Programmers vs Dream's Fake Manhunts

[Sponsor] Linear helps teams track issues, manage projects, collaborate on docs, and ship products - and the crazy thing is their software is actually fast and a joy to use. Learn more: https://trm.sh/linear [Description] The Standup crew puts their detective skills to the test with a quiz about Dream's supposedly "uncut" Minecraft Manhunts. Can programmers spot the suspicious edits, hidden clues, and inconsistencies? [Sources] http://www.youtube.com/⁨@rubyswolf9767⁩ https://www.youtube.com/watch?v=WsWNWzSwVUA https://dartjs.com/joltcracker https://mcsrc.dev/ http://github.com/misode/mcmeta [Hosts/Guests] Prime: https://thestanduppod.com/contributor/theprimeagen Teej: https://thestanduppod.com/contributor/teej-dv Trash: https://thestanduppod.com/contributor/trash Casey: https://thestanduppod.com/contributor/cmuratori [Partners] Linear: https://trm.sh/linear Sentry: https://trm.sh/sentry [Socials] x.com/thestanduppod instagram.com/thestanduppod tiktok.com/@thestanduppod facebook.com/profile.php?id=61573286173150 [Other Places to find us] Spotify: https://open.spotify.com/show/01A062kejnXFkJE01bjN5J RSS: thestanduppod.com/feed.xml Clips: @TheStandupPodClips Website: thestanduppod.com

• 0:00Speaker 00

  All right, Casey, hit us with your blocker.

• 0:01Speaker 01

  Hold on.

• 0:02Speaker 02

  First off, I have to start the startup.

• 0:09Speaker 01

  Uh, anyways, I'm sorry. Welcome to the stand-up. Today I have with us legendary game dev Casey Miratori, Tej Dev.

• 0:17Speaker 00

  With two blockers.

• 0:18Speaker 01

  Oh, gosh, we already have the blocker. Okay, thank you, TJ, and also Trash Dev. Hey, Trash Dev. Casey, tell us, what are your blockers? What is this Flew Shop blocker? Let's put this in progress right now.

• 0:28Speaker 00

  No, no, no, I want to do the other one first.

• 0:30Speaker 01

  Oh, gosh, okay.

• 0:31Speaker 00

  The Flew Shop blocker is-

• 0:32Speaker 01

  You'll notice that, you'll notice that I delivered the Flew Shop blocker back faster than I assume William Sonoma has ever done anything. If this is the thing I'm thinking about.

• 0:39Speaker 00

  It's exactly. So, a lot of people have been asking me.

• 0:42Speaker 01

  Oh.

• 0:43Speaker 00

  There was a bit of a thing that happened with William Sonoma two years ago, actually. For people who don't know, I wanted to send Jonathan Blow a box of peppermint bark for Christmas because I know he likes peppermint bark and I thought that would be a nice thing to do.

• 1:05Speaker 01

  That is a nice thing to do.

• 1:06Speaker 00

  You know? I was trying to be a good person on Christmas, all right? And send someone some peppermint bark. That's all I was trying to do. And William Sonoma, they, they just flat out lied to me. They, first of all, they said they were gonna ship it. They never shipped it. The like, the shipping, like, place just said "label printed." It never changed from that. And then, six months later, they changed the status to just be delivered on the William Sonoma end, but it had never, it was still just label printed and then it just disappeared from the, from the shipper's people. They're like, "We've deleted this record because we never received it back."

• 1:46Speaker 00

  And so, I was wondering, I'm like, I sent a box of peppermint bark. Which is going to finish first? The 10-year project that involved creating a new language, a new engine, and either one of the largest or the largest puzzle game ever created, will that ship first or will the peppermint bark ship first? And now, last week, we got the answer. Order of the sinking star is out and John has received no peppermint bark after two years. Zero.

• 2:19Speaker 01

  And just to put this in perspective, John's language also iterated and made macros or like compiled time programming in, in a much more expansive way than people thought humanly possible. He iterated on human design faster than bark could be shipped.

• 2:34Speaker 00

  Yes, like, you could meta-bark before William Sonoma could bark, basically.

• 2:40Speaker 01

  Thing.

• 2:41Speaker 02

  Where's your money?

• 2:43Speaker 00

  So, here's the thing. I decided, I decided that I'm just like, I'm gonna be a good employee, right? And I'm gonna be, you know, you just don't complain about your teammates when William Sonoma obviously should be fired. You take matters into-

• 2:58Speaker 01

  William, William, William.

• 3:00Speaker 00

  So, what I am doing, what I'm doing right now, and I'm gonna go to the store today, I bought a bark, box of peppermint bark myself.

• 3:13Speaker 02

  Literally says that, huh?

• 3:15Speaker 00

  And I am going to try, and, and this is important, I'm gonna try to ship this to Jonathan Blow because maybe it's just impossible. Like, maybe shipping peppermint bark really is harder than making like the largest puzzle-

• 3:25Speaker 01

  It could be illegal.

• 3:26Speaker 00

  Maybe it's impossible. I don't know.

• 3:27Speaker 01

  Peppermint across state borders? That could be a crime.

• 3:29Speaker 03

  It's like the TSA of, of shipping is like, they're like, "Well, one time a guy tried to take out the US Postal Service with peppermint bark. We can't have that happen again."

• 3:39Speaker 02

  I have a question. I don't know what peppermint bark is. Is it in the shape of a dog or is it like the bark of a tree?

• 3:45Speaker 03

  Oh, I love that trash.

• 3:46Speaker 01

  Wow, it's, that is beautiful. This is-

• 3:49Speaker 00

  It's a very, that's a very good question.

• 3:51Speaker 01

  Especially Trash, he just put his entire thumb into his mouth and sucked off frosting from a, like, from a donut he's eating. This man loves trees and does not know what bark is.

• 4:01Speaker 02

  I hate peppermint, too.

• 4:03Speaker 01

  But it doesn't have to be, okay.

• 4:05Speaker 02

  I know, I've never heard of bark.

• 4:06Speaker 01

  Casey, sorry, Casey, go. It's migration season. Nothing like seeing tickets fly done for the winter. Look at, look at that guy go. Look at that, look at that guy go. Finally, a ticketing system that can keep up with my agents.

• 4:19Speaker 01

  Hey, cut the jokes. Linear did the impossible. It's helping me stay organized and that's why I'm so glad to have them as a partner of this channel. So, if you want to support this channel or just get your work organized, check out Linear today. Links in the description.

• 4:33Speaker 00

  No, but I don't, I think Trash has a very valid point here and I, that I was about to point out and, and I don't think, I don't think he realizes just how prescient he is because I believe it's probably bark of a tree based on the look of it would be my guess, right? But what do you call that right there?

• 4:53Speaker 02

  That's why I got confused.

• 4:54Speaker 00

  It's a freaking dog, dude. That's a dog, right?

• 4:57Speaker 02

  That's when I got confused. I got confused when I saw that dog and I was like, "Oh, maybe."

• 5:01Speaker 03

  Quick question? Do you think they're AI generating the top of the boxes now? And AI got confused.

• 5:08Speaker 00

  I think these might be older than that, but that would make sense. However, uh, well, we'll get to that. I'm also sending him the advent calendar, which also has two dogs, just like the other one, two dogs on it.

• 5:22Speaker 02

  Yep.

• 5:23Speaker 00

  And Trash-

• 5:24Speaker 01

  Is that one exclusively peppermint bark or does it come with various other barks?

• 5:28Speaker 00

  I think it's just peppermint bark. I don't know.

• 5:30Speaker 02

  Where are the barks, are there?

• 5:31Speaker 00

  I don't really know. But here's the thing-

• 5:34Speaker 03

  There's at least chocolate, I feel like.

• 5:35Speaker 01

  Yeah, there's like, uh, that, uh, that peanut brittle is kinda like a form of a, of a bark. Okay, what?

• 5:40Speaker 00

  I don't know what the heck this was supposed to be, but I'm like, John's getting that, too.

• 5:44Speaker 02

  See, that's what I was envisioning when I saw the cover of the peppermint bark.

• 5:46Speaker 01

  Okay, but I feel, Casey, I feel like you're kind of going on hard mode, though, because previously you just did bark and now you're doing bark, big bark advent calendar, plus a dog.

• 5:54Speaker 00

  Yeah. I, I figured, look-

• 5:56Speaker 02

  It's getting stolen.

• 5:57Speaker 00

  Here's the thing. I, I like doing fault tolerant programming. It's one of those things that I enjoy looking at and analyzing. It's, I think it's fun. And so, what I have learned in fault tolerant programming is that if you need something, if you're really trying to get something to work that is extremely difficult, like shipping peppermint bark to Jonathan Blow apparently is, then you do it multiple times, right? You have multiple things that you send there and yeah, maybe the dog doesn't get there, but the tin, the box does, right? The advent calendar gives us 24 separate days of possible peppermint bark that could succeed. Just think about that.

• 6:38Speaker 01

  Is it 24 or 25?

• 6:41Speaker 00

  I don't think you get one on Christmas. Do you? Maybe you do.

• 6:45Speaker 03

  It depends.

• 6:46Speaker 01

  I'm looking up right now.

• 6:46Speaker 02

  I don't think so.

• 6:47Speaker 00

  Does it say? Does it say?

• 6:49Speaker 01

  Oh, look at this. It's actually, it's actually a split. So, there's two schools of thought. 24 is the traditional one.

• 6:54Speaker 00

  Yeah.

• 6:55Speaker 01

  Whereas 25 is a, is the modern commercial one.

• 6:58Speaker 00

  This one looks like 24.

• 6:59Speaker 03

  Just people these days, they want to have their cake and eat it, too.

• 7:02Speaker 00

  Yeah.

• 7:02Speaker 03

  You know what I'm saying?

• 7:03Speaker 00

  They want to have their bark and eat it, too.

• 7:05Speaker 03

  Yeah, exactly.

• 7:06Speaker 01

  This guy, they disgust me.

• 7:08Speaker 03

  Disgusting.

• 7:10Speaker 02

  I don't think I've met anyone that is Ian Bark before.

• 7:14Speaker 01

  I love peppermint bark.

• 7:15Speaker 00

  Well, Trash.

• 7:16Speaker 01

  Peppermint bark is one of the most delicious treats of all time.

• 7:18Speaker 00

  I love it, yeah. I'm with you, Prime.

• 7:21Speaker 01

  Also, peppermint Hershey kisses, very good.

• 7:26Speaker 03

  Trash, we'll get you some for next week. Some non-peppermint versions, too, because I'm with you. I'm not a huge peppermint guy in all different situations. I don't mind it as a candy, but I'm not a big like peppermint plus chocolate kind of situation guy.

• 7:37Speaker 00

  They have other barks, yeah, like caramel bark and things like that.

• 7:39Speaker 02

  Yeah, looking at them.

• 7:40Speaker 00

  So.

• 7:41Speaker 03

  That's what I, that would, I would be down for that. All right.

• 7:45Speaker 02

  Interesting.

• 7:45Speaker 03

  Trash is really considering it, though. Look at how much thought is going. I've never seen him this focused on the stand-up in my entire life.

• 7:52Speaker 01

  Dude, Trash just gave us his X high face right there.

• 7:56Speaker 00

  Yeah, he just turned the model and-

• 7:58Speaker 02

  I, I think I've seen this before, but I don't think anyone's ever called it bark. It looked like, it looks familiar, just like, I don't know.

• 8:06Speaker 03

  I'm aware of your game.

• 8:09Speaker 00

  Yeah. What do you got, Prime? You got a blocker today.

• 8:11Speaker 01

  I got, I got a blocker. I got a small blocker, but before we do that, let me just move this one to-

• 8:15Speaker 02

  Should it be done, not interview?

• 8:17Speaker 01

  We're gonna, okay, you're right. We, we can go down to done.

• 8:20Speaker 00

  Are you even a scrum master? Because I'm starting to think that you don't really know.

• 8:24Speaker 03

  He's a scrum master.

• 8:25Speaker 00

  I don't know how this works at all.

• 8:26Speaker 01

  All right, Casey, I saw something today on the internet.

• 8:28Speaker 00

  Okay.

• 8:29Speaker 03

  Oh, boy.

• 8:29Speaker 01

  And I thought of you immediately and I started thinking, okay, I wonder what Casey would say because, correct me if I'm wrong, typically when we're designing new chips, it's like a multi-year process of taping it out or whatever that crap's called and designing and I'm sure there's FPGAs involved and figuring out everything that goes into designing a chip that at the end-

• 8:49Speaker 03

  Salt and vinegar.

• 8:50Speaker 01

  Salt and vinegar.

• 8:50Speaker 03

  Jalapeno pepper.

• 8:52Speaker 01

  Some of us are barbecue, baked kind of boys, you know what I mean? This is very confusing. But then I saw today-

• 8:57Speaker 00

  Usually, I mean, just to be clear, usually it's about because of all the validation and test, like, it's like, you know, typically there's this really long lead time because you, you know, you don't want to put a crappy chip out there that has all these bugs in it, right?

• 9:08Speaker 01

  Yeah, you don't want to be, what, what was the, what was the lake? Raptor Lake or whatever? What was the lake from Intel that was just like filled with non-stop people talking about how terrible? It was like a few generations ago, I think?

• 9:17Speaker 00

  That was, yeah, you're talking about there was a, a voltage issue on those, but then there was like the Pentium divide bug would be another classic one that they had, uh, that was like, okay, yep, that, you know?

• 9:29Speaker 01

  Yeah, I don't want a divide bug, okay? That does not sound good. All right, so, I saw this today and it said, "These two people from Google are developing chips with AI, claiming it's gonna take the process from multiple years down to potentially one day."

• 9:47Speaker 03

  Sick.

• 9:47Speaker 01

  And I was curious, Casey?

• 9:50Speaker 00

  Yes.

• 9:51Speaker 01

  In our world, what would that do? Like, let's just theoretically locate it, just ignore all the actual practical implications of everything. If you could go from idea to actual like, "Hey, this design works in one day," what would that do to Moore's Law and all these, like, these, you know, more, again, Moore's Law, I'm also kind of hand waving here, the theoretical double improvement every 18 months, even though I know that's not what Moore's Law is? Like, what would this do to the world if this was actually true?

• 10:21Speaker 00

  Um, well, I guess I don't really know what part there would be speeding up exactly for that. I mean, it would help if I read the article at least, uh, just so I know like what they're talking about. But, um, I, I mean, obviously, shortening lead times means that you would get access to the thing that you would have had, you know, in a, in a year, you could get access to it now, right? And I guess what I would say is, unfortunately-

• 10:57Speaker 00

  So, so one thing you could definitely say that you would get is a one-time bump in that all the things that would have been a year or two years out, we could get them sooner. So you'd have a one-time bonus of like shrinking that horizon.

• 11:08Speaker 01

  Okay, we'd zap into the future, everyone would just be twice as fast. Or whatever.

• 11:12Speaker 00

  Well, not, not these days.

• 11:14Speaker 01

  Not these days.

• 11:15Speaker 00

  Um, yeah, maybe if, if this was, had been something that had been done back in the day. Uh, but the other thing is presumably, um, presumably that's a bonus, you know, even after that, there's, there's some bonus to it where it's just like, it could, the industry can be more responsive, right? Because currently, you know, there's, there's not really that big of a problem with having two-year lead times in CPU design or something like that these days. It's not like that's a huge issue. But the part, the part where it becomes, you know, very, uh, you know, maladapted is when something shifts in what people wanted the chips to do and the industry can't respond as fast as it would like, right? So.

• 11:58Speaker 01

  So AI going from like not really big to, hey, we need to build a hundred billion dollar data center yesterday is a, uh, is a non-traditional amount of growth.

• 12:08Speaker 00

  Sadly, no, not like that. Uh, and the only reason I say sadly, no, not like that is because that is just a pure capacity problem and this would not help that at all, right? Uh, so, so what I more mean is like, uh, say, say on the AI front, you were building all of these CPUs and you know, you didn't assume that people needed very much like neural, like, like tensor throughput. And then now AI is all the rage and everyone wants tensor throughput, even in like their laptop chips. Well, now you could adjust that balance very quickly. Like, instead of waiting two years to change how much of the diarrhea is dedicated to neural, now we could just change it right away. And like the chips that ship, you know, in two months or something can have it because

• 12:50Speaker 00

  the design, we just redid it and the fabs, you know, we, assuming that this article is corrected, I, I'm just.

• 12:57Speaker 01

  Yeah, no, we will pretend. I, I know, I led, that's why I led with, let's pretend like everything is, because I was just curious, like, what does it do? Does it fundamentally change the, the slope of improvement? Or is it going to be one of those things where like, okay, now everybody needs these tensor new calculations and everything? So we have a, we have like a perceived speed up right away, but it's more like we just realized what we could have already had, we just simply didn't plan for that. And so therefore it's like, oh yeah, this is what it would have been.

• 13:22Speaker 00

  Yeah, I assume, um, I mean, it's really hard for me to say because not being a hardware designer, like I can't necessarily say all the things that this might help, but like in general, if you design hardware, it's not like you don't have the ability to experiment with new designs before, before actually starting the process of like taping out a real chip, right? Uh, so I'm not sure how much this would actually impact, like the, the overall quality of chips. That's a better question for a hardware designer. So, but, but it clearly, at least to my mind, I'm like, it clearly just means you could be more responsive because if now you're just saying like, look, after we decide what we actually want to do, it only takes us a few days to like vet that design or something,

• 14:06Speaker 00

  then obviously we could make sure that, you know, the, the lead time on orders to TSMC is, is not going down at all because of that, as far as I know, there's still some lead time, but at least you know that when you place your, uh, you know, you, you've got your allocation TSMC, when you're going to tell them what you want to actually build at whatever time that has to go in, you've cut your, your time down to just whatever the length of the process is, but you know, between those two steps. So getting the, you know, telling TSMC what you're actually going to, to fab, them doing all of whatever has to happen to fab it, and it coming out the other end, that's not going to change, but the, the lead up to that would, would presumably, right?

• 14:44Speaker 01

  Okay.

• 14:44Speaker 00

  So that's my best guess.

• 14:45Speaker 03

  And there's only so many wafers, there's only so many wafers in the world and then time.

• 14:48Speaker 01

  And Trash eats a huge amount of wafers, I know that. He may not know about Bark, but he's a wafer guy, it seems like.

• 14:54Speaker 03

  Yeah, Trash, thoughts on, thoughts on wafers versus Bark? We'll have to have that test maybe next week.

• 14:58Speaker 01

  Yeah. Do you think AI could improve that process? Your Bark to wafer speed. Okay.

• 15:05Speaker 00

  Yeah, and I would, I would assume that like, yeah, I mean, at the moment, most of the problems with chip supply right now is literally capacity. It's not really design, like design, obviously you're, you're always looking for ways to improve it, so I'm assuming that this is a good thing. I mean, I haven't seen it, so I don't know. Uh, so it's not like, uh, speeding up design is a bad thing, obviously it's good, but, um, at the moment, that's not the problem, right? The problem is just that building these physical facilities and the ASML machines, there's only so many of those that are necessary for doing the patterning. It, you know, it's just, it's a, it's a capacity constrained industry at the moment and design is not the bottleneck at the, at the moment.

• 15:49Speaker 01

  Okay, I was just curious because when I hear things such fantastic, like, oh, this process used to take us two years, now it takes us one day, it's just like, oh, that's like a thousand X improvement, so what does that buy us? And I just, I, I have no way to even reason through that question because it just sounds, it sounds too fantastic to be real. Cause I would just assume if there's any other process that you could be like, it's instantaneous now, it would just immediately make a big change. Just kind of like, like software, like we build our own, a bunch of own crazy tools internally that are awesome, that we would have just never, you know, built before, so there's obvious changes there.

• 16:22Speaker 00

  And like I said, maybe there is, yeah, maybe there are some, uh, ways that this helps them, uh, other than being more responsive, I don't know. Cause like, again, not being involved in hardware, it's hard for me to say like, you know, how much, how much of the, how much of your ability to improve the design is a problem with this, you know, is a problem with that lead time versus, you know, just the fact that, well, you know, um, we have to kind of be slow because when we tape out a chip, that chip has to ship into the real world and we actually have to produce a certain number of them at scale, right? And so it's like, it's the commitment to that, that is the thing. We want to be really sure this is actually what we want. Okay. I don't know.

• 17:04Speaker 01

  Yeah, that makes sense.

• 17:06Speaker 00

  Probably my guess is, um, I mean, cause they already use AI in designing chips, you know, for doing like layout stuff and things and optimization. Nvidia's already, um, you know, multiple times claimed that they've shipped, you know, cert, like, like they're, they're actually like, their organization of things like their, you know, arithmetic logic unit goes through AI layout tools and stuff, but that's been true for a very long time.

• 17:29Speaker 01

  Okay. All right.

• 17:30Speaker 03

  That was true, it's been true for like quite a while.

• 17:32Speaker 00

  Longer than like LLMs and that stuff.

• 17:35Speaker 01

  Yeah, I knew network routers and all that, they also use AI for a long, long time to write the, I've heard they write code of areas, kind of like how to move the packets around by AI. So this is not shocking that.

• 17:46Speaker 00

  Yeah, machine learning for things like, you know, routing is a very common thing you would want to do, right? Uh, and so, so it's not like they weren't already using AI to speed up that process. So again, and it's hard for me to comment on this because since I haven't seen the article, I don't know what specific things they're talking about, uh, having improved. One thing it might, it might mean is people who work on the validation teams might, might be losing their jobs, that could be bad, right? Like, because a lot of that stuff is, you know, it's man, it's manual or partially manual, right?

• 18:20Speaker 01

  Well, I'm glad we're still using tape though, 3M will always be a company. All right. Tape it out is still a massive thing. All right, so TJ, we do have a very important thing because it does seem like you have something for us. I didn't know what to put it as, so I just put it as the amazing game, because I don't even know what it could possibly be. I know that's, I, I did not look at what they were doing. I even walked into a meeting where everyone hid their laptops. So it was clearly whatever's going on, this is going to be fantastic.

• 18:47Speaker 03

  Hey, everybody, and welcome to the stand-up today.

• 18:50Speaker 00

  All right, all right.

• 18:51Speaker 03

  Um, okay, so here's, Prime, I'll do the intro. Uh, so this will be a little bit different.

• 18:57Speaker 01

  First off, to interrupt you, I already did the intro, I welcomed everybody but Trash because you interrupted me. And so now this is a second intro, buddy.

• 19:06Speaker 03

  Yeah, well, that's probably true. Okay, so recently, recently, a big drama in, uh, the Minecraft community went on and as I was watching the video about it, I was struck by there was a bunch of fun programming concepts in the drama. So I thought we could do a little quiz game show.

• 19:29Speaker 01

  Okay.

• 19:29Speaker 03

  Where all of you will get to participate as you go along and, um, experience some of the, the things along the way. And then, um, we'll just sort of talk about some of the programming things that are happening as we go, because there's actually some fun, uh, programming stuff. So that's kind of why we're doing that here today. Uh, and also just like, it's one of those things where it's hilariously awesome how they found a bunch of these things here, and so I wanted to go over it. So if you guys don't know who Dream is, he's probably the biggest Minecraft YouTuber ever. Uh, and he's got this series called Manhunt, where they're all.

• 20:05Speaker 01

  The most elusive prey, they say.

• 20:07Speaker 03

  Allegedly, allegedly, uncut, real, raw videos. Okay? That's, but, and I say allegedly in the sense of he claimed it, what's the word instead of allegedly that I'm looking for?

• 20:23Speaker 00

  No.

• 20:23Speaker 01

  Purportedly. Purportedly.

• 20:25Speaker 03

  Purportedly. Purportedly, thank you. Um, and so of course, every time someone makes a claim on the internet, everyone immediately wants to prove them wrong. So people have been trying for a long time to prove that these aren't real, okay? There's like a whole bunch of drama for like literally years for a bunch of these, whatever, that's not the main point of the video. Here's the basic premise. I'm gonna play a little clip. Ideally, ideally, this is also gonna play on your guys' screens for you. If it doesn't, then hopefully then this works for mine. So we'll see what happens. Of course.

• 21:01Speaker 00

  I had one clarifying question.

• 21:03Speaker 03

  Yeah, of course.

• 21:04Speaker 00

  What is the?

• 21:07Speaker 01

  Minecraft.

• 21:08Speaker 00

  When you say uncut, why is it impressive that they were uncut? Because the guy could go eat some food or like, what is the, what is the bonus to it being uncut exactly?

• 21:18Speaker 03

  So in each of the videos, it's set up that Dream, the main guy, is the host, is trying to escape like four to five other people hunting him down in the game.

• 21:27Speaker 00

  Okay.

• 21:28Speaker 03

  And so they go and do a bunch of very like insane stunts and he does like crazy parkour. They go and fight stuff and they're like chasing after each other. With the premise being like, hey, it's a two hour video or something.

• 21:42Speaker 03

  Like, and all of this really happened. So the drama is elevated, right? It's like reality TV. It's supposed to be like, it's reality TV, uh, but on the internet. Does that make sense?

• 21:51Speaker 00

  Well, reality TV is heavily edited, but they're saying this is not for some, okay.

• 21:55Speaker 01

  This is reality TV.

• 21:58Speaker 00

  Okay, real reality.

• 22:00Speaker 01

  Yeah.

• 22:00Speaker 03

  Right. Um, and honestly, that part doesn't really matter. We're just gonna get eventually to the code parts, cause I think that that's what's fun. But like, that is the premise. And so here, I'm gonna play a little bit from the original video that kind of broke this news. So here you go.

• 22:14Speaker 04

  New evidence was just discovered that suggests Dream cuts his supposedly uncut man hunts in order to fake them.

• 22:19Speaker 03

  Fake.

• 22:20Speaker 04

  And not just once or twice, but forty-six different times across the three original videos.

• 22:24Speaker 03

  This was all- The other thing of course is, since he says this and claims it, then like, that makes the drama higher, right? So he's claiming it, cause he thinks it's, it makes it bigger, so then people wanna stop it.

• 22:34Speaker 04

  Found using new tech that allows anyone to fairly easily search for hidden cuts in any Minecraft video. Because normally they're almost impossible to see. I mean, try spotting the hidden cut in this clip. Pretty tricky, right? But what if you somehow had the ability to see how long the game had been running for? Now, try paying attention to this stopwatch.

• 22:56Speaker 04

  I've been playing the game for two minutes here, but then suddenly during the cut, it jumps up to three minutes. Clearly giving away the fact that I hit over a minute of failed attempts.

• 23:08Speaker 04

  Then, for this next cut, this time it jumps down.

• 23:12Speaker 03

  Part of the reason I had to include this beginning part is because I just need to get my statement out there. Lego Batman, amazing movie. Trash, have you seen it?

• 23:21Speaker 02

  The Lego movie, but not Lego Batman.

• 23:23Speaker 03

  Lego movie's really good.

• 23:25Speaker 02

  I've seen the Star Wars Lego movie, or series or whatever.

• 23:27Speaker 03

  But did you watch Batman? Cause if you haven't watched it with your kids, you'll love it.

• 23:31Speaker 02

  I haven't. I will.

• 23:33Speaker 03

  Okay, I just needed to make sure my public statement about Lego Batman was on the record. Thank you. Okay.

• 23:38Speaker 04

  And that's because I had to call it a day and close down the game, only coming back to it later, starting a new session from zero. There's just one problem though. Minecraft never actually shows us this stopwatch.

• 23:54Speaker 03

  So, that's the basic premise of what's gonna, of what they're gonna do. We're gonna get to, uh, we're gonna walk through a series sort of of steps along the way here. And eventually tell you how they found out, just from the info that's currently on the screen right here, how to reconstruct, basically that timer. Uh, from like Minecraft internals.

• 24:15Speaker 01

  Okay, so just, can I, can I just throw something out there? Just so I understand it, which is that, we just watched a video that he made that looked uncutted to the untrained eye. But it actually was cut several times and he even quit for a day and came back. And so he is an, this person is claiming to be an expert not only at detecting it, but also he can create it very, very well. Like, that was fully believable for me, the rookie.

• 24:39Speaker 03

  Right. Yes. Watching the video, you don't know that he took like a hundred shots and missed a bunch of times and then made it go together, right? So he's do, effectively, Dream is doing Minecraft trick shots, right? And he's claiming they're a hundred percent real. And then people are getting upset, cause they're saying it's not real. Okay. So, we're gonna have a quiz along the way. And as we go along, each of you is gonna get a chance to guess where or how a cut is supposed to have happened during this clip. Okay? So we're gonna watch a clip. You guys will get to guess how they detected that there was a cut. And then I will reveal the right answer. So that, what?

• 25:24Speaker 00

  So these are-

• 25:24Speaker 03

  Oh, thank you.

• 25:25Speaker 00

  These are the things I'm supposed to be watching for on the screen is-

• 25:28Speaker 03

  These are the four items that you're supposed to be watching for. We don't act, don't worry also, we don't actually have twenty-nine different clips. We have some, I have some slides later showing the code.

• 25:37Speaker 01

  Yeah, it says cut one out of nine. Don't worry.

• 25:38Speaker 00

  What is the phrase lag spike refer to?

• 25:43Speaker 03

  Trash. This is your expertise.

• 25:46Speaker 02

  Like, hey, Kay, see?

• 25:49Speaker 00

  No, no, but I mean, when we say lag spike in this case, are we literally just talking about the frames like lagging? Or is it Minecraft, is it, is it frame drop or is it some other thing like?

• 26:02Speaker 03

  They're saying, so occasionally Minecraft especially, you know, JVM Minecraft, has, uh, you know, some problems on the server. Uh, with a multiplayer server, right? Cause these are multiplayer games. Uh, he has his real friends coming and joining the game and they're playing, they're playing, uh, remotely and stuff. So, okay.

• 26:22Speaker 00

  I'm sorry. I hate to be that guy, but I'm just saying, so if you have a game where the client and server are decoupled, you will not see a framerate drop in that case. What you'll see is something like a guy teleport or something like, so I'm just asking, in Minecraft, what does a lag spike look like?

• 26:36Speaker 03

  Oh, okay. It'll look like teleporting.

• 26:38Speaker 01

  Okay, so you rubber bands. I believe we call that rubber banding.

• 26:41Speaker 00

  Got it.

• 26:42Speaker 03

  Yep. Yeah. Um, okay. All right. So I'll click play and hopefully this is gonna work here.

• 26:53Speaker 01

  All right, I think I got it. I think I got it.

• 26:56Speaker 03

  All right, so then we'll stop. Open answers.

• 26:58Speaker 02

  Can I click? Oh, I can.

• 27:00Speaker 03

  You should be able to click.

• 27:02Speaker 01

  Yep, I clicked.

• 27:02Speaker 03

  And it should keep me updated when you guys click.

• 27:05Speaker 00

  I have no idea.

• 27:05Speaker 03

  Okay, this is working so well, I can't believe it. Let's see what happens. Locked and reveal.

• 27:14Speaker 03

  Prime thinks, not that guy, thinks it's particles vanishing. Just watch the video again, Prime.

• 27:20Speaker 02

  It's pretty odd. You see it.

• 27:23Speaker 03

  You can literally see it. You can see it jump. It just jumps in the video.

• 27:30Speaker 03

  Here.

• 27:31Speaker 02

  It's right in the beginning, right?

• 27:33Speaker 03

  Yep. Good call, Trash.

• 27:36Speaker 02

  Right here, coming up.

• 27:39Speaker 01

  Okay, see, I thought it was, I thought he-

• 27:41Speaker 02

  I honestly thought that was me.

• 27:43Speaker 03

  Right.

• 27:43Speaker 02

  I thought it was me that spiked, but then you played it again and did it. I was like, okay, that's-

• 27:47Speaker 01

  See, I thought it was gonna be, cause this is like so intense that someone mapped out how often the torches go like this. And they're like, actually in the background, when you looked at that brief second, one went down, but then the sparkle, you know, I was just like, oh damn, okay, that's-

• 28:02Speaker 03

  Prime, that was exactly what I wanted you to think and why I put this one as the first one in the quiz.

• 28:09Speaker 01

  I was looking at the stars real hard. When I saw that, I said, okay, TJ's lying to me. That's, I thought this was the red herring.

• 28:17Speaker 03

  I see, I see what you're saying.

• 28:18Speaker 02

  It was just too obvious to ignore that spike.

• 28:21Speaker 03

  Yep. Okay. Well, that's okay. We have some more cuts. You have some more chances to guess, Prime, okay? So, uh, here's where I have a new confession though. Uh, this, this presentation's actually just a way for me to ask Casey a bunch of, uh, questions about video games. And Casey can explain to me-

• 28:40Speaker 01

  Legendary game programmer Casey Bortori.

• 28:41Speaker 03

  Legendary game engine and developer, uh, Casey Bortori. Of what the heck is going on when we, this is also not just for me, but for the audience. This is also turning this into an educational opportunity, right? What happens during the lag spike? This is a proper client server architecture, like the server is running somewhere else, uh, for Minecraft. And like, how do you make those happen less or worse? Or, uh, like appear less worse, I'm saying.

• 29:07Speaker 00

  Well, again, it really has to do with where the, you know, I, I would imagine that people use the term lag spike now. This is why I asked that clarifying question to refer to a lot of things that might have nothing to do with the lags, but that, that lag might not be the best term for or whatever. So lag spike is now like vibe coding, right? It, it, it has its own meaning that may not exactly be the meaning that, that was, it was originally created for, so who knows? We're just basically talking about something that's a hiccup, right?

• 29:34Speaker 01

  Yes. My kids say the word lag all the time for things. I'm like, kids, that's not what lag means. You're using the wrong term here.

• 29:39Speaker 00

  So, so who knows, right? But in general, what happens is, uh, you have a, obviously in a video game, you have a certain budget. That you have to get everything-

• 29:52Speaker 01

  Three hundred five million dollars.

• 29:55Speaker 03

  Oh no, not-

• 29:57Speaker 00

  In sixteen milliseconds or thereabouts, right? You're trying to get everything to the GPU that needs to go to the GPU to like get the next frame kicked off and whatever. And those are supposed to come out at a regular sixty frames a second, or sometimes higher if you're trying to, you know, run it at a faster framerate or whatever.

• 30:14Speaker 00

  And so what you have to look at is what are all the things that could happen in there, right, that could cause a problem? And there's two broad categories. I won't bother going into too much detail so that we don't have an hour long podcast on just this. But one category is there's actually something happening on the client's side in the code that's trying to get that stuff done in the 16, 16 milliseconds. It's 16.6666 or whatever, but you know what I'm saying?

• 30:46Speaker 03

  Yeah.

• 30:48Speaker 00

  You have something happen in there, which actually means that you missed the time when you would have had to have gotten the data to the GPU to present the next frame in a smooth fashion. Right? And so that can look like any number of things, just like there were some computations that that took too long. It can also be that you handed stuff off to the GPU and the GPU took too long. So it had too much graphics processing to work. That's very common now, right? It's got too much graphics processing work to do. It doesn't make its, you know, 16 milliseconds that it then has to to hit that frame rate. Right? You got this like pipeline of things that all have to sort of daisy chain. They each are getting 16 milliseconds in sort of a rolling window.

• 31:27Speaker 00

  And so if anyone misses there, just through the code not running fast enough on the client's side, that's one place you could obviously get a spike. The other place you can get it is if, so you have a network in between these two things. You have a server on the other side. Obviously, the server can have the same problem. So the server could be in a situation where it takes too long to produce whatever the thing is it needs to send to the client. Right? Or the network could have a problem, right? It's not getting things fast enough. What happens in those cases is that the client, when it's getting this information back, it's seeing, you have to imagine, right? Like on your machine, you are effectively in a time window.

• 32:10Speaker 00

  You could, depending on how you want to conceptualize your head, you're in a time window that's a little bit ahead of the information you're getting by some number of milliseconds. Right? Cause the server is sending you stuff and you, for example, know that you, you know, your computer knows you pushed the spacebar to like jump or something. You pushed it right now. The server can't possibly know that because you've got to send a packet out to it and it's got to send the packet back. Right? And so on. So the way to think about it is every game engine has to have a strategy for what it's going to do to mix this variable set of event timing together to produce something coherent. Right?

• 32:51Speaker 00

  Now, one thing that you often see is the client will just simulate its own world and it will take your inputs and simulate them at the time that they happen. And it will simulate what it thinks the server is probably going to do. The server meanwhile is simulating the, like the authoritative version of the world that takes everyone's inputs and simulates them fairly, hopefully. Right? And so effectively what you've got is two constantly diverging states of the world, the one on the server and the one on your client. And the client has to do something to bring those two back together. Right? So one way that you can also therefore have a lag spike is if that code is unable or not written properly or whatever, or is unable to, based on the circumstances,

• 33:35Speaker 00

  generate something smooth and seamless that hides the difference between those two things. So a classic example in a really like case that's, you know, no one's doing anything wrong, but you just end up in a situation that's sort of untenable. I'll describe. You could be playing a game, you know, like a battle royal game and you're running with your gun or whatever. And there's a guy, your teammates running.

• 34:01Speaker 00

  Your guys running.

• 34:01Speaker 03

  Relatable.

• 34:02Speaker 00

  Your teammates running right next to you. Right? You're both running and you just, you're running together and then all of a sudden they like either rubber band back out of your view immediately or just disappear. Right? And you're like, what happened there? It's like, well, the last information that the client got from the server was that you were both running in that direction. Then you lost network connection to the server for a little while. So it just extrapolated where it thinks the guy's going, but really like, you know, a half second ago, he broke right. Okay. When eventually your client gets the information that he broke right a half second ago, it has to like interpolate those positions. It just has like, oh crap, I thought he was here. He's way the heck over there. Right?

• 34:45Speaker 00

  So that's another way you can get a lag spike and it could appear in all sorts of ways, depending on how the client tries to like, show you a smooth transition between those two things. It could just hard cut, in which case it's just a jump like whoop, he's gone. Or it could try to like smoothly interpolate, in which case he goes like wow, you know, like, right? But, but that's fundamentally the, so those are usually the two broad categories. You're looking at something in one of those two things.

• 35:10Speaker 03

  Great. Thanks, Casey. Any other follow up questions? Trash, Prime.

• 35:15Speaker 01

  That was legendary.

• 35:16Speaker 02

  Very legendary.

• 35:17Speaker 03

  Okay, legendary.

• 35:18Speaker 01

  The one I don't like is when it happens to you on Apex, because I used to play Apex quite a bit. You'd, you try to run and it's like you're running insane. It's like you're running in a dream because it constantly is like bringing you back, bringing you back, bringing you back. That's tough.

• 35:31Speaker 03

  That's like when there's this one hill by our high school, uh, where Kayleen and I went, you know, when we were.

• 35:38Speaker 01

  Or TJ was the star football player. Could throw a football over a mountain.

• 35:43Speaker 03

  Drummer in the marching band. Okay, Prime. I quit football. I was on, I, I could have played varsity, okay, Prime? And I quit football to play, uh, play in the band. But there's, she did cross country. There's this one hill by the school that she'd always be like, oh, I hated running up this hill. So every time we go up there, I always make the car go really, really slow right as we get to the top.

• 36:09Speaker 01

  She gets a re-experience.

• 36:10Speaker 03

  It like feels like.

• 36:11Speaker 00

  He's like, I hate when you do this.

• 36:15Speaker 03

  Oh, man. Okay. We've got another one here. So options for giving this cut away. Helmet glint changes, HUD disappears, absorption harps reset, or the bobber jumps.

• 36:27Speaker 01

  What's a bobber? That we, I love that we also.

• 36:30Speaker 00

  You'll see, you'll, it'll be obvious.

• 36:32Speaker 01

  I love that half the game is just playing against what TJ put in our minds to seed our reality.

• 36:36Speaker 00

  Yeah, yeah.

• 36:40Speaker 03

  No comment. All right, here we go.

• 37:00Speaker 03

  Okay, so.

• 37:01Speaker 00

  What, what was the bobber?

• 37:03Speaker 03

  So there's, that's the fishing line. So helmet, glint changes, HUD disappears.

• 37:08Speaker 02

  You gotta do it over time.

• 37:09Speaker 03

  resets or the.

• 37:11Speaker 01

  HUD disappearing is obvious because that just happened.

• 37:25Speaker 01

  Okay, I think I, I think I saw what happened. I'm like pretty sure I know what happened.

• 37:30Speaker 00

  I have no, I didn't see the bobber that I'm supposed to see, so I'm kind of.

• 37:33Speaker 03

  Okay, the bobber Casey's right here. This, he's throw, he's fishing. He fishes at the beginning.

• 37:38Speaker 00

  Yeah.

• 37:39Speaker 02

  Just in the very beginning.

• 37:41Speaker 01

  Okay, can you, can you just show it to me for a second? Show it to me for a second. Okay, keep going.

• 37:45Speaker 03

  There's pause at the beginning.

• 37:46Speaker 01

  Oh no, keep going with that, with that. Okay, I thought it was something completely different. I just want to see it again.

• 37:51Speaker 00

  I don't know. Yeah, okay.

• 37:54Speaker 03

  All right, yeah. All right, I'm gonna open the answers. It's okay, it's okay if you get them wrong. It's just a quiz. It's not a test. Okay, it's just a quiz. It's not that big a percentage of your grade for this semester.

• 38:03Speaker 02

  I want to win, TJ, okay?

• 38:05Speaker 00

  I, I legitimately have no idea. That would, that, that one is definitely at my sort of.

• 38:11Speaker 02

  I think.

• 38:12Speaker 00

  Having never played Minecraft. I'm like, I didn't even know what was going, I'm like, I don't even know what's on the screen right now.

• 38:17Speaker 02

  I'm with you, but I think.

• 38:18Speaker 00

  I have no idea what I'm looking at.

• 38:20Speaker 02

  I'm ready.

• 38:21Speaker 03

  Yeah, that's all right.

• 38:22Speaker 00

  Um, so I'm just, I'm just, I'm just picking a random.

• 38:25Speaker 01

  See, I thought the place he was digging was reset. Cause he liked, when he, when he went out, it really was the HUD disappearing.

• 38:31Speaker 03

  It really was again. And this time, well, we have a little bit of extra info, I believe.

• 38:36Speaker 04

  Triggered this full screen, but there's also a technical way to prove that this wasn't just going full screen like he claims it is. You see, Minecraft has what's called the GUI scale, which is an integer pixel scaling factor for how large GUIs render. For example, at a GUI scale of two, every texture is scaled two times such that every pixel from the texture becomes a two by two of pixels on the screen. At three, it's a three by three and so on. That means that the inventory should get increasingly sharper as you increase the resolution. As each texture pixel gets mapped to several screen pixels, which reduces the overall effect of the constant scale compression artifacts, making the image bigger while the noise stays the same size, makes the noise look smaller and less noticeable in comparison.

• 39:20Speaker 04

  But when we actually look at the video, we see the exact opposite. It gets blurrier as it zooms in, indicating resampling. Resampling.

• 39:30Speaker 03

  All right, so there you go. So that's, so there was two things there, but I didn't want to ask you guys, did it get blurrier as it zoomed in? Cause that one's a little trickier. So the point for this one that he's making, which I just thought it was fun because Prime, we had to deal with the slot at the tower, trying to get pixel perfect fonts. You remember this?

• 39:47Speaker 01

  Yes, I spent my life on, I still think about pixel perfect fonts to this day.

• 39:51Speaker 03

  Pixel perfect fonts. It's really hard to do and have them be readable on a card sized item. It's really hard. So. This is what happened. If we go back to the clip, well, actually I don't know how to go back that much into the clip. We're not going to do that right now, but we think about, I didn't add, I forgot to add. I forgot to add a player thing inside of this. Okay, cut me some slack. It's working pretty good.

• 40:13Speaker 02

  It is nice.

• 40:14Speaker 03

  He's saying, oops, guys, I went full screen. That's why randomly everything else in the game disappears. You can't see my bottom hotbar. You can't see my items, anything else.

• 40:25Speaker 02

  But what's not fair is I don't know when the HUD's supposed to disappear or not.

• 40:30Speaker 03

  It never.

• 40:31Speaker 02

  So it's always there.

• 40:31Speaker 01

  Cause I've done fullscreen in many games and typically when you go fullscreen, the game doesn't get dramatically worse out of nowhere. So that's why I was just so confused by.

• 40:38Speaker 00

  Interesting, interesting. Okay.

• 40:39Speaker 03

  Right. Yeah. So that's okay. That's okay. We are going to get to some actually interesting, uh, some actually interesting cuts, uh, in a moment.

• 40:48Speaker 01

  That one was interesting.

• 40:50Speaker 03

  That, that is good. That is the pixel, the pixel thing is good. Yeah, me too, Trash, because I'm rooting for you. Okay, Trash, I'm rooting for you. All right. Uh, all right, let's go, we'll go to the next one. Here we go. So achievements, so achievements in Minecraft are when you accomplish something new. It does like, it'll pop up sometimes on the top or it'll send a message in the bottom left. Items in the chest change. His aim is off, so like where the pixel thing is or what he's holding swaps.

• 41:19Speaker 02

  Swaps hands?

• 41:20Speaker 03

  So everyone's watching. Swaps hands, right, swap hands.

• 41:23Speaker 00

  So it goes from the left hand to the right hand.

• 41:25Speaker 03

  Right, that would be what a held item swap. Yep, that would be, yep. So here we go.

• 41:32Speaker 03

  We don't have a audio for this part because he narrates what goes wrong here. So just.

• 41:46Speaker 00

  Hit me one more time?

• 41:47Speaker 03

  Okay, hit me one more time.

• 41:47Speaker 02

  I need to see the options again.

• 41:49Speaker 03

  Yep. Okay, alright Trash, I can hear you.

• 41:52Speaker 02

  I have feedback for future games.

• 41:54Speaker 03

  Chest items change.

• 41:55Speaker 02

  Feedback for future games.

• 41:56Speaker 03

  Aim is off.

• 41:57Speaker 01

  Yeah, having the...

• 41:59Speaker 02

  We need the choices with the clip.

• 42:00Speaker 00

  Means that like, he's not pointing at the chest and it opens or something like that? Or what is that?

• 42:05Speaker 03

  Right, the aim is off as in it shifts sometime where it couldn't possibly have shifted, right? Like, showing that there is a cut, right? Showing that there's a cut.

• 42:14Speaker 00

  Understood.

• 42:21Speaker 03

  Nice parkour.

• 42:29Speaker 03

  I know, the "Oh, let's go", it's alright.

• 42:32Speaker 03

  Okay. Some of these are, they're getting better, okay? The cuts are getting better, not all of them are just massive lag spikes. You can guess, you can guess.

• 42:40Speaker 01

  Hmm. I didn't even see the achievement pop, so that's how stupid I am.

• 42:44Speaker 00

  I saw the achievement pop, but I don't know what that achievement is, so I'm not sure what that means.

• 42:48Speaker 03

  Oh, I see, yeah, that's tough, Casey.

• 42:49Speaker 00

  So like, I can't really get that one, but...

• 42:53Speaker 03

  Okay, here we go, let's see.

• 42:55Speaker 02

  Please, please.

• 42:57Speaker 03

  Casey!

• 42:58Speaker 02

  That was a lot, you just got lucky.

• 43:00Speaker 00

  Well, I thought when he went to the chest, it kinda didn't exactly, I don't know, that was all I could get, but...

• 43:05Speaker 03

  Yep, here you go.

• 43:05Speaker 02

  He told where it closed for like, no reason, basically?

• 43:07Speaker 04

  We can tell this because he's aim jumps. What I mean by an aim jump is when your cursor suddenly moves despite being completely still, both before and after. If this jump were just the start of a continuous motion, then that would be a whole different story, and it's something normal that happens quite often.

• 43:23Speaker 03

  Alright, that's good enough here, though, okay? Alright, following, nothing crazy on that one. Trash, you're really letting me down.

• 43:31Speaker 02

  Well, okay, okay.

• 43:33Speaker 03

  What, oh what, Trash, it's okay.

• 43:35Speaker 02

  Well, I heard the pigs, and it felt like that achievement popped up a little late. I'm not, I don't know what the timing there is, but that's what I was... I was timing the snort, the pig snort to the...

• 43:45Speaker 00

  Okay.

• 43:45Speaker 01

  Yeah.

• 43:46Speaker 03

  Okay.

• 43:47Speaker 01

  Man, I am bad at this game.

• 43:49Speaker 03

  That's alright, that's alright, Prime, that's alright. Here we go, that Prime, that just shows every programming trickshot you've ever done, we didn't put in edit it. So that's good, you're just an honest programming trickshooter, okay?

• 44:01Speaker 01

  Or I'm like, I'm open AI level employee trickshooter is all I am.

• 44:07Speaker 03

  Alright, this one, hotbar, that's the thing on the bottom, between your different items, you swap between it, it's your current inventory, right? Um, debug menu, there's like...

• 44:15Speaker 01

  TJ, a hotbar is actually a thing at Whole Foods. It's where it's...

• 44:19Speaker 00

  It's like where the soups and stuff are.

• 44:21Speaker 03

  Or Air One, if you're Trash, probably.

• 44:24Speaker 02

  I was gonna make an Air One joke, but I did not.

• 44:26Speaker 00

  Yeah, yeah, yeah. Upscale, bro, LA times!

• 44:30Speaker 03

  Right, yep. Debug menu closes, that's the, uh, there's like, gonna be some text in the top left that tells you like, where you are in the world. That's the debug menu. Uh, C is a lag spike again, in the colloquial sense, Casey, of course, or D, all of the above, okay? Alright, so let's watch.

• 45:01Speaker 03

  Okay, we'll watch it again.

• 45:02Speaker 02

  Wait, what, options, options, hold on.

• 45:05Speaker 03

  Oh yeah, of course, Trash, sorry.

• 45:05Speaker 02

  Hotbar changes lag spike, debug menu.

• 45:08Speaker 03

  Debug menu closes, lag spike, or all of the above.

• 45:11Speaker 02

  Okay, I think I got this.

• 45:13Speaker 03

  Yep.

• 45:20Speaker 01

  Oh my gosh, that's so hard, oh no, I can't tell.

• 45:24Speaker 02

  Got it, got it.

• 45:26Speaker 03

  Okay, alright, okay, open answers. Okay, wow, I love it, I love when you guys are, you're ready to go. Great job, everybody!

• 45:34Speaker 01

  Yeah, yeah.

• 45:35Speaker 03

  Great job, okay? Alright, we'll let him explain.

• 45:39Speaker 04

  And yet another inexplicable lag spike, conveniently occurs where 13 minutes go missing, Dream suddenly jumps forward, his debug menu closes, and his slots change. Yet another cut.

• 45:50Speaker 03

  So there you go.

• 45:51Speaker 01

  That was a pretty rookie cut, honestly, because that item swap change, like, that's...

• 45:55Speaker 02

  That was blatant.

• 45:56Speaker 03

  And...

• 45:57Speaker 00

  And the debug menu just stopped having the X and Y up in the corner, so that's pretty...

• 46:03Speaker 03

  One of the things we're gonna get into too is like, a bunch of the stuff in the hotbar and some other things, they've got some animation timers and some other things, which, part of the reason that I thought this would be fun to discuss with, uh, Casey. So, we've got another one here, okay? Ready? Debug values jump. Uh, so, the values up at the top left could change.

• 46:23Speaker 01

  Oh, I love that B and C are in competition with each other. That's like, how much of our ability to understand Minecraft really works here.

• 46:29Speaker 00

  What is item name, where are the item names being shown? Like, when you hover over an inventory slot, or...

• 46:34Speaker 03

  Yeah, so right here, oh, like, right here.

• 46:37Speaker 00

  Okay.

• 46:37Speaker 03

  Yep.

• 46:38Speaker 02

  Lingers, okay.

• 46:40Speaker 03

  Item name lingers, or a small lag spike, okay, on this one.

• 46:42Speaker 02

  How are they supposed to work?

• 46:57Speaker 03

  Okay, we'll, we'll...

• 46:58Speaker 02

  Hold on, time, I feel like this...

• 46:59Speaker 03

  Debug values jump, item name?

• 47:01Speaker 02

  Like, Minecraft player, because I don't know the behavior of item name.

• 47:07Speaker 03

  That's okay, Trash, that's okay, we're gonna get to the programming stuff soon, we're almost there. Then you'll have a chance.

• 47:17Speaker 01

  Like, can you do it one more time? Hold on, one more time.

• 47:19Speaker 03

  Yep, of course, of course, of course, I love, get your investigation on, Prime.

• 47:23Speaker 01

  I'm feeling good about this one.

• 47:26Speaker 00

  I couldn't tell this one, sorry.

• 47:30Speaker 01

  Alright, I think I know.

• 47:32Speaker 03

  Okay.

• 47:34Speaker 03

  Do you guys want one more? One more chance, are we good?

• 47:36Speaker 01

  I feel confident.

• 47:37Speaker 02

  What was the fourth, what was the fourth small lag spike? Okay.

• 47:41Speaker 03

  Small lag spike, here, we'll just open the answers, we'll open the answers.

• 47:46Speaker 01

  Feeling confident, I think I can even explain it.

• 47:49Speaker 03

  Oh, okay, I love Prime, I love that. Prime, call your shot beforehand.

• 47:54Speaker 01

  Yeah, I'm gonna call my shots.

• 47:55Speaker 02

  I didn't answer yet.

• 47:57Speaker 00

  No, no, I answered it.

• 47:58Speaker 03

  Not yet, Prime, Casey and Trash still have to guess.

• 48:00Speaker 01

  Alright.

• 48:01Speaker 03

  But I'm saying, before I reveal what the right one is, I want you to confidently explain.

• 48:05Speaker 00

  I don't know, this was a sublime guess, I have, I had no idea.

• 48:10Speaker 03

  Alright, so here we go, locked answers, alright, Prime?

• 48:13Speaker 01

  Alright, so, if you looked at the bottom part, I was looking at the item names, and out of all the times, one of them hard disappears fast, and the rest fade out. And so, the dirt hard disappear seems to me that it was a fast vanish versus the other ones fade.

• 48:27Speaker 02

  I thought because he might...

• 48:29Speaker 00

  So I guessed, I just blind guessed that one. Okay, that, I probably should not receive credit for that, for that one, just to be clear.

• 48:36Speaker 02

  I was so close to picking B.

• 48:39Speaker 00

  I looked even, and I couldn't tell.

• 48:40Speaker 01

  Did I get it?

• 48:41Speaker 03

  Here.

• 48:42Speaker 01

  Did I get it?

• 48:42Speaker 03

  You did, Prime, you did. Here, I'll even open this up for you, you guys can still see my, see my screen, right?

• 48:48Speaker 00

  Yeah, yeah, yeah.

• 48:49Speaker 03

  Uh, didn't go, it didn't, oh, did it not go to the right? Oh, did it not? Here, one sec, let me try and go back to this.

• 48:54Speaker 02

  I'm so upset.

• 48:56Speaker 03

  Uh, for some reason, I can't, I can't see where's, oh, man, okay, we'll come back to this later, or I'll find it later. Uh, oh, it's one of the, it's one of these constants, it's one of these constants here, I don't remember which one. Uh, there's one though, where it says that it takes 10 ticks always.

• 49:12Speaker 01

  I thought I saw auto fade speed. Yeah.

• 49:15Speaker 03

  Oh, did, okay, did we, did we?

• 49:17Speaker 01

  Yeah, just look down, go down a little bit more, look, okay.

• 49:21Speaker 01

  Uh, keep going, I saw auto, auto fade speed, there you go, auto speed, fade speed factor, oh, that's...

• 49:26Speaker 03

  Oh, right here. Oh, but it's a factor or something else, that's okay, we'll get to that, we'll come back to that later. But, here we go, we'll play, we'll play that explanation.

• 49:33Speaker 04

  He changes direction and starts tiling to the side where he opens his inventory to grab his pickaxe and cuts upon closing. It's pretty well done, but he missed one little thing. The item label for his crossbow disappears as he cuts instead of fading out like it normally does.

• 49:46Speaker 03

  Yep.

• 49:47Speaker 01

  Not dirt.

• 49:47Speaker 02

  I did the lingers because the diamond pickaxe faded slowly. I don't know if that was supposed to happen.

• 49:52Speaker 03

  Nice.

• 49:53Speaker 01

  There's two fades, there's two fades, so that's how I know.

• 49:56Speaker 03

  That's, yep, nice, Prime, there was enough info in there to deduce it.

• 50:01Speaker 02

  I can't get last place, I can't get last place. Next one.

• 50:04Speaker 03

  Legendary game dev Casey is doing really well here.

• 50:08Speaker 02

  I have to get this right.

• 50:09Speaker 01

  That's why. Alright, 6 out of 9, let's go.

• 50:11Speaker 03

  Trash, you do, bro, you do, okay? You do. Alright. What gives this cutaway? Achievement pops late, again, we already discussed this. Coal burns too fast, so you put coal into a furnace, it burns over time, okay? Render distance changes, right?

• 50:26Speaker 01

  Oh, that's crazy.

• 50:27Speaker 03

  Once you're in, once you're inside of a, uh, inside of a chunk in Minecraft, there are like, these squares, and it sort of only loads part of the world at a time. Like, and you can see other distances. It shouldn't change, generally speaking, there's some exceptions, but the Minecraft people will tell you when it can change, that's okay?

• 50:45Speaker 00

  I've got a new teammate.

• 50:47Speaker 02

  Let's go!

• 50:48Speaker 00

  New teammate arrives, Molly Bean, we're gonna do this, we're unstoppable now.

• 50:54Speaker 03

  Okay, or the ingots are smelting too fast, okay? The ingots is in the metal that's inside the furnace.

• 50:59Speaker 01

  The little metal thing, you go smelt it.

• 51:01Speaker 02

  Got it, got it, got it.

• 51:02Speaker 03

  Yep, so you... So he built a tower up to the top, and his friends are trying to come up and knock him off to kill him, okay?

• 51:12Speaker 03

  That... So he's cooking various items in the furnace or smelting, you can cook, uh, raw meat to make it not raw, and it's better for you.

• 51:20Speaker 01

  Yep.

• 51:21Speaker 03

  Um, not usually steak tartare is not the main like, food item. So he towered really high, and he's basically building stuff to protect himself.

• 51:41Speaker 02

  I need choices and then another.

• 51:45Speaker 01

  Trash is taking this way too seriously.

• 51:47Speaker 02

  I just can't get last.

• 51:48Speaker 03

  Okay, there's an achievement and it pops too late, or no. Coal burning too fast, render distance changes, or the ingots, the metal.

• 51:56Speaker 01

  It smelts too fast.

• 51:57Speaker 03

  Any of these four, we'll watch one more time, okay? This is also an important clip to understand because a clip is coming up right after this. Or not right after this necessarily here, I'm just gonna mute this so you can, I can talk over this.

• 52:11Speaker 03

  And basically in like a hunt like this, you build this huge tower so that he can have the high ground. Of course, everyone knows after watching Star Wars, that's very important. And he can like actually take time to do the crafting he needs to get his items, okay? So he's like watching his friends come up to hunt him down while he has this thing built up really high, okay? Which also means you can see very far and some other things like that, okay?

• 52:38Speaker 02

  So is he pretty much dead meat at this part, at this point? Cause as soon as.

• 52:42Speaker 01

  Why just one more time? Great question. Just for, just for, okay, fine.

• 52:47Speaker 03

  Oh, yeah, yeah, no, I can play it again.

• 52:48Speaker 01

  I just had one more question.

• 52:49Speaker 02

  I don't know where to look.

• 52:53Speaker 03

  That's okay. The achievement's also sometimes popping this bottom left area down here.

• 52:57Speaker 01

  We can be done, I don't even know what I'm looking for anymore.

• 53:01Speaker 03

  That's okay.

• 53:01Speaker 00

  Yeah, I got nothing.

• 53:03Speaker 03

  Yeah, that's okay, that's okay.

• 53:06Speaker 03

  Trash? You can do this, buddy.

• 53:10Speaker 02

  Ah, let's go.

• 53:12Speaker 03

  Okay, lock it in, reveal.

• 53:14Speaker 00

  Let's go.

• 53:15Speaker 02

  Prime. Damn it!

• 53:20Speaker 03

  Trash, you were so close, you were so close. So, uh, here we go, we can look at the cooking recipe for, uh, this, uh, this tells you some stuff about the cooking. This one though is the one that I wanted to click open here. This is, uh, the data for iron ore and it takes 200 ticks to cook iron ore into an iron ingot, okay? Minecraft's 20 ticks a second, so that's 10 seconds. Okay, more ingots come out of the clip than they had time for. So what was funny is as they were doing this analysis for the, the actual final thing that we're about to reveal soon of how they found a bunch of these cuts, they also just found a bunch of other things like this.

• 54:00Speaker 04

  When I got to the tower, I proposed that someone cross-check the furnace timings. And so Leila went ahead and did so. And what she found was shocking. There's a whole 9.3 seconds of footage missing from the video after he loaded his safety. That means he somehow smelted almost an entire extra piece of iron. When he opens the left furnace here at 1420 and nine frames, it smelted 0.56 ingots. He then empties the left furnace here and then 13 seconds later, he opens the furnace again and witnesses it smelt its third ingot only 15.1 seconds after he opened it the first time. Do you see the problem here?

• 54:43Speaker 03

  There we go, that's what we already went over, so.

• 54:45Speaker 01

  Dang.

• 54:46Speaker 03

  Right? So that's, so they're getting pretty serious about this, which I have to say I love. I love that they're figuring out the timings of these and me like, just literally not possible, just literally not possible.

• 54:57Speaker 00

  This is better than like the, than like the baseball, like instant replay New York. Like check the umpires thing, like these, they need to hire these people.

• 55:07Speaker 03

  This would be a, these guys would be the ultimate review officials.

• 55:11Speaker 00

  Yeah, they clearly care way too much about it.

• 55:15Speaker 03

  I like it personally.

• 55:17Speaker 00

  I have no problem with it, I'm just saying like, wow, that's a lot of like scrutinizing some random YouTube video, but okay.

• 55:24Speaker 03

  Yeah, yeah, yeah. So, alright, so the ingots were smelting too fast. Alright, Casey, I got another Ask Casey question for you. Okay.

• 55:35Speaker 03

  A lot of Minecraft recipes, as we showed just briefly here, you know, they're kinda in this, they're just like declared like this is some data, right? When in like making a game or building an engine, when would you do kinda like one versus the other? Try and do stuff more with like declarative, just here's JSON blobs, whatever, versus like, cool, we're just gonna write out code?

• 55:57Speaker 00

  Well, uh, there's a lot, there's a lot of context that's kind of necessary to determine which one of those is likely gonna happen.

• 56:07Speaker 00

  You know, I would highly recommend not using JSON, but if you're just talking about like, we could write this in a data format that wasn't gonna cause problems, so let's just say data versus code. The question is usually, uh, comes down to a couple of things. The first one is, do you need users to be able to modify this thing? So that's an easy question you can ask that lets you know right away, like if users have to be able to modify this thing, then it had better be in data because usually don't want to ask users to code. However, there is a second, there is a corollary to that is like, okay, if the answer to that is yes, do you give your users a scripted, like, do your users have access to the code?

• 56:48Speaker 00

  Cause sometimes that answer might be yes as well, it's like, yeah, it's like this modding thing and they can all have access to it, or we just, we give them all a subscription to, to OpenAI and they're supposed to decompile the code. I don't, who knows, like, if you have some explanation for why the user could somehow modify it without that, then maybe that's not true. So that's thing one. Thing two is, uh, kind of the more important part, which is, do you need this thing to be represented in a way that will be inconvenient in data? And this is something that a lot of people kind of, um, I guess I would say they don't appreciate the ramifications of the decision they make here very often. Because usually they're just like, well, I can just data drive it, it's, that's always better, right?

• 57:32Speaker 00

  And the problem with that is that when you decide to data drive something, what you're effectively saying is that you're going to restrict sort of the, the like, class of things that could happen in this part of the game. So it means that that thing that you're just setting the numbers on, it will behave the same as all the other things do, just with different parameters, right? So you're homogenizing some part of your game usually when you do this, right? Assuming that there's more than one thing that can be, can be set this way. And so what can unfortunately happen in some cases is if you decide to do this data driving prematurely

• 58:13Speaker 00

  and you haven't thought it through, you can get in this situation where like, oh yeah, there's like 500 pieces of armor in this game and they all feel exactly the same, they just have slightly tweaked like numbers, right? And so the other thing that you kinda have to know from a game design standpoint, and I'm the wrong person to talk about, you know, the finer points, so I'm not a designer, but is, have you ensured that moving to data driving won't remove some algorithmic expressivity that you may have needed in the design? To make these things feel more tangibly different by making it not just a number, uh, but more of a sort of an algorithm. Now you can still, you can patch over that by having like, in your data driven, you can have like type fields

• 58:54Speaker 00

  that help select different algorithms, so you know, there's ways of trying to like, you know, merge those two. But again, it's just, it's part of the sort of gestalt, like thought process you should be going through is like, how am I gonna make sure that I'm not just gonna end up with, you know, um, a billion different types of armor that are all feel exactly the same, right? So, uh, so those are the things you should think through. In general, the only real hard do and don't is just never use JSON, that's a terrible idea. It's a very, very bad format, uh, for storing things that you would be using at runtime, in my opinion.

• 59:25Speaker 03

  Thanks Casey, Trash, Prime, any questions to follow up there?

• 59:29Speaker 01

  What is the, besides for like, because,

• 59:34Speaker 01

  one thing I understand about JSON is that it kinda, it, you know, by being self-discovery, it kinda stinks and there's like a lot of things you have to do, you have to do validation of things coming in, you also, you know, it just like puts memories and goofy locations, all this kinda stuff, and I know that there's reasons why it's bad, generates a lot of garbage, but why is it bad overall for stuff as opposed to like a one-time cost? Like, okay, hey, we read all this data and it kinda, you know, it kinda stinks, but we, then we put it into the right order and everything's super fantastic and now everything's great, why does it continue to suck? Or what is the problem beyond that?

• 1:00:06Speaker 00

  For the reasons that you're saying, uh, is exactly why it sucks, right? Because if you just have some, you know, um, a very typical thing that you'll see in games is there's just some like binary blob format, so there's no parsing time, it just loads immediately. And then it can also have like a nice little custom editor, usually game and studios have a little custom editor that lets you just edit those things and it, it'll check like for validity and stuff like that. JSON has none of that, right? JSON is just an arbitrary text format, you put whatever in it you want, which means that 100% of your tooling now is like in runtime in the game, doing that validation on a thing it loaded out of text it had to parse. I hope you got the parser right, I hope you didn't have any like problems in there, I hope it's fast enough and not slower than load times like Grand Theft Auto, for example, um, and so on.

• 1:00:51Speaker 00

  So it's just, it invites just a tremendous amount of unnecessary work that you're doing. And even when you do that unnecessary, unnecessary work so that, you know, it is fast enough and, you know, you have validated, it's still just not that convenient because now the user has to edit this JSON, which editing JSON is not a pleasant experience, not a good, uh, experience for the user anyway. So I just see there be no real benefit to JSON other than it's a standard. And maybe, you know, if you have a really overwhelming reason why you want, uh, that standard to, to be used in your game, then maybe you have a case, uh, but otherwise I just think it's, it's a pretty poor decision in, across pretty much all axes that you might be considering it, right, that might be considering.

• 1:01:32Speaker 03

  Okay, well, start scoring some points, please.

• 1:01:34Speaker 02

  I know, I'm trying, that's all I'm thinking about right now.

• 1:01:38Speaker 03

  Alright, here we go.

• 1:01:40Speaker 00

  Oh man, these are getting hard, okay, alright.

• 1:01:42Speaker 03

  They're getting harder, they're getting harder, okay? The lava pattern, right, so there's little particles that come off of the lava and like the top texture of the lava can change over time, so does it jump in some unreasonable way? Uh, the hunger bar, so in Minecraft, uh, on the hunger bar, it like moves up and down, uh, and like, kind of just like wiggles, it jolts up and down.

• 1:02:02Speaker 00

  Oh, okay.

• 1:02:03Speaker 03

  So did it, did it do something unexpected here?

• 1:02:05Speaker 00

  Okay.

• 1:02:06Speaker 03

  Does the axe glint jump? So when you, uh, enchant something in Minecraft, uh, it starts getting sparkly effects, obviously. That's why it's sold a trillion copies.

• 1:02:17Speaker 00

  That's what enchanting means, I mean, if you've played video games, you know that enchanting means there's some frickin' sparkles, okay?

• 1:02:23Speaker 01

  It's real, it's quite enchanting to see something sparkle.

• 1:02:27Speaker 03

  Or is it that there is a cloud who jumps in the background, okay?

• 1:02:31Speaker 02

  Took a screenshot of these and I'm having, I have them left side of my screen.

• 1:02:34Speaker 03

  Oh, Trash, good thinking, dude, good thinking.

• 1:02:37Speaker 00

  He's locked in, I keep saying locked in, it's my new word.

• 1:02:40Speaker 03

  Well, he is though, he is.

• 1:02:41Speaker 00

  He's locked in.

• 1:02:44Speaker 01

  I would assume there's no clouds for this particular one.

• 1:02:48Speaker 03

  You never know, sometimes there's gasps and some other things.

• 1:02:50Speaker 02

  Uh, did it phrase for you?

• 1:02:52Speaker 03

  Nope, that's, that's the end, that's the end, let's start it over. We'll watch it a few times, it's getting tougher.

• 1:02:56Speaker 00

  It's a really short.

• 1:02:58Speaker 03

  Yep.

• 1:03:01Speaker 02

  Seen it.

• 1:03:02Speaker 03

  Okay. Alright, we'll restart one more time and I'll... Do you guys need to see the beginning part again? Or are you guys...

• 1:03:08Speaker 00

  I didn't need a few times because I was expecting kind of a longer clip. I'm sort of settling into it now.

• 1:03:13Speaker 03

  Yup, yup.

• 1:03:14Speaker 01

  TJ, when you did talk about some of these things, like you brought it up for a brief moment. Is this the one that I would pre-know?

• 1:03:21Speaker 00

  Alright.

• 1:03:21Speaker 01

  Because you did mention one thing.

• 1:03:22Speaker 03

  I don't know, I'm not gonna tell you, Brian. You can take the gamble about whether it's the thing that I mentioned.

• 1:03:26Speaker 01

  I'm trying to do the opposite. I'm trying to say like, hey, maybe you should like, I should intentionally pick wrong.

• 1:03:30Speaker 03

  Yeah. No, no, I'm saying, I'm saying, Prime, you should do the one that you think is the best one. Is everyone... Do you guys need another time or you're good?

• 1:03:39Speaker 01

  Cut. I actually just wanted my question answered so much, I didn't even look at what I was doing.

• 1:03:44Speaker 03

  Alright, here we go.

• 1:03:46Speaker 03

  Casey?

• 1:03:47Speaker 02

  Are you kidding me? No way, dude.

• 1:03:49Speaker 00

  That was pretty, it was pretty brutal. Like, it was up at the top and it goes back down.

• 1:03:54Speaker 02

  Okay, go back, go back.

• 1:03:55Speaker 03

  Casey! What a legend! Casey, what a legend! Alright. Okay, what a legend. Here we go. We can examine some Minecraft code over here. Um... Casey, maybe, um... You can walk us, you want to walk us through this code?

• 1:04:12Speaker 00

  If I can see it, let me try and make that big on my screen. Oh yeah, I can see it, okay, sure.

• 1:04:15Speaker 02

  I want to see the clip again, this is BS, there's no way.

• 1:04:18Speaker 03

  Okay, Trash, we'll watch the clip one more time here.

• 1:04:20Speaker 02

  I was looking for that.

• 1:04:21Speaker 00

  It goes dark, it goes dark too. It makes it hard to see because it goes dark, like, the screen kind of dims because he's in an UI or something.

• 1:04:29Speaker 02

  I thought that was just part of the...

• 1:04:30Speaker 00

  Yeah, Casey, to hide it.

• 1:04:31Speaker 02

  I need the clip, clip, clip!

• 1:04:33Speaker 00

  Go, okay, show me.

• 1:04:34Speaker 03

  Okay, okay, well, okay, we'll play the clip.

• 1:04:37Speaker 04

  ...right in the face. The very axe Dream stole earlier. It's enchanted, and enchanted items have a continuous, smooth, perfectly looping glint animation. If there really was a cut, then maybe this animation would suddenly jump. And so I went back, frame by frame, and there it was.

• 1:04:56Speaker 00

  See it?

• 1:04:57Speaker 04

  Clear as day. The enchantment glint on his axe just completely jumps as he closes his inventory.

• 1:05:03Speaker 03

  Oh!

• 1:05:04Speaker 04

  This part's highlighted, and now suddenly this part is instead. I mean, it's a pretty blatant cut.

• 1:05:08Speaker 02

  If you look at the lava, there's clearly a big blip in the lava too.

• 1:05:15Speaker 03

  That could be, that could be, but we're gonna do best answer only, sorry Trash. So, you can actually see right here, here is, I believe this is the right code here.

• 1:05:25Speaker 01

  It looks like setup glint texturing.

• 1:05:27Speaker 03

  Yup, that's the glint texturing. Um, I don't know. Casey?

• 1:05:32Speaker 00

  Wow, okay, interesting. Alright.

• 1:05:38Speaker 01

  Why is the same, why is the same number... Why is 11,000, why is 110,000 divided by 110,000 a thing?

• 1:05:48Speaker 00

  Well, I mean, he, so, he's trying to produce a, um, you could think of it as like, things that want to be at a different period. Right, so, so the mod is like...

• 1:06:01Speaker 01

  Oh, whoopsies, I just realized there's a mod right there, yeah, my bad.

• 1:06:03Speaker 00

  Do you know what I'm saying? So like, the mod is producing something that goes in that period because it's milliseconds, right? So it's something that's periodic on that scale. And then the divide turns it into a uniform float so that you can, you know, do, multiply it by what you actually want to do it by, right? So it's trying to produce a number between 0 and 1 at that point.

• 1:06:22Speaker 01

  Yeah, and of course, this is the, this is the important bit for finding the cut.

• 1:06:26Speaker 00

  Yeah, that was, that was the really interesting part to me was the fact that this just uses getMillies. So, yeah, not how I would have done it, but, you know, that, that means you're locked to probably a real, the real clock. I don't know, I don't know, uh, Java, so I'm not sure, is that like an actual millisecond timer that's external from the game? Or is that, did the game implement that? Is UtilTheGames class, I don't, I don't know.

• 1:06:47Speaker 01

  And does matrices run backwards there? Because I always thought it was, identity, scale, rotate, orbit, translate. And that looks like we did translate, rotate, scale.

• 1:07:00Speaker 00

  Um, oh, I don't know. So sorry, could you say, could you say what, what you were saying?

• 1:07:04Speaker 01

  I do, I always do scale, rotate, orbit, translate, if you want to like, place things. Is rot is how I remembered my head. When you want to place something out there, you don't screw it all up where you accidentally orbit when you meant to rotate.

• 1:07:16Speaker 00

  Um, I, I would, I would agree with you, but I was just trying to, what I was just trying to say there is, it's, it's about the order you want them applied in though, right? So, he's trying to produce this kind of weird effect of this thing shimmering through it. So, that, it may just be that he wants the translation, he wants to move it, then spin it, then scale it. Might be what he was trying to do, right?

• 1:07:36Speaker 01

  Okay.

• 1:07:36Speaker 00

  Uh, you know what I mean? He might want the translation to be scaled, I think is what I'm trying to say.

• 1:07:42Speaker 03

  Casey, I got another one for you here, which you have already alluded to, right? Which is that the glint is running on real time, I believe, uh, later. No, I do have some of the code later, I think. It is like, similar to Java's, uh, just like, literally the, a regular clock situation, not like, part of the game clock or anything.

• 1:08:01Speaker 00

  Uh, can I kind of say one more thing about the, uh, Prime's question first?

• 1:08:05Speaker 03

  Yeah.

• 1:08:06Speaker 00

  So, uh, just in case this matters to anyone out there who's confused about matrices, uh, I just also kind of wanted to add, we could, we could talk about matrices sometime. It's been a long time since I've talked linear algebra, but I just wanted to point out, so when you think about how matrices work, it's also important to remember that it depends on how you're imagining the multiplication is working. Like, how you decided to implement it, right? If you use the standard math notation, and you implement your library so that it multiplies matrices the same way that they would be multiplied if they were written out, uh, mathematically. So like, you know, when you're reading left to right in math, you would read left to right in your code. So you're, you're implementing the same operators that a math textbook would do it.

• 1:08:48Speaker 00

  Then, it's really just about, uh, the order in which you're imagining the multiplication happening. And if we, could we go back to that snippet in a second that was there?

• 1:08:57Speaker 03

  Yeah.

• 1:08:58Speaker 00

  So you'll notice here, there's, nothing's written out. It's like, dot translation.

• 1:09:03Speaker 03

  Yeah.

• 1:09:03Speaker 00

  So, to me, like, I don't even know, it's one of the reasons I don't, I really dislike that format of writing math as well, is when you do dot translation, I don't know, is that a right multiplied translation? Or a left multiplied translation? Because if you took two matrices, and you had the translation on the right, then that would apply the translation first. But if you had the translation on the left, that would apply the translation second. And so when you do dot translation, it's like, now I need to know what your convention is. Is your convention that you, when you say dot translate, that you're concatenating on the left? Which means the translation will happen after everything else that's in this matrix? Or is the convention that the thing will happen on the right, which means it will happen before everything else that happened on this, on thing, right?

• 1:09:45Speaker 00

  And so, I, I, you know, I, I don't mean to criticize this particular thing, like, it's not like some huge crime that it's doing this, but like, this is just one of the reasons that I prefer not to write code that way. Because then I know that anyone else who doesn't know my convention, they read it, they have no frickin' idea what this does. They could guess that it's probably on the left translation, because you would imagine that they're going to go like, the things that happened before should happen first, right? But it's purely, that's purely optional. That's not a, an aspect of matrices, that's an aspect of which, which order you decided to do the multiplication. Does that make any sense, what I'm saying? Hopefully?

• 1:10:18Speaker 03

  Yup.

• 1:10:18Speaker 00

  Okay.

• 1:10:20Speaker 03

  So, I feel you.

• 1:10:22Speaker 01

  By the way, Casey, I know you haven't done linear algebra for a while, but linear algebra, kind of the rage, all the rage these days. It has feelings now.

• 1:10:29Speaker 03

  Pretty big deal these days.

• 1:10:29Speaker 01

  It kind of has feelings, so you should say sorry to Dario.

• 1:10:33Speaker 00

  Only, only if we restrict it to being like, frickin' 4 bits, or like, 6, not if we do it in full 32-bit float apparently, right? You gotta keep those bits low.

• 1:10:43Speaker 01

  Grituitous 32-bit float.

• 1:10:44Speaker 00

  Yeah, god, all those bits. Can't afford them.

• 1:10:47Speaker 03

  Just going to waste.

• 1:10:48Speaker 00

  Can't afford the memory, can't afford the bandwidth, can't afford any of it. Okay, the glint runs on real time, not game time, what should run on the wall clock? Um, so, in my opinion, uh, the way that I always write this code, there is no such thing, like, I would not even allow any module to, in the game code, to compile with access to a wall clock. Full stop.

• 1:11:14Speaker 01

  Okay.

• 1:11:14Speaker 00

  That's just like a hard no, it's like, that's not even in the includes, like, you're not, you don't have access to that module, right? And the reason for that is because I like to treat the game code side of things as effectively, from, from like, if you think of it almost like, from this, this layering perspective of like, operating system, engine, game code. I want the game code to be completely reproducible. It should have no way, right, ideally, of getting entropy from anywhere that I have not pushed into it as part of the update call.

• 1:11:53Speaker 00

  So, it can have something that runs on wall clock time in quotes, but that wall clock time is going to be some value I called wall clock that I stuffed in the head-end structure that gets, that's propagated out. And it may not be wall clock time. Why? Because I want to do things like, record runs of the game, play them back, and have them be reproducible. Exactly. I don't want to have the, like, oh crap, now I've got to go like, put this piece in that like, intercepts calls to the system timer, like, through the thing, like, I don't want to deal with it. And I just, it's easier if nobody ever does that in the first place, right? Because otherwise, you've inherited this mess. So, so typically, what I do at the start of things like this is, I basically have this thing that's like a timing struct that gets fed forward.

• 1:12:38Speaker 00

  And it's got things like, how many audio samples do I want you to think have elapsed? How much wall clock time do I think, uh, want to last? And how much game clock time do I want you to think has elapsed? And those are just completely set by me, and they may or may not be truthful, and everyone else is just expected to use them as authoritative.

• 1:12:56Speaker 03

  That makes sense to me. We made this mistake, uh, in Mordoria round 1, so...

• 1:13:01Speaker 00

  Yeah, yeah, it's an easy mistake to make, it's one of those things that you learn over time, because you're like, oh crap, I want to start doing these things. And like, it's the same as don't call a random number generator that's not under your control, right? It's like you, the random, random number generator is another input into the system. It's supposed to actually be under your control that produces a, a thing that you know what it is, right? Uh, so, anyway.

• 1:13:22Speaker 03

  Right.

• 1:13:22Speaker 00

  Yeah.

• 1:13:24Speaker 03

  Alright. Okay.

• 1:13:26Speaker 00

  Alright.

• 1:13:28Speaker 03

  Things are about to get really tricky.

• 1:13:30Speaker 00

  Uh oh.

• 1:13:30Speaker 03

  Okay. Cause now we're introducing the possibility, maybe not even a cut in this one. Okay. Is it A, the hunger bar shakes in a way that is impossible? Does the axe glint jump in an impossible way? Or is it the clouds? Or this time, is it actually a perfect cut?

• 1:13:54Speaker 00

  Okay.

• 1:13:55Speaker 03

  Let's play it.

• 1:13:58Speaker 03

  Oh no. Okay, well, there we go. It says.

• 1:14:04Speaker 03

  I'm gonna, I'm gonna mute it since he tells you what it is here. I didn't realize that I put the wrong clip in for this one. My apologies. Okay, there's a clip, there's a cut right there where the three to four goes. Uh, so you don't pick D, okay? Don't pick D.

• 1:14:19Speaker 02

  We know what it is. We know what it is. This one, this was, uh, everyone knows what it is now.

• 1:14:23Speaker 01

  I like how it's zoomed in.

• 1:14:24Speaker 03

  Clip it, fine.

• 1:14:25Speaker 02

  I don't know what it is.

• 1:14:26Speaker 03

  Put it. No. Okay, well Trash, you don't get another try.

• 1:14:29Speaker 00

  Trash. Yeah, that's a great point. That's a great point, Trash.

• 1:14:33Speaker 01

  You could not have another try after that one.

• 1:14:35Speaker 00

  Here, here, we're, we're gonna play.

• 1:14:37Speaker 03

  I trusted you to tell me whether my Pokemon were valuable or not. I, I even added, I forgot that I didn't mute this. I even added this little black bar over this spot.

• 1:14:47Speaker 02

  I don't even know what that's supposed to be.

• 1:14:51Speaker 03

  It's blocking him saying the words.

• 1:14:53Speaker 00

  Oh.

• 1:14:54Speaker 03

  But this is where, this is where it should give it away, Trash. It's even better.

• 1:14:57Speaker 00

  Can you get it now, Trash?

• 1:15:00Speaker 03

  Okay, okay Trash.

• 1:15:02Speaker 02

  I was looking like elsewhere.

• 1:15:03Speaker 03

  Lock in your answer here.

• 1:15:04Speaker 00

  Yeah.

• 1:15:05Speaker 03

  Okay, lock in your answer, Trash. Okay.

• 1:15:07Speaker 00

  Yeah. There we go.

• 1:15:09Speaker 02

  I was actually kinda terrified. I was about to.

• 1:15:13Speaker 03

  Alright. Here's where things get very fun, okay? Here's where things get very fun.

• 1:15:19Speaker 04

  Well, as we reach the most impressive hidden cut I've ever seen, the jolts narrow this down to a one-second window. From here to here. He jumps down, stops for a split second, cuts, and then continues, not leaving anything on screen to give it away. However, based on his food level of 17, these jolts should have really happened 52 ticks or 2.6 seconds apart, not one. And while I think.

• 1:15:49Speaker 03

  So, this is where it gets really fun.

• 1:15:51Speaker 00

  I never would have caught that.

• 1:15:53Speaker 03

  Yeah, yeah, right, I know. So this is where this gets really fun, uh, because, well, he's, we're gonna show the code in here. I'll, I'll come back to the MSRC thing. This clip right here, which I'm gonna start right here, this, if you recall, this is from back at the top of that big tower. This is the most important moment maybe of Dream's career. This, uh, this is called the boat clutch, okay? So watch what happens here. He gets knocked off by his friend.

• 1:16:22Speaker 03

  So there's a bug in Minecraft where you can place a boat down and enter the boat, and if you place it down fast enough, you won't take fall damage. Normally, he's about to die here. Crafts the boat.

• 1:16:33Speaker 00

  Okay.

• 1:16:34Speaker 03

  In air. Crafts the boat in air, right?

• 1:16:41Speaker 03

  And so this is supposedly, right? This is being presented as this happens in real time.

• 1:16:47Speaker 00

  Organically, right.

• 1:16:49Speaker 03

  So the options are water flow jumps, the hot bar with his items on the bottom changes.

• 1:16:54Speaker 00

  Got it.

• 1:16:55Speaker 03

  Something about the boat desyncs or the sun moves in the sky. Okay, we'll watch one more time here. This is our last clip.

• 1:17:05Speaker 02

  There's not even a sun, right?

• 1:17:06Speaker 03

  Okay.

• 1:17:07Speaker 00

  Alright.

• 1:17:09Speaker 03

  Oh, sorry. You wanna watch?

• 1:17:11Speaker 00

  I got it. Or I, or I think I got it.

• 1:17:14Speaker 03

  Lock your answer, Trash. Lock your answer.

• 1:17:16Speaker 00

  I'm not so sure about my answer anymore.

• 1:17:20Speaker 01

  Unfortunately, I can't tell it with the rest, so I'm just gonna guess this one.

• 1:17:23Speaker 00

  I'm sticking with it anyway. I'm sticking with it anyway.

• 1:17:25Speaker 03

  No!

• 1:17:27Speaker 01

  I did not see that one.

• 1:17:28Speaker 02

  I was looking specifically at the hot bar.

• 1:17:30Speaker 03

  Here you go.

• 1:17:31Speaker 00

  I lucked out. No one else got it. I got lucky again.

• 1:17:34Speaker 04

  Point to YouTube compression as a possible culprit. And sure enough, this can actually lead to the two hot bars looking different, but in the exact opposite way to what we see here. Since the hot bar at the bottom of the screen is darkened by the inventory overlay, changes there can sometimes get compressed away, which can lead to the bottom hot bar being delayed behind the bright, clear hot bar in the actual inventory GUI. But here, it's the exact opposite. The hot bar at the bottom of the screen somehow sees into the future.

• 1:18:04Speaker 03

  That's the main one for this one here. He just, they spliced it on top of it, probably. They probably spliced this on top, had something that hid it from his inventory, and then he clutches down at the bottom to the boat. So anyways, that one doesn't have anything to do with the jolt cracking. Here you go. Prime, we'll talk about the randomness at the end. This was a fun, you know, there's some fun stories about how you shuffle and introduce people to new randomness. So we'll come back to this in a second. Alright, so. Here's what's going on here for the hunger bar. Basically, this guy, Dark Spark, the one whose video we've been doing, shout out to them. They made this thing called the jolt cracker. And effectively, what happens is each of these things has a chance to be up or down, right?

• 1:18:49Speaker 03

  So up is the pluses, down is the minuses. They move up or down. And as you have different hunger amounts, it changes. And here's where it gets fun. They have basically this big modulo that they take this item over, right? And it gets a new seed. The seed is based on the particular time in the tick in the game, right? So they're setting the seed based on the tick inside the game, right? And then they're doing this big modulo cut here across using this number, okay? And what happens with that is you can work backwards from the pseudo-random of the numbers,

• 1:19:32Speaker 03

  where you can say, cool, what produces this pattern? Since it's based on the tick amount and this modulo number here, you can actually work backwards to figure out what tick a particular up and down categorization is going to produce this pattern. So since you can produce the tick, you can find that original, uh, stopwatch timer that we brought up at the beginning of the game. So basically, this functions as a rough clock that shows you the count of the game at various points along what's going, because it's pseudo-random based off of the tick.

• 1:20:17Speaker 00

  That's hilarious.

• 1:20:18Speaker 03

  And so what you can see as you're going, what should the video look like? It should be a monotonically, yeah, go ahead, Prime?

• 1:20:26Speaker 02

  No, I didn't say anything. I didn't say anything.

• 1:20:28Speaker 03

  Okay. It should be a monotonically increasing line. It's just the slope is just one for all of time, right? If it's truly uncut. But instead, what you see is like, oh, maybe they played the game for a while and then had to restart because they ran out of time that day and played again for a while, played again for a while, right? And so the jolt cracker can tell you from watching this. Um, and so then basically, right? They go through here, you put in the different spots in the video, where game tick could possibly produce this jolt. Um, so they made an awesome little website that you can go through and you can download the JSON from these and you can click through to the different spots in the video, so you can go prove it yourself. All of the code is open source.

• 1:21:11Speaker 03

  They built, I don't have it in here, but they built a small, little computer vision thing that basically can look, you know, very simple, but it just looks at this spot of the screen.

• 1:21:21Speaker 00

  Oh my god, dude.

• 1:21:22Speaker 03

  And can tell you what tick it needs to happen from.

• 1:21:24Speaker 00

  They really don't want this guy to get away with cutting Minecraft videos, huh?

• 1:21:27Speaker 03

  No!

• 1:21:28Speaker 01

  Dude, but the thing is, is that I think that it's, if most people cut Minecraft videos, it's not a big deal. You just can't do the boat save. And say it's uncut, right? Like, that's, I mean, that's.

• 1:21:39Speaker 03

  So one of, you can't be the best.

• 1:21:41Speaker 00

  There's two big. We gotta let Trash go.

• 1:21:42Speaker 03

  Alright, Trash. Alright, see you Trash.

• 1:21:44Speaker 02

  Oh, okay.

• 1:21:45Speaker 03

  Bye Trash.

• 1:21:46Speaker 02

  I got, I got four minutes.

• 1:21:48Speaker 03

  The thing they talked about though, Prime, besides just, you know, proving that people are lying, which isn't objective good, proving that people are lying, that's good, uh, is also there's a lot of people who splice Minecraft speedruns. And so this is gonna be a crazy tool for the Minecraft speedrun community, because it's gonna be like, good luck, bro. Good luck beating the jolt cracker. Like, um, which is gonna be crazy. So I think that's pretty much, uh, that's pretty much it for this. We can talk maybe a little bit about, um, you know, some, some stuff in terms of randomness.

• 1:22:26Speaker 00

  Trash, that last, Trash, that last.

• 1:22:27Speaker 02

  Oh, that hurts me.

• 1:22:28Speaker 03

  Yep. Yeah, Trash has a meeting. He just needs to leave. I, I, I should have put something in where it put legendary if Casey won, but, um, I didn't, I didn't think about that.

• 1:22:36Speaker 00

  Well, now there is a real legend. It's like, remember how he won that game?

• 1:22:40Speaker 03

  Yeah, that's true. Um, I do wanna say before we just talk to him, other final things. Shout out Dart Spark. Uh, we'll leave the link to the video for his original thing, the jolt cracker, some of the code and some of the other things there. Uh, the, the video is awesome. Great YouTube video. Amazing. Very well edited, very fun to watch. Uh, takes you on a really great journey. So just wanted to at least make sure we do that before the last second of the video, so.

• 1:23:03Speaker 02

  This was insane. That was like really, really interesting.

• 1:23:07Speaker 03

  Yeah.

• 1:23:09Speaker 03

  Casey, this is just a joke question that I wanted to have, at least before we go, we can just sign off because Trash has to leave, which is if you were designing the game, how would you make a tamper evident? And then I wanted to do the reveal of flipping it. If you were Dream, how would you have hidden the cuts better?

• 1:23:24Speaker 03

  Oh, wait, one more thing.

• 1:23:26Speaker 00

  Okay.

• 1:23:26Speaker 03

  Kate, Casey, I need to, I need to know which one of these you think is funnier. Can you read these on my screen right now or no?

• 1:23:32Speaker 00

  Okay, yeah. So John Carmack says that demo involved a non-trivial chance of embarrassing failure. I'm only going for the baseline model.

• 1:23:40Speaker 03

  This is for the, okay.

• 1:23:41Speaker 00

  Roadster. And you said that makes sense. I imagine that's why they don't call you John Rocket Mac. Okay.

• 1:23:48Speaker 03

  Yes, I gotta chuckle.

• 1:23:50Speaker 00

  Uh, here's what I would say is like, uh, John Carmack was the guy who did Armadillo Aerospace.

• 1:23:56Speaker 01

  He did do a rocket company for a while.

• 1:23:58Speaker 00

  I kinda feel like the better burn there would have been something like, you're only saying that because you're jealous that you, you know, or something like that.

• 1:24:05Speaker 03

  Well, I wasn't trying to burn him. I'm just making a dad joke.

• 1:24:07Speaker 01

  Alright, cuz.

• 1:24:08Speaker 00

  Well, that succeeded.

• 1:24:09Speaker 01

  Car Mac.

• 1:24:10Speaker 03

  Okay, good. Because John Carmack's awesome. That's the only reason I wanted to put it.

• 1:24:14Speaker 01

  We're fans of John Carmack.

• 1:24:16Speaker 00

  TGTV, breaking absolutely huge oversight from Elon Musk revealed in tonight's test reveal. It makes absolutely no sense to call the world's first flying car the Roadster. It should be called the Skyster. Tesla fans in shambles after this.

• 1:24:29Speaker 00

  I, I don't know, man.

• 1:24:31Speaker 01

  It's A.

• 1:24:32Speaker 03

  Casey, you can pick A, B, or C. You can just put neither. That's fine, but I have to get your opinion.

• 1:24:35Speaker 00

  See, we're going with C.

• 1:24:37Speaker 03

  Okay.

• 1:24:37Speaker 00

  I think we, dad jokes, I guess what I would say is the,

• 1:24:43Speaker 00

  I'd wanted more of a pun, bad pun feel than I'm getting from this. I realize you were trying, but I, I don't know. I want a better bad pun.

• 1:24:51Speaker 01

  These are like badass and cringe, TJ, so you just didn't do a good job. Sorry.

• 1:24:54Speaker 02

  That's literally the definition of a dad joke.

• 1:24:57Speaker 01

  Except for Rocket Mac.

• 1:24:58Speaker 00

  Maybe the A, Rocket Mac was pretty decent. Here's the thing, I'm gonna go with A. I'm gonna go with A. A has more of a pun to it that actually makes sense. So we're gonna go with A. It's A. The answer is A.

• 1:25:06Speaker 03

  Okay.

• 1:25:07Speaker 00

  The answer is A.

• 1:25:07Speaker 03

  Nice. Awesome.

• 1:25:09Speaker 00

  The answer is A.

• 1:25:09Speaker 03

  Okay. Thanks everyone for joining us on this episode of the Stand Up. If you liked this, you can make sure you watch it on Spotify or on the main channel, which maybe that's where you're watching right now, where we have every full episode unedited at, well, just a little editing. Um, you guys don't, do not check the hunger bar on this.

• 1:25:26Speaker 00

  Do not check the hunger bar when he says unedited.

• 1:25:31Speaker 03

  Oh, man. Bye, everybody. Thanks.

• 1:25:33Speaker 00

  Bye.