this thing will have much worse memory bandwidth. I mean, a third of the memory bandwidth is not good, right? Which I think is roughly what it would get. Again, take it with a grain of salt because we had all of zero seconds for me to look at this stuff. But the trade is that because this is using LPDDR, right?

• 50:42Speaker 01

  It can have a ton of memory, can have 128 gigabytes of memory. Whereas a RTX 5070 is like, I don't know, it was like 12 gigabytes. It's like, it's like a tenth of the memory. So you're making a trade and I assume that this is a trade that AI people are willing to make because

• 51:00Speaker 01

  although it'll be slower, so evaluating evaluating any particular part of your model in theory would probably be slower on this chip because it's got one third of the memory bandwidth for pulling it in. So it can only go at one third the speed of pulling it in if it's memory bound, which again, I don't study this stuff, but my understanding is that AI workloads often can get memory bound, depending on what you're doing. But on the flip side, you can fit a much larger model in there, right? You can fit 10 times larger model. So I'm assuming that's a trade that AI people would be happy with. Otherwise, why did they make this thing, right?

• 51:38Speaker 00

  Can I ask a dumb question really quickly?

• 51:40Speaker 01

  I probably can't answer if it has to do with AI.

• 51:42Speaker 00

  No, it has to do with memory and specifically why you said like one tenth the memory. Why does the low power enable more memory versus the high power counterpart? Is it purely just a power draw problem or is there something I don't understand?

• 51:58Speaker 01

  I really couldn't tell you. So I would assume that it is probably more to do with cost than anything else and/or market segmentation, meaning I don't think there's anything in particular that's... Well, you know there isn't because you could go buy a 5090, right? And what does that have? 24GB, 32GB?

• 52:20Speaker 00

  I assume it's 24.

• 52:21Speaker 01

  I don't remember what they have, but they've got way more. And that's on a consumer-level card. You're not talking about buying like a Blackwell for the data center.

• 52:26Speaker 00

  It's a 32.

• 52:28Speaker 01

  32, right? So obviously, you could have stuck more GDDR on there. It's just the cost is higher and the, you know, so I think they're trying to get 128GB into something that doesn't cost you 10 grand. If you wanted to go pay data center prices, you can get that. I mean, you can get high bandwidth memory for that matter. HBM modules on their like actual data center parts, they can do it, right? So it's really, I think it's just a cost issue. In terms of the manufacturing of these things, I imagine GDDR is just a lot more expensive than LPDDR. That'd be my guess, but I don't know.

• 53:04Speaker 02

  Probably a little bit for laptop form factor, too. I mean, it's just like, you'd rather not have necess... Like, I don't know how much power draw, but I imagine

• 53:15Speaker 01

  Right.

• 53:16Speaker 02

  Going faster means more power, more heat, bad battery life, etc.

• 53:22Speaker 00

  TJ doesn't really concern himself with battery life. I don't know if you've seen his pictures, but he has a battery life that he could run for days on a laptop.

• 53:28Speaker 02

  Well, I didn't show, I didn't post that picture, but I can grab it if you want, probably.

• 53:33Speaker 00

  It's a pretty good picture. So, Casey, I did verify the RTX Spark has 300GB a second for the unified memory system, whereas the 5070 has 670 or something like that.

• 53:44Speaker 01

  Yeah.

• 53:45Speaker 00

  It has over double of whatever it is. And then the 5090 has like 1.7TB.

• 53:50Speaker 01

  300 sounds high for the Spark. Where are you getting that from?

• 53:53Speaker 00

  Uh, this is, this is the summary from an Instagram Verge post, and then Tom's hardware as well as reporting, apparently, that exact number.

• 54:04Speaker 01

  Um, interesting.

• 54:05Speaker 00

  Right here. So Tom's hardware is saying it's 300GB a second, right here.

• 54:09Speaker 01

  Because it looks like 273 on the, oh, mm.

• 54:14Speaker 00

  Okay, so they may have done a little bit of rounding.

• 54:17Speaker 01

  It's like 273 versus 672. So, I mean, yeah, I, maybe, maybe saying a third is, is being a little bit harsh, but I wouldn't say double. So it's more than double.

• 54:31Speaker 01

  Let's take a look at what it actually is if people want to get that, uh, that specific about it.

• 54:40Speaker 01

  2.5, so split in the middle, like.

• 54:43Speaker 00

  Split the baby.

• 54:45Speaker 01

  So 2.5. And again, that, that's the trade, right? You're just trading like, okay, we got, you know, we're gonna have less speed at pulling this stuff in from memory, but you're gonna get a lot more memory and presumably that's what you want. Otherwise, again, I don't know why they would build this, right? It, presumably that's what people want. Because if it wasn't, I, then what's, I don't know what the pitch for this thing is, right?

• 55:06Speaker 00

  So I, I think, I, I think I understand more of what this pitch is going to be and kind of the general direction Microsoft's taking, is that they kind of, he said a very interesting thing, which I just wanted to bring up in here.

• 55:16Speaker 01

  Who's he?

• 55:17Speaker 00

  Well, I'm about to show you. His name is Satya. And he said the following. He said, "Our goal is to deliver un-metered intelligence to every home and every desk with Windows." And then they do this RTX Spark, and part of RTX Spark is that you can run a bunch of local models yourself. So you don't have to necessarily rely on the cloud models. You can run models yourself as the general kind of tenor or goal, it seems like, for where Windows is going. And, uh, there's a very specific phrase in there, which is "un-metered intelligence." Uh, Sam Altman, for the last two years, has been saying that he envisions OpenAI to be a metered intelligence. It should be no different than your electricity, all this. So this kind of seems like, hey, they're kind of going against this general idea. And instead, they're moving a lot of it down into the individual's laptop and then building harnesses.

• 56:02Speaker 00

  So my personal guess is that what Windows is wanting to do with these larger memories, but a bit slower, is that you can run a bunch of these smaller models to do mundane tasks, simple tasks. And then have a cloud subscription to run a model up in the cloud that doesn't need nearly as much power because they're just telling little models what to do. And you're running like four models in parallel to go and do these basic operations. And so thus, that's why they're fine with these, say, lower bandwidth memory, these four to 10 tokens per second kind of production. Because it's not really, they don't really have to care about that because they have multiple going and they have the big powerful one up in the cloud. And they're just trying to win the basic consumer by reducing costs and reducing data center.

• 56:45Speaker 00

  So that's, that's what I think is, is generally happening here. And that's why they've chosen this general architecture.

• 56:52Speaker 02

  Well, I, I mean, I got two, two things that'd be fun to chat about for it. I think the, the first one is, is, is it not a little bit weird? Like that it feels like a direct shot at OpenAI that like Microsoft has a really big stake in and has spent a lot of money on it. It feels like, like Sam's things been the, it's like, it's too cheap to meter or like it should be treated like utility. Or we should like have UBI and tokens. That's, they kind of gave up on that train, but that was, I recall talking about it on, on the pod. So I'm interested to hear, you know, sort of what people are thinking for like, is there some trouble in paradise over there? What's going on? You know?

• 57:33Speaker 00

  Well, we know, we do know from the court leaks that Satya was a major reason why Sam, you know, held on to a CEO ship of OpenAI. And so there is something, it's, it is very interesting that he's using very, because we all know that Satya does not tweet a tweet that has not been looked at by human resource.

• 57:52Speaker 02

  Especially not this one.

• 57:53Speaker 00

  By, yeah, especially not this one. Like they've really, this is a very intentional tweet, which makes it seem very intentionally against whatever that vision is of the world. Which to me, you're, I think. I think part of this is also in response to like the data center hate is that generally speaking, people seem to be very anti data center in our current in our current world, whether it's founded for good reasons or bad reasons, people generally are. And so having the ability to have agents run on your computer as opposed to up in the cloud means less data centers. Thus, it feels more tenable, I guess. And so my two guesses is that there is trouble in paradise when it comes to this because Windows has also created their own frontier model. And so there's like 900 things going on to where it seems like they've went from

• 58:39Speaker 00

  competing like competing against the other models using OpenAI to saying now we're competing against OpenAI with our own models. And we're like directly going against it. And their contracts, I think, run out 2032. So maybe there's something about that.

• 58:53Speaker 01

  Trash, thoughts?

• 58:55Speaker 03

  No, I was gonna say, I'm actually, I've been thinking about doing local AI stuff. We have like a local AI group that I'm part of, and they all have like these like DGS sparks are just way too expensive. So I feel like this was very good timing. Wait, wait, Trash, turn around. What?

• 59:12Speaker 03

  What?

• 59:12Speaker 00

  Yeah, too expensive.

• 59:13Speaker 03

  You said too expensive. Dude, DJX sparks is like 5,000 or like almost $5,000. It's crazy.

• 59:20Speaker 01

  Trash, do you know how many Pokemon cards I could buy?

• 59:23Speaker 03

  It's it's a fraction of this, but

• 59:30Speaker 03

  no, no, but like to be serious here for a second, I think it's pretty neat. The one thing I don't like is like, there's no Linux support. Now I'm a Linux knob now, thanks to the last

• 59:41Speaker 03

  I was like, I was like, no Linux support. I'm not even take a part of this, part of this show right now. But apparently it might be coming. But I think the interesting thing is, I'm curious how, because I already know a lot of these companies are underwater with as far as revenue goes. So if you're, if I have this computer, I don't need to subscribe to ChatGPT anymore. I can do my mundane tasks. I feel like most consumers aren't doing anything crazy besides actual enthusiasts and developers. So your average consumer doesn't need that like crazy power. So they can run a Qwen-32 model on their machine and get away with it. And then if they're really, I don't know much about what Microsoft's releasing with their seven frontier models. I'm assuming they're shoving that down their employees' throats right now, and it probably sucks.

• 1:00:22Speaker 00

  They did cancel block code, so.

• 1:00:24Speaker 03

  Yeah, exactly. I'm assuming that's, that's definitely like leading up to this moment.

• 1:00:30Speaker 03

  But anyways, I'm just curious how it's gonna work out for Anthropic and ChatGPT. Cause I know like the big thing was like, we need more compute. We don't have enough compute, but now everyone has it on their machine now. That's no longer gonna be an issue. And also, you're gonna lose all those subscribers. And that's the next issue. So I'm just curious how ChatGPT is gonna like respond to this? And if they're gonna think like they're potentially in trouble here. I don't know.

• 1:00:53Speaker 02

  Yeah, the other, the other trash which you mentioned, which I was kind of interested in is I'm wondering how much of this is like a. You know how when you, someone says no to you, so you just reframe it like it was always your idea? And that you're really happy that you can't do option A and now you have to do option B? Like, I do wonder if some of it is like, hey, we thought we'd be able to like build all these data centers and everyone would be like stoked about it and we'd get to like do all that stuff? And then that doesn't seem like it's gonna be possible. They can't get the electricity. They can't get anything else? And so is, are they doing like a pivot to try and be the first, uh, like, you know? First past the finish line on these sort of like home AI setup things because they can't get

• 1:01:37Speaker 02

  the data center stuff approved, right? You know what I'm saying? Like in terms of like, well, we need to get AI'd everybody, obviously. Uh, but we can't do the data center thing, you know, is.

• 1:01:46Speaker 03

  I'm sure that's a part of it, but I, I feel like they're probably working on this well before like the hate on data centers potentially.

• 1:01:53Speaker 02

  Well, some of these have been, but like the packaging is very different, like packaging it inside of a Surface laptop. Like they've been building all these components, but putting it together in one is interesting.

• 1:02:04Speaker 03

  Yeah, yeah. I, I'm curious how it's gonna shake out. I don't know that. I don't think they mentioned prices. Did they, Prime?

• 1:02:11Speaker 00

  I don't, I didn't see the prices. I have, I watched.

• 1:02:15Speaker 00

  I, I watched the keynote and the keynote didn't seem to mention prices or at, at least that I remember.

• 1:02:19Speaker 03

  I couldn't find prices anywhere. I'm assuming it's just gonna be kind of insane.

• 1:02:24Speaker 00

  Hundred twenty-eight gigs of RAM? Even if it is low, low power along with a fifty, seventy? Come on, come on.

• 1:02:32Speaker 01

  Well, I mean, I can't, is like.

• 1:02:36Speaker 01

  What would even be the explanation for how it would be cheaper than a DGX? Like, the thing that Trash doesn't want to buy because it costs too many Pokémon cards. Like, you take some of the networking out of that thing, but now you gotta pay for a screen and a keyboard and chassis and all this stuff that you didn't have to pay for before, right? And a battery. So how does it get cheaper? Like, it seems like it will just cost the same as the thing you can already buy, right? I mean, I don't know. I don't know why this would be cheaper.

• 1:03:06Speaker 00

  I'm gonna throw this out there. Is that there's a huge amount of people that are just Windows Andys, shall we say? They're not really into the computer side of the computing. They don't really want... Because if you've ever played with local models, at least last time I did it last year, there's a lot more work to it than simply, you know, plug it in. Hey, I'm up and running. And I think that's what Microsoft is selling. Is this really kind of like, hey, oh, you're just in Excel. You're just in these things. Well, here's our integrated. It runs locally, so it's fast. You don't even have to pay money. You just have to, hey, just buy a Microsoft 365 and pay the once-a-year fee for that, and you're good to go. It runs all locally. It's all yours, right? And so to me, they're making a play to purchase people in by saying it's free, effectively.

• 1:03:50Speaker 00

  But do you have to buy all their services anyway? So they're like, still winning in the end. It's just giving people the access.

• 1:03:56Speaker 01

  Well, I'm just pointing out, for this particular product, I'm not really seeing how this comes in at a price point below, you know, even $3,000, possibly higher. So it doesn't seem like... Like, this is not something very many people are going to buy, I wouldn't think.

• 1:04:11Speaker 00

  It's geared towards programmers predominantly. Like, that's what, at least their presentation, really made it seem like a programmer thing. The person that did the examples talked about how they got all the, like, effectively Unix tools, like grep is now in Windows. All these things are now in Windows. They showed they have like 53 tools or something like that. And so it did seem like it's geared towards the working professional as a working professional computer.

• 1:04:33Speaker 01

  Right, and that's why I was saying, like, so this isn't really something that you're going to sell to, you know, somebody who just runs Windows and doesn't know anything, right? Like, it's not really targeted... At them, it's only targeted at people who already know what they're doing and are using the command line and stuff, right? I think.

• 1:04:49Speaker 02

  Casey, you're not thinking about this, you're not thinking about this in the right way. Okay? So here, we were talking about TAM earlier, obviously the TAM for this podcast and the company that it represents, everything that's ever lived.

• 1:05:03Speaker 01

  Yeah.

• 1:05:03Speaker 02

  Um, that coincidentally, everything that's ever lived, they can be programmers now, Casey, because of AI.

• 1:05:09Speaker 01

  Right.

• 1:05:10Speaker 02

  You haven't even thought of that, probably.

• 1:05:11Speaker 01

  True.

• 1:05:12Speaker 02

  So, everybody at the company is a programmer now, in 2026. So you just have to convince the CEO of the company that since everyone's a programmer, they need a programmer computer, and therefore, they need to get this computer for every single person at the company so that they can run it.

• 1:05:31Speaker 00

  Hey, Casey, do you want a $500 million token bill? Or would you rather just pay, let's just say a million dollars for a few laptops? And then it's all run locally? What do you think, Casey? Well, I'm just saying. Hey, I'm just saying.

• 1:05:44Speaker 01

  Laptops or Prime, a thing that looks like a griddle that you can, that you can get. Did you see that?

• 1:05:50Speaker 02

  I saw the people's doing the picture of the cooking on it.

• 1:05:53Speaker 01

  It looks like a hot plate that you could cook on, which I think is kind of fun. Like, can you make an egg on this thing by putting a, like, a piece of, like, Teflon on top of it? I want to know.

• 1:06:02Speaker 00

  Uh, it did, it actually did remind me very, I, I bet, I bet you I even have one in here, an Xbox One X, the top to the Xbox One X, they just lifted the design of that and just put it as the, like, the grit, the grading or whatever, the grit on top of, the griddling on top of this one.

• 1:06:21Speaker 01

  Here's what I want.

• 1:06:22Speaker 02

  Yeah, hit me.

• 1:06:23Speaker 01

  And I'm dead serious. I want someone to take one of these things, the griddle one, I want them to put a solid metal piece in place of the waffle pattern, and then I want them to create a thing called Bacon Mark. And Bacon Mark is basically an AI throughput benchmark, where you put a piece of bacon onto it, and then the benchmark result is just a picture of the bacon after the workload has run.

• 1:06:54Speaker 00

  Okay, okay.

• 1:06:55Speaker 02

  Is it, and how do you determine if it's a good result? That the bacon's still raw?

• 1:06:59Speaker 03

  Literally.

• 1:07:00Speaker 02

  The rawer the bacon, the better the result?

• 1:07:02Speaker 03

  Yes.

• 1:07:03Speaker 01

  Yes.

• 1:07:03Speaker 02

  Right?

• 1:07:04Speaker 01

  I mean, maybe, or, or, i- i- i- but yes, like, the, if the, if the thing computes the same, if the, if it's computing the same answer, I guess, then the m- the more, like, intact, the more virginal the bacon.

• 1:07:16Speaker 03

  It should be.

• 1:07:16Speaker 01

  The more efficient this AI is.

• 1:07:19Speaker 02

  Unfortunately, Casey, this, we did miss the boat on this one. This would have been perfect, perfect for, uh, Apple to release, because then we could have called it Let Tim Cook. Okay? That would have been the perfect.

• 1:07:30Speaker 01

  Too late. Too late. He's gone.

• 1:07:32Speaker 02

  Gone, it's past, it's over.

• 1:07:35Speaker 03

  Yeah.

• 1:07:35Speaker 01

  Yep.

• 1:07:36Speaker 00

  Unfortunately, they did let Tim cook, and he got kicked out of the kitchen.

• 1:07:41Speaker 03

  He couldn't handle the heat.

• 1:07:42Speaker 01

  His cooking days are over.

• 1:07:43Speaker 00

  Couldn't handle the heat.

• 1:07:47Speaker 00

  Huh. I, I do think this is really interesting, though. I am very, I'm very curious where this is going. Uh, because it, there is something that's very, like, unique about this moment, that Windows did have a three and a half hour keynote, and this was obviously like the, kind of the centerfold of this. I'm not sure if that's the appropriate term, now that I think about it, because I do remember that song.

• 1:08:06Speaker 01

  Yeah, let's, soft to the centerfold. Do you really want that, Prime?

• 1:08:08Speaker 00

  Yeah, I just started replaying some 80's music in my head. Uh, I don't think I can undo that. I can't, I literally can't put the centerfold back in the magazine at this point, so we're just gonna keep on going.

• 1:08:18Speaker 01

  Okay.

• 1:08:18Speaker 00

  And, uh, someone mentioned that they never mentioned any of their pro- any other programming apparatus, other than AI, except for VS Code once. And so, like, they are fully going towards the pro- the, the programmer, but they're going towards them exclusively through model usage. Like, there, they don't talk about TypeScript, .NET. Like, .NET, they've always talked about .NET. They're not talking about .NET anymore. They're not talking about C sharp, they're talking about models.

• 1:08:48Speaker 01

  AI.

• 1:08:48Speaker 00

  AI.

• 1:08:49Speaker 03

  I'm curious how those models are gonna be. One of, I said their flagship one is like equivalent to 4.8 or something?

• 1:08:56Speaker 02

  It honestly enhances you a lot and gets distracted.

• 1:09:01Speaker 00

  Goes on X and tells me about things that are happening.

• 1:09:04Speaker 02

  Okay, I did, I do have to say, so Casey 4.8, Opus 4.8 came out, so I gave it a test run. And it was hilarious because the very first message it sent back to me, I haven't been using Claude models in quite a while, the very first message goes back, it goes, excellent insight, that, that's straight to the point. And I'm like, ah, bro, bro, put the code in the bag, bro, stop bending my tokens, not telling me how smart I am. I already know that, okay?

• 1:09:34Speaker 01

  Glaze, gotta get a little glaze in there. Gotta, gotta get a little glaze on that code, doughnut.

• 1:09:39Speaker 02

  It was crazy. Immediately, it was like so obvious. And I was like, this is why people are falling in love with Anthropic. It tells them how smart they are every time they send something to it. Oh, man.

• 1:09:54Speaker 00

  Yeah, as far as models go, as like which one's better, honestly, at this point in my life, I can't really like, besides for like some kind of weird implementation details, like some will create many tests, some will create one test, like there's like kind of these weird like kind of variations in that. As far as performance goes, I'm having a harder and harder time disambiguating which one is better, right? Like I can't, if you were just show me a function, a snippet of code, it'd be hard to understand which one's better. So I'm sure the Microsoft one does fine, right? Like I'm sure you wouldn't even, you would probably not be able to understand at a, at a more micro level, the difference.

• 1:10:30Speaker 01

  I could also imagine, um, if I used my powers of imagination to really imagine in an imaginative way.

• 1:10:38Speaker 00

  Yeah.

• 1:10:38Speaker 02

  Yeah.

• 1:10:39Speaker 01

  Uh, that if your goal is to program on Windows, which at this point, I don't know.

• 1:10:46Speaker 00

  That's the imagination part, Casey. Okay.

• 1:10:47Speaker 01

  That's the imagination part. Microsoft obviously has a huge history of source code control check-ins to Windows code bases that nobody else has because they're all internal, they weren't done publicly. And so their AI is presumably trained on all of the entire development of the NT code base and all the development of the Windows 95 code base and all of the development of Microsoft. Office and all that stuff, uh, which nobody has as a training data set, except people who maybe like hack sorted out of Microsoft or were paid to smuggle it or whatever happens in AI circles now. I'm sure it's all very illegal and that they're all fine with it because they've convinced themselves of something. But I would imagine that Microsoft is probably the only people who can train on that.

• 1:11:31Speaker 01

  Unless OpenAI got the rights to train on that as part of one of their, you know, sleazy backroom deals that they do or whatever. However, I would just like to bring this back to

• 1:11:43Speaker 01

  another point, which is, again, I think it's worth underscoring that this is a non-announcement. Like the only thing that's actually like, as far as hardware is concerned, this is just an announcement that Microsoft is like getting Windows to run on this thing. Right? Like, that's it. Because there's no, it's just a spark, like, there, I guess you could say that the laptop part is a form factor announcement. It's like someone's gonna put this, this thing you could have bought from Nvidia since last year, right? We're putting it into a, into a laptop for you if you prefer a laptop. Like, that's the hardware announcement. There isn't another thing right there. It's not a new chip. It doesn't have new capabilities. That's it.

• 1:12:28Speaker 01

  So really, the announcement is more of a software announcement. It's like Microsoft got Windows on ARM running on this thing. That's really the announcement, right? I mean, and, and laptop form factor. So, so I make sure that that's got like, there is no hardware. There's no chip thing here, right? I mean, are we all?

• 1:12:45Speaker 02

  Yeah.

• 1:12:45Speaker 00

  I think the chip thing is just like how small and integrated it is, right? With unified memory as opposed to having its own dedicated V RAM, right? Isn't that like the?

• 1:12:55Speaker 01

  No, that's the same, that's the thing Trash was talking about that he didn't want to buy because it was too expensive. It's the same chip, same, same memory.

• 1:13:02Speaker 00

  I'm an idiot, sorry.

• 1:13:04Speaker 01

  So it's true, we took, we took this thing that you could have bought since last year, and we're putting it into a different thing now. That's the, that's it for the hardware side of things, as far as I can tell. I didn't see any announcement of like, a new, actual capability, or a more modernized chip, right? So it's really more of a software announcement, like the, the, the, the actual new thing here, as in terms of what's gonna change for you, if you, instead of buying one of these Nvidia little desktop boxes. Uh, we'll call it a Mac Mini Compete, or whatever. Instead of buying one of those, uh, the only difference really is that now you'll be able to run Windows, right? You'll be, like, before, you couldn't, you couldn't run Windows on those, I don't think

• 1:13:46Speaker 01

  Windows on ARM worked on that, on that little Nvidia box. But it's the same box, just in a different form factor now.

• 1:13:54Speaker 03

  It's a nice little skinny laptop.

• 1:13:56Speaker 02

  And, Casey, correct me if I'm wrong, this was, there was some sort of, like, Jonathan Blow is tweeting, like, that, what if it was like, a completely different style of computer, right? That was, like, in terms of the things of, like-

• 1:14:11Speaker 01

  Yeah, and it definitely isn't that, like-

• 1:14:13Speaker 02

  Right, there were the mysterious- The mysterious tweets, and all of these secret little signs and stuff, and you're like, oh wow, they're gonna release, like, a GPU computer, or something insane, right?

• 1:14:24Speaker 01

  No, yeah. Whatever John Blow wanted, he definitely didn't get it, because literally the chip, because literally the chip is the same chip that you could have bought, it last October, right? So, so, there was no, like, they didn't announce any new, like, architectural, interesting things, where John would have been like, oh, that's a step towards the thing I wanted, like, no, there was no step, at all.

• 1:14:47Speaker 02

  Well, 'cause they were doing the mysterious, like, oh, Nvidia sent the date, and Microsoft sent the date, and you're like, whoa.

• 1:14:53Speaker 01

  No.

• 1:14:54Speaker 02

  Is this gonna be, like, the first time-

• 1:14:55Speaker 00

  It was a location in Taiwan, if I'm not mistaken.

• 1:14:58Speaker 01

  It was the location, it was the location of the, uh, it was the, the GPS coordinates of where they were gonna give the keynote, right?

• 1:15:04Speaker 02

  Right, yeah, yeah, yeah. So you thought it was gonna be like, whoa, I bet their stock's gonna double after this, like, it's gonna change the game.

• 1:15:11Speaker 03

  We'll have to look at the stock, actually.

• 1:15:13Speaker 02

  Oh yeah, trash. Let us know.

• 1:15:15Speaker 01

  Yeah, so that's the, the long and short is if you want this thing early, just go buy an Nvidia, one of those little Nvidia Spark boxes, whatever they called it, the RTX Spark or something. I don't remember what it was called.

• 1:15:25Speaker 02

  Yeah.

• 1:15:26Speaker 01

  The, just go, you can literally buy this chip right now if you want it.

• 1:15:30Speaker 00

  By the way, you can see, you can see it in the stock price. It actually did directly affect the stock price. And then after everybody saw it, it directly unaffected the stock price.

• 1:15:42Speaker 00

  People are like, literally, I mean, look at this. This is, you're talking about somewhere in the teens, four, four, you know, 15, we'll just call it 415 for average right there, all the way up to 460. So a 10, 15% boost, all the way back down to now 427. So.

• 1:15:59Speaker 01

  Oh, well.

• 1:16:00Speaker 00

  Oh, well. I mean, so to be fair though, I am going to at least try to glaze it a little bit here.

• 1:16:05Speaker 01

  Yeah.

• 1:16:05Speaker 00

  Uh, I think that trying to prevent, like, first off, 128 gigabytes of unified memory and then saying this 120B plus, this is, this is very misleading as far as what you can run locally. But excluding all of that, uh, the idea that you should be able to run your stuff locally and then that, that, like, that's a good idea. I think is a really good message to be said, because, you know, the whole Darios of the world talking about how everything's dangerous and they should own everything you're doing. There are so many companies that have really like competitive data about a activity, right? They have users who pay them money, who send them specific data. They have like all this good stuff. And they have to effectively ship everything off to OpenAI if they want to use any sort of these models to be able to do any sort of kind of

• 1:16:49Speaker 00

  cool things that they say can't afford a team of data scientists to figure out how to do. And they don't want to share this data with anybody. I'm thinking of like mining factories. I'm thinking about just a lot of just more kind of brick and mortar tech kind of combo places. And the ability to run things locally, I think, becomes really appealing to them. And so this idea of moving towards it, I think is really good. I mean, I, I personally think this is a good direction overall for people. Because you stop having to be tied to large companies. And I think that that is overall a good thing.

• 1:17:20Speaker 01

  I would agree with that and like it would be nice if Microsoft decided to go that direction for some reason.

• 1:17:27Speaker 00

  They're not, they're gonna triple E it.

• 1:17:28Speaker 01

  They don't seem like it because you know, Satya Nadella is famously the, the person who came in and basically like, I mean, I'm not gonna say killed Windows, but he was definitely the person who came in and said, "Look, cloud is, cloud is what we do. That's where we're gonna make all our money is subscriptions, software as a service. We don't really want Windows to be good because the better, the better things are locally, the less we can charge you remotely." That's Satya Nadella.

• 1:17:53Speaker 00

  He'll have killed Xbox too.

• 1:17:55Speaker 01

  What? Well, yeah, but that, yeah, but that's kinda separate from this, this thing. But, so if he's actually now changing tack on that and he's gonna turn Microsoft back to an actual personal computer company, that would be fantastic. I don't believe it, uh, but if that actually happened, that would be a great outcome.

• 1:18:16Speaker 00

  Yeah. So that's, that's, so I'm trying to like be at least Microsoft positive, because I'm always pretty negative on Microsoft. I think they've, they've somehow just fumbled the bag so hard for the last decade. I cannot believe that I think Steve Ballmer is a better CT or CEO than Satya. Like, I'm saying it right now. That's my personal opinion. Uh, Halo 3 and he was ruling during that time. Absolutely loved the man. Uh, and so it's just, this has to be a good, and that's what I'm hoping for, is that people realize the, the need for local and perhaps I will push Microsoft, but Microsoft, they did it with VS Code, which is they built a bunch of cool open source stuff and then they just immediately started EEEing people. I assume this thing is gonna be the, the exact same thing, like,

• 1:18:57Speaker 00

  "Oh, all these cool local models." But if you really want to actually use it with Windows, you have to pay for all these services to be able to build your soul or your, your memory thing and your management and all this kinda stuff, right? And so it's just gonna be, they're gonna re-EE you again. They're just in the embrace season of local AI right now. And someone's asking what EEE is? Embrace, extend, extinguish, right? You become the good guys. Everybody comes to you. You extend it and make it, uh, really, really functional, but then it becomes also like privatized to where it's very hard for anybody to compete with you, but you're also now extracting value. And then you extinguish by diving your prices and effectively killing everybody else out while you can then reap all the benefits of

• 1:19:40Speaker 00

  being a single monopoly winner. Classic Microsoft, by the way.

• 1:19:46Speaker 00

  So, I'm trying to give them a little bonus, okay? Hey, I hope they're good.

• 1:19:50Speaker 01

  Alright, now that the second donut has been glazed.

• 1:19:55Speaker 00

  Uh, we were gonna talk about Nvidia as well, like the Spark. Is there anything interesting in that? Or is it really just purely a...

• 1:20:03Speaker 00

  This is just a... It's just the same thing, cuz I was reading through it and there's nothing that really feels, uh, interesting about it. It just, I mean, they really do love DLSS. And I know that that just, just, just ruffle some serious feathers out there. If you mention DLSS, SSSS. That's all I know.

• 1:20:20Speaker 01

  Yeah, I mean, this, this thing is not, uh, like I said, it's not a new chip. If you wanna know how it performs with GDDR, or, no, it's still, no, in fact, it's still LPDDR in the other box. So you can literally just go see how the thing performs. And I guess we don't know what kinds of concessions might have to be made for like the laptop version. So, maybe the laptop version performs worse for some reason because of something that has to happen there because of the power. Uh, that part I don't know. But, in general, like if you wanna know how this thing works, well, you can just go look at the benchmarks for the thing that we already have. And, I believe the benchmarks for that thing, put it somewhere around a 30/80. Uh, so like if you wanna know how your graphics would run on this thing, it's like a 30/80, roughly.

• 1:21:05Speaker 01

  Um, which is, which is fine, uh, performance. I think, like, again, take all this with a grain of salt, but I think that's roughly where the Spark is. Uh, and again, that's just because, as we said before, you know, games care a lot about the memory bandwidth on a GPU. Um, and so, you know, having not that much memory, uh, you know, on a GPU is actually fine if the bandwidth is really high. That's why we don't, typically gamers don't have 128GB GPUs, right? That's not a common kind of SKU. Uh, and that's because, you know, games are used to fitting in 8GB, 12GB, 16GB now. Um, so they just don't need it as much.

• 1:21:42Speaker 02

  I can't wait for everybody to have 128GB so that games can start taking up that much space.

• 1:21:48Speaker 01

  That's all that'll happen. That's exactly right. All that'll happen is, is now the games will just take 128GB.

• 1:21:54Speaker 00

  They're all gonna ship jobs to digits.

• 1:21:56Speaker 03

  It's gonna be so good.

• 1:21:56Speaker 02

  Sorry, sorry for ruining everyone's day. You guys thought it was bad downloading 100GB for Call of Duty. They're gonna have it be several terabytes and it's gonna be 128GB of RAM required.

• 1:22:09Speaker 01

  And uh

• 1:22:09Speaker 03

  So good though?

• 1:22:10Speaker 01

  And I believe, I mean, I, I think the, again, didn't have time to actually look at this myself. But, uh, vague understanding is, you know, so it's, it's MediaTek, right? I mean, presumably you guys have heard of MediaTek, right? They make cellphone CPUs.

• 1:22:28Speaker 00

  TV CPUs too.

• 1:22:29Speaker 01

  And TV CP, they make CPUs that are in a lot of

• 1:22:32Speaker 00

  1.8 gigahertz quad cores, right? Something like that?

• 1:22:35Speaker 01

  Well, it, it depends entirely on which one of their

• 1:22:39Speaker 00

  I thought that was one of their newer ones, but anyways, yeah, sorry, keep going.

• 1:22:44Speaker 01

  Uh, they're the, like, they make this, the CPU, Nvidia doesn't make the CPU part of this thing. It's two, it's actually two, like, kind of, um, things on a package where the, the Nvidia side is the, the part with, you know, the GPU stuff. And then the CPU side is made by MediaTek. Uh, and it's got just the, the standard, the CPUs that they make, right? So it's, it's a, I think it's 10 cores of the X, of the X line and 10 cores of the A line. So that they have like, there's like 10 performance cores and 10, like, more efficiency kind of cores. That's, that's what it is. So it's, it's not a mystery, like, you can just go look, like, like I said, they, they made it sound like this was this big, grand unveiling of stuff.

• 1:23:29Speaker 01

  It's all stuff, we already know what it is, you can already look up the benchmarks. It's not as good, for example, as a Qualcomm, like, the Qualcomm Windows on ARM CPUs. Those CPUs are like better for performance, I believe, than these, than these are. So on, on the CPU side of things, if you think of one of those Qualcomm notebooks, that's probably like best case CPU performance for, at least for like, single threaded. Like, the things that you would normally be looking at for CPU there. Maybe, maybe if you're, if you're doing like a big multi-threaded Lurkla, maybe it would change. But, so it's, again, it's a really kind of a nothing burger from the standpoint of like hardware announcement. There just wasn't much to look at here, which is good because Prime only gave us about 15 minutes to look it up.

• 1:24:11Speaker 00

  Casey.

• 1:24:11Speaker 03

  Wait, what was... What was...

• 1:24:13Speaker 00

  What?

• 1:24:13Speaker 03

  They said something about 1,000 games. Like, what... What does that mean? You know what I'm talking about?

• 1:24:18Speaker 00

  Yeah, I'm looking at it. Over 1,000 RTX Enhanced and Accelerated Games and Apps are available.

• 1:24:25Speaker 01

  They probably just mean that they tested them and they run, because, remember, they just... They all have to run through a compat layer, because nobody ships ARM-compiled binaries. So, they have to run through a compatibility layer that makes the X64 code run on ARM. Right?

• 1:24:41Speaker 03

  Mm-hmm.

• 1:24:41Speaker 01

  Um, so, they probably just mean that that's...

• 1:24:44Speaker 02

  Or like, it's got DLSS support or something, or it's got some of the other, like, AI capa... Capability enhancement things, maybe? I don't know.

• 1:24:53Speaker 01

  But yeah, I don't think there's anything special about that, because again, like, you know... That's... That's just what you would expect from a Windows PC. Like, what you expect from a Windows PC is all the games.

• 1:25:03Speaker 02

  All the games run.

• 1:25:04Speaker 03

  Exactly.

• 1:25:05Speaker 01

  So, it's not much of, again, not much of an announcement, but I don't know what they mean there. Yeah. I didn't see that part of the announcement, so...

• 1:25:10Speaker 00

  Okay, well, I'm watching a little video of it going. And this one looks like you're playing on a PlayStation 1, and this one looks fantastic, so I don't really know what's exactly happening.

• 1:25:19Speaker 01

  Oh, is that that... Is that that new, um... The...

• 1:25:22Speaker 00

  I don't know what's going on here.

• 1:25:23Speaker 01

  Thing they're gonna do for DLSS 5 or whatever, where they just... Where they just replace the graphics with different graphics that are, like, better?

• 1:25:28Speaker 03

  Looks like Minecraft on the left.

• 1:25:30Speaker 00

  Yes. Yeah, so this thing is just, like, slowly zooming in. Yeah, I'm not... I'm not really sure... What this comparison is, because I... I don't think I've played a game that's looked like this. I... I'm actually replaying Dark Souls 1 right now, and it doesn't look this bad. So, I... I don't really know what's going on here. I don't really get this example, but whatever. I... I guess I'm not a... I'm not in the know on whatever that is.

• 1:25:50Speaker 03

  You're a boomer.

• 1:25:51Speaker 00

  I'm a boomer.

• 1:25:52Speaker 02

  Yeah, okay, boomer.

• 1:25:55Speaker 00

  Alright, well, um, I think that's about, uh... About it, you know? Appreciate you guys.

• 1:26:01Speaker 01

  Great work, everyone. I think we really... I think we really knocked it out of the park this time.

• 1:26:06Speaker 00

  I think we did. Honestly, the floating-point operation thing was super interesting. Talking about the local agents, I think is actually... Not only interesting, it's actually important. I think it's really good. I'm hoping that more people realize that they don't need to be beholden to these big companies.

• 1:26:20Speaker 01

  That would be great if... If the industry go in that way, because that's much better than the current thing.

• 1:26:24Speaker 00

  Yeah. I'm... I'm against current thing.

• 1:26:27Speaker 01

  I am against the current thing and for the new current thing.

• 1:26:30Speaker 02

  Me? I support current thing.

• 1:26:32Speaker 00

  Aw, dang.

• 1:26:33Speaker 02

  So, look now.

• 1:26:34Speaker 00

  Yeah, I actually hate you, is what that means, CJ.

• 1:26:36Speaker 03

  What is the current thing? I don't even know.

• 1:26:40Speaker 02

  Trash, that's actually, smashes red button.

• 1:26:45Speaker 02

  Yeah, red button trash.

• 1:26:46Speaker 00

  That is red button behavior right there, Trash.

• 1:26:48Speaker 03

  Oh, are we talking about the police, the police thing again?

• 1:26:51Speaker 02

  No, I don't know what the police thing is, but we're not talking about it for sure.

• 1:26:55Speaker 03

  I thought that was our grand first blue button. I thought that's what we were talking about.

• 1:26:57Speaker 00

  Yeah, that has nothing to do with police, but I really appreciate you.

• 1:26:59Speaker 02

  The thing we really need to talk about is the minifig situation.

• 1:27:01Speaker 00

  Dude, the minifig situation is insane. Should we talk about that next week?

• 1:27:05Speaker 03

  What's minifig?

• 1:27:06Speaker 02

  Fender guitars.

• 1:27:07Speaker 00

  Oh, my gosh.

• 1:27:09Speaker 02

  Can we do a slideshow as the blocker for next week and show Casey, uh, the, the Lego situation? Brick by, brick by, brick, brick by, what is this?

• 1:27:21Speaker 00

  Okay, how about this one? Trash and Casey, if you see anything about Lego news.

• 1:27:26Speaker 02

  Great point.

• 1:27:26Speaker 00

  Okay. Just navigate away. I will prepare a roughly 10 slide presentation.

• 1:27:31Speaker 02

  Okay.

• 1:27:32Speaker 00

  About it. By the way, TJ, there's been some-

• 1:27:34Speaker 02

  More than that. There's a lot of pictures required. 10 minutes.

• 1:27:36Speaker 00

  There's a lot of pictures, so I'm gonna have to figure out how to like make it so that the blocker isn't the entire episode.

• 1:27:41Speaker 02

  I think it should just be the entire episode.

• 1:27:43Speaker 00

  It might actually have to be, honestly.

• 1:27:45Speaker 02

  Just do it. Just do it. Casey, you will like it. It will, you will be shocked. Just do it, Prime. Just do it.

• 1:27:50Speaker 03

  Lego Pokemon right there.

• 1:27:52Speaker 00

  Yeah, I know. I'm very interested. Hey dude, I got a lot of Lego over here, but it's, it's a very interesting story, but TJ, there's a lot of, uh, lawyers and other lawyer tubes coming out saying that the inverse party is completely incorrect. And that minifigs is the correct one. And so there's like a lot, there's a lot of drama going on and I've been trying to keep up with all of it, so I'll be glad to do a whole like-

• 1:28:14Speaker 03

  I have no idea what you're talking about.

• 1:28:17Speaker 00

  Dude, you just ca-

• 1:28:18Speaker 03

  What's a minifig?

• 1:28:19Speaker 00

  Oh my gosh. Again, Trash, do not even Google it. If you Google minifig, you'll be led into the entire situation, so you gotta stay away from it.

• 1:28:27Speaker 03

  No, I'm just kidding.

• 1:28:28Speaker 00

  Is that your wife's name?

• 1:28:30Speaker 03

  I have a, a-

• 1:28:31Speaker 02

  No, his wife's name is Hey Google.

• 1:28:33Speaker 00

  Alexa! Your wife comes in? Please, could you play? Yes, sir.

• 1:28:41Speaker 02

  Play Despacito and she starts playing it on the piano.

• 1:28:44Speaker 03

  Imagine someone just walks into my frame and just does whatever. That'd be insane. That would be good.

• 1:28:50Speaker 00

  Alright, um, yeah, so-

• 1:28:51Speaker 02

  We'll do that next week. Pitch Deck.

• 1:28:52Speaker 01

  The Receipts app is working.

• 1:28:54Speaker 00

  Pitch Deck, I don't think, honestly TJ, I don't think we can Pitch Deck and do mini figs. The mini fig situation is so the-

• 1:29:00Speaker 01

  Mini figs, next time. Mini figs is next time. That's what we're doing.

• 1:29:03Speaker 03

  Dude, I've been holding my pee for like 45 minutes.

• 1:29:06Speaker 01

  Alright.

• 1:29:07Speaker 02

  Alright, that's it. Bye, everybody.

• 1:29:09Speaker 00

  That's the end of the episode. Bye, everybody.