Raising Funds For World Domination (Day 317)

[clears throat] [snorts] All right, welcome to raising funds for world domination day number 317. So, for today's daily update, there are no updates on the anticipated appeal because it's Saturday. Uh, I'm pretty sure there's still that uh new block sponsored block type available in store. Uh, today I finally did some recording on the next long form video. This one will probably be a bit shorter. Uh, and I edited part of that. Um, so that'll be to come out soon. And other than that, we will continue today with our work on the terminal dock mining simulation game. So the current focus of the work on the game uh involves trying to get toward uh multiplayer support and the current focus is to add some unit tests for the chunk loading model. Um this actually I haven't had tests in um I haven't really been testing it before. And there's also some uh technically some bugs in it. Uh and then it also needs to be generalized so it works with multiple clients. [clears throat] So I kind of have it almost doing some meaningful tests. Um, I think there also may be some issues with like flow control where it's like send sending off maybe too many requests at once. [snorts and clears throat] Says, "Good evening. How's it going? Yeah, maybe the fact that this is a list is actually uh I kind of don't like the behavior when it starts up. It it looks like it almost looks like it lags a bit and I think that might be because of this >> [clears throat] [snorts] >> going well. Had a productive day. Spent about two hours with Claude and Chrome. MCP updating LinkedIn. [snorts] Oh, interesting. Using MCP to update LinkedIn. That's not a use case. case I would have even thought of Canada generation. [snorts] How's it going? [clears throat] [snorts] You know what? I might change this to not use a list. >> [snorts] >> to interview me. Went to my current LinkedIn profile, fixed a bunch of stuff. Who knows if it helps, but maybe Yeah, I I actually think um [clears throat] sometimes just having like well maybe a large language model could do this, but I think sometimes it helps just to have um like someone who's unfamiliar with your page or thing like look at what you're doing because I do see sometimes um like when you're trying to advertise yourself. This is like especially true in business. If you're trying to advertise your like portfolio or what you do or your focus or whatever, um people almost always overestimate how much they're they think other people are aware of what they're trying to communicate. Like if you have a website or something or a signup link or portfolio or something and you you want people to see it, it's like very common for somebody to have like a LinkedIn and they there's like just a link like an about, you know, my website link or my resume thing or some reference to something. um and or even like they post a YouTube video, they're like, "Hey, I'm looking for this or that thing. Please uh check out my portfolio." And then you're like, "Oh, wow. That's interesting." And you try to like find their portfolio and you can't find the can't find the link. You can't find their website or whatever. Or even like um for my stuff, I I have like Patreon. My Patreon link is on every single page. It's like an automatic stub. And if you go to my channel, it's it's there. But people still uh I'll get people will randomly comment like, "Hey, have you ever thought about starting a Patreon page?" I'm like, "I have one." And it's linked in literally every video. So sometimes just kind of like the getting the the context because to [snorts] you it's familiar but not to everyone else. Let's just make that one. >> [clears throat and cough] [clears throat] [snorts] >> watched a good YouTube video a while ago but differences different platform types apparently Tik Tok, Twitter, Instagram are good for finding content creators Patreon and Kofi are good for getting paid. [clears throat] Yeah, every platform has a specific focus and importance. Um, Twitter Twitter has changed dramatically. Um, like in the last I don't know six months or something. Um, I would say it almost seem it almost seems Twitter is almost uh exclusively around news like breaking news and stuff. So, and to the extent where like if you post something that's not about breaking news or like what's you know the current thing then uh like if it'sformational content then it it just won't get seen. Um Tik Tok also Tik Tok got acquired by um Oracle recently so they changed their algorithm a lot. >> [snorts] >> Instagram also u has been changing it it depending on the laws and the administration could change a lot uh and there's certain platforms where I think it's like basically baked into your algorithm that even if your content gets engagement they're like trying to curate a certain type of audience on the platform so even if your content is good it'll never actually get seen. >> [snorts] >> U but yeah there you can monetize on Twitter but uh I don't think anybody actually makes anything on Twitter. Um Pat Patreon is um I don't really use it for any I just use it as like a payment processor basically. They don't really support cross linking. Yeah. every platform wants. I think that's pretty much b an assume it's a good assumption for every platform. If you link to a different platform on any other platform, then you'll get downranked in that post. Yeah, it'll still get seen. But yeah, [clears throat] I need anytime I like anytime I advertise something or like try to monetize those that post will always do worse. The real Sheree. Good evening. Is that a tea? Uh, yes it is. Um, orange pico tea or uh Earl Grey, I think. [clears throat and cough] need to be stuck out for a few weeks. That's too much time to be out of tea. >> [snorts] >> We're on my shopping list now. Hopefully get some very soon. [snorts] Nice. What's your uh what's your favorite kind of tea? I'm a fan of black teas personally. I know some people really like fruity teas, but I I always find they're really sour. Can cannabis sativa that's not a type of tea. Maybe maybe remember http 418. I am a teapot. I [clears throat] forget the lore behind that. [snorts] I know it's one of the official HTTP codes. I forget why they added it. [clears throat] [snorts] Zach 4378 says, "I'm gay." Congratulations. That means you're happy. Utah teapot. I I don't don't I'm digging in my brain. I can't uh conjure up anything for that. Utah Tapod is one of the default 3D models. Very famous. Oh yeah, I think I remember. It sounds like one of those things that I learned and then forgot about. >> [snorts] >> Should [snorts] I have done this? Because I guess if I want to send one element at a time, I can just do it. [snorts] You know what? I'm going to revert what I revert all my changes that I just did. >> [snorts] >> I was thinking maybe I'd do that for uh better flow control, but I I can just send a list of one element and then if I need to go the other way, I may refactor it. Let me end up unrefactoring this again. [clears throat] This is what Google says. But 418 [snorts] the humorous intentionally designed client signed error code indicating that the server is a teapot and refuses to brew coffee. I seem to recall the um one of the first webcams ever. The first like live stream was of a teapot. Someone set that up to uh so that everyone could monitor the coffee pot in the other office. It's very easy as a software software dev to create work for yourself. [snorts] Should have done this is a question for the ages. Yeah, I thought about it. Um, I kind of would the the UI when it first loads it seems like it locks up. I don't know. [snorts] I'm not sure where where where exactly that is. It would uh and I had thought for a while that there's something there's some kind of flow control issue in the chunk loading model. >> [snorts] >> Okay, now that I'm trying to actually think about how to meaningfully test this. I realize it's going to be kind of harder to because it also it depends because there's basically two threads working in tandem. The testing thread and then the actual trunk loading thread. And depending on race conditions, they're going to have to they're going to have to synchronize at some point through at multiple points throughout the testing. [snorts] Was it a live stream to see if if an unw if an unwatched pot refuses to boil? No, I think it was um it was something it was it was either like it was something about checking if there was coffee in the pot, if there was any coffee left. Does modern Java use things like promises? Uh, I think the the latest version the latest versions of Java I've seen are um it almost seems like they're trying to make it into a functional language. I I actually don't keep up with modern Java at all. To be honest, it's getting windy out. Um, okay. So get a random set of regions. >> [snorts] >> So I guess this maybe should block >> [snorts] >> What's going on? Uh, we are currently working on some unit tests for a terminal based video game. It's a live thinking stream. Yep. Now let me think about thinking. So ideally I want to test like all the all the preconditions and callbacks and stuff. When I tell it to update the regions, it ideally should send it should call back with the signal for any chunks that have become pending and that's necessary so I can update the UI so to give an indication that they're they have question marks and then once they load then I do another call back. And because it's in a different thread, I would have to wait. But it also means if I check the progress in every loop and I end up with a bug then it will deadlock and I will never see progress. Oh boy, we got a super chat from Jame James Goodbridge I assume. Thinking is not allowed under your current subscription. Please upgrade to a plan that allows thought. Thank you for your cooperation. Oh, this is this is the new new model where you you will not own anything and be happy. You can't you don't even not even allowed to own thoughts. You got to subscribe. [snorts] Thank you very much. Last Friday, I asked my teacher if I should learn specific programming language or multiple languages. He said, "Nowadays, AI can write code for us. So, we shouldn't waste too much time on programming. We should learn more about how to use use AI to do. Do you agree, Robert? [clears throat] Um, ask your teacher if you um uh what what kind of teacher was it?" Uh, is this a a school teacher or or or is it a programming teacher? [clears throat] Doesn't sound like a great teacher. Uh, it's it I mean it's hard to say. Um, uh, so high school. Yeah, it's it's um [snorts] it's it's hard to say. The um I do think um like knowledge workers are going to get hit very hard, but probably not necessarily for reasons necessarily about like the AI metals replacing them. Um I don't know these there's so many things changing so fast all at once. Um so as it is today like right now I'm not using AI coding. Uh and people every stream people ask me about like hey why are you not using cloud code and stuff and I I know I do agree it's it probably is better. I just haven't had time to experiment with it. I've a few times I've gone on like um what is it? Copilot. So that there's a link in the description. You can go to the GitHub for this and you can go on copilot and you can ask it questions like um I've asked that a few times like you know optim u give me an example of an optimization to this function or one of the last things I asked it was um add a block give me the code to add the uh a new block type for metallic gold to this game. And that should be very very easy to do. Uh it also has like lots of examples in the code. Um and the the answer it gave me I kind of thought like oh this is a softball question but it completely was wrong. And uh like the answer it gave was like not even close. like we tried in inheriting from like a class that doesn't exist and stuff things that like that like that should be like AI is absolutely smart enough to solve that problem. Now if you like ask it and point that out but I asked it for to do this and it it didn't do it and people keep saying like ah this it's so amazing it's going to take over but like you know where where is it? Um, and I think people see I do see a lot of potential in it. Like it does seem it almost does seem like we're on the cusp of being able to like, you know, really get crazy on the exponential curve, but at the same time, uh, it's a common fault of like entrepreneurs to to because like if you're an entrepreneur, you see the potential. You see like, oh, it's going to get better. It's we're gonna make so much money and stuff. But it's also true that entrepreneurs will um constantly say, "Oh, we're gonna it's eventually going to get crazy and then um it's gonna we're going to replace all our employees and stuff." And then it just it never materializes. Um I could see it one day getting pretty crazy. Like it it's probably just going to be like every year it's going to get it's going to get better and better and better. Uh but as of today, like we still need coders and there's tons of people saying like it's gonna, you know, it's going to replace all all uh everyone in the next six months and stuff and then it just keeps not happening. Um, a big component of the layoffs recently is actually just because it's a cover for um, like if companies come out and say they're laying off programmers because they anticipate lower revenue, then that makes the stock price go down. But if they say we're laying people off because we're more efficient, we replace them with AI. The stock price goes up because if you just do the calculation like if I tell you I run a company revenue way down, that means earnings per share is down with the same um you know for the for the same unit economics. But if I say revenue is the same, but we now have half the employees, the profit goes way up by you just take subtract those salaries from the balance sheet. So that's another thing that's going on. Um I kind of think you'll probably have to do like everything like ideally if you want to go into this profession and master the craft, you'll have to learn the programming language and you'll have to learn AI. Um so I think it it is a good idea to to learn AI and look how to use it to solve problems. Um but you should also uh ideally know how to solve them like independently of the AI. like um I think the people who are going to do best with AI are the people who kind of don't trust it and always assume it's like it's lying to you and it's like showing you mistakes and stuff and it's your job to find the flaws in everything it's saying and that it's like a pathological liar all the time. Um yeah, I think that's that's the the right strategy to survive in this profession. Although it's I mean the economy is so crazy that we keep we keep um like the the government we keep having pandemics and wars and stuff like when COVID happened the entire economy shifted everyone went inside and then every like the demand for software skyrocketed and we didn't have enough engineers so then we had like record enrollment in that profession. And all the people who went in who got you know went to the profession who started studying at 20 like 21 or something those people graduated and are graduating now or like the last couple years and now we have like record low demand. So like it's the worst it's the worst of both worlds. They're they have the most competition in the weakest job market. right now. But then of course like because everyone's, you know, pessimistic and stuff, I bet enrollment is dropping like a rock right now. And so if you go into the computer science profession when everyone is like maximum pessimistic, that'll probably have a better outcome because when you graduate, you're going to be graduating with less peers and the economy might be recovered. But then also we have wars that are going on and the other variable is immigration. If immigration if the government decides like oh let's bring in a trillion people to compete with you then suddenly uh you know it's going to get really really difficult for you and you don't uh you have no control over that and you also it's it's impossible to predict. Oh and like government stimulus too. the government might just like print five trillion dollars into the economy and then uh pump up all these AI startups, but then also they might pull the plug on it and then they'll they'll all crash. So, it's um yeah, it's getting harder to reason about what's going to happen. >> [clears throat] >> kind of feels like it's in the category of flying cars promised in the past. [snorts] Yeah, it's there are some aspects of um AI where I think oh this this could get like like a runaway crazy thing but um at the same time I think a lot of the most profound >> [snorts] >> the most like profound takeover scenarios with AI are all going to be like capital limited. Like I think we could even get get to a point like let's say you had a general an actual general AI and it realized like wow I can do all this amazing stuff and figure out better than human solutions. The first thing it would probably do is it would ask for more resources. So it would say like can you build a data center here or can you like um can you like order all these robots and stuff and to the time to deploy that capital and to build the infrastructure that the AI wants is going to take time. So that that's going to kind of slow down even in a a general AI singularity scenario and which we're we're not at that yet. Like one of the other things that kept me away from actually trying out AI seriously is I would go and I would look at the plugins for for actually using AI and they all had like one-star reviews and people were like this is total garbage and I can't even copy and paste with this and they're like, you know, this is using the old version of the model. it's like the results are terrible and stuff and so I'm like I'm not I'm not going to waste my time with some AI plugin that you can't copy and paste. So given that we were at that given that given we were at that point like six eight months ago um I don't know it's probably not not going to take over I think uh actually you know I think she said the she or he suggested that to um the right strategy is to to learn how to use the AI but I Uh I think knowing how to write code and like knowing um about how code is structured and design patterns and stuff that allows you to specify the language to request the AI to do stuff. So, for example, um if I'm if I want to get an AI to write an OS for me, if you have no idea about writing an operating system, you probably don't know that a context switch or like preemptive context switching is a is a very important part of an operating system. And so, if you don't know that phrase or what that means, then you won't be able to ask the AI for to give you that and if you don't know of like the common problems and pitfalls that you can have um like for example it's you uh if you've ever written a preemptive context switch before you might know that forgetting to save some of the registers is a common bug and then if so if you ask the give me a give me the context switch code for a preemptive context switch for this CPU and then you it gets an an answer back and uh it doesn't save the registers. You can say like this is wrong. You're not saving the registers. But if you don't know anything about the result, then you won't be able to you won't be able to respond and correct the AI. [clears throat] Is ML AI the next frontend engineer? Uh, I think it'll it'll probably invade every role back end and front end. What do you think about business opportunities for AI cyber security? [snorts] There probably would be some uh if you could get really good at like if you could learn the patterns for like really good fuzz testing and um I guess like if you're probably already good at cyber security, you'd probably be very very good at asking an AI to help you do cyber security much more efficiently. Put it this way, AI might be powerful. Software has never been more complex. Expectations have never been higher. Yeah. The other thing I said uh I've been said repeatedly about my impressions of it so far are um um AI works. the output you get with AI is very good. Um if if the solution space to your problem is very large and any specific any random point in that solution space is acceptable that's the scenario where AI is good but the scenario where it's really it's like garbage is when you have a very large solution space and there's like exactly one potential solution that's acceptable. Um, often the AI is really bad at finding that. It'll maybe get close to it. Um like an example of of like a big solution space uh is I said like um write a write a pro write a computer program that'll wish you happy birthday and it looks fancy and there's like many interpretations like just is you know you could open happy birthday but maybe it's not centered and it's like the color is kind of weird but there's some like balloon icons or something or a cake or like any combination of those things is acceptable. But if you say if you ask an AI write a computer program uh that performs the absolute um fastest uh uh [clears throat] memory copy function for arbitrary data structures and it considers like overlapping regions stuff that's the kind of thing where uh the very often the AI will it'll give you something that's close but it's just like subtle it's subtly wrong but it's catastrophically wrong. So you have to figure out how to fix it yourself. [clears throat] Is that why gold is so valuable? Because trill trillion dollar trillion dollars can be printed in a snap. Yeah. I guess uh that's why uh probably commodities are going up so much and we got the war going on. There's also um I think there's also like some uh I don't know that the [clears throat] people the powers the bee who have all the money are doing some shady business with the market. There's some market manipulation with, you know, the billionaires and people who are in a pay grade above us. [snorts] Just prompting it wrong, bro. Just vive coded an entire OS. It's called Windows 11. Nice. You should sell it. Get rich. That's why Windows 11 sucks at working with display port monitors for USB 4 doc. Yep. It's all all your fault. The war is a great cover for the files not being released. Uh some of them are released. I I saw there was a thing that's like even more of them were people are talking about them today. There's a I mean there's so much completely insane news going on now it's like impossible to process. In Canada here, we also had um this absolutely historic decision by the government, by the court in BC to like to acknowledge uh to acknowledge title Aboriginal title to like the entire city of Vancouver. [snorts] And now all these like private land owners are like there's a a completely unanswered legal question about who has title to their property because tit like western style property law is exclusive ownership of property. It's like all these pe it's and the the Canadian media barely covered this at all. So for people in that situation, it's like number one [clears throat] saturation point. It's been constantly reached for a while. Yeah, it does seem like we're uh in a saturation point. People are just there's so much going on all at once. People uh people are can't cope. No one owns anything. What is just steward of a property for a period of time? Well, I mean even that is uh in question now. [snorts] Same thing with land ownership in Australia. Just all a name means nothing really changed here. Well, I mean the uh the the thing in Vancouver this just this just happened like a few days ago. So, there's a whole bunch of answer questions, but like this is probably going to decrease people's property values a lot. [snorts] Gor Danny says, "Hello there. How are you doing?" I'm pretty good. How are you? Okay. I haven't I haven't done any coding yet. I did some coding and then I perverted it. >> [snorts] >> I guess the test doesn't make a lot of sense unless I actually wait every every iteration. Yeah. I'm good. Pretty amazed by how how deep math can get. [snorts] Yep. It gets really deep. You should should learn the Reman hypothesis. Good job. We distracted Robert. Yep. She ball. Ah, says, "Yo, how's it going?" [snorts] Chat wasn't this alive since 1987. That was a while ago. >> [clears throat] [snorts] >> Fort train Fortnite. Those are different things. Can you please explain what you're doing? Um, I am working on some unit tests for a terminal based video game that I'm working on. So, I have uh I guess the tests run right now, I think, or not. H I broke it. Uh yeah. Anyway, um so I'm writing tests for um the chunk loading model of a terminal based video game. Maybe that's because probably because I didn't deploy it. What is the reman hypothesis? How do you feel about the fast fiat transform? Uh [snorts] the fast fio transform is um actually the it actually was on my to-do list at one time uh to make a video and the video it was going to be titled uh uh the fast fo transform the most the like the most important algorithm um and actually uh veritasium made a video and I think it has the same title and uh I can't remember I can't remember if I watched the video but um yeah Veritasium has a video on fast varia transform um and it is a uh it's a very important algorithm. [snorts] Yeah for transforms and lelass trans. I wish I had time to like mess around with that stuff more. There's quite a bit of deep um actually if you um if you Google search for how to multiply numbers with the FIA transform uh you might still find my blog post. Actually, I guess I could probably do that. [snorts] We should [snorts] all become CEOs. We could uh bro, why don't you two become a CEO like in that video we were talking about? Uh so the problem the problem is there's too many entrepreneurs trying to do that but because that's unproductive what ends up happening is a lot of a lot of people try to do that and then it it doesn't even work because then they they like make some stupid product and then it it doesn't they can't raise the rounds and then they try to get the contracts But like they don't have the connections. So for like for every one person trying to do that, there's probably 10 people who are trying to do that model and failing. And then when you when you try to do that and then fail, now you have no money and all and you like kind of burned your your bridges. [snorts] Can you suggest anyone interesting viruses like stuckset which is not that popular like I love you or want to cry [snorts] suggest any more interesting viruses uh I don't know are you looking for like documentaries or or just research projects um there's the Morris worm actually I have a video I have a video on the pinky command which references the finger command which was the target of the Morris worm back in the day. [clears throat] Ever since Veritas Veritasium did their math math has a whole video specifically the part about caner diagonalization argument I pretty much stopped watching their content. I I don't think I watched that video. That's a topic I I'm not uh not well verssed on. Can you please send your GitHub account? Uh there's a link in the description. Got the toilet command. It's free. I've got fire. Fire. Yeah. So many types of text you can do. I if I remember correctly, there's a I think there's another command like the toilet command. Uh, just Google search for like toilet command alternative and I think there's another one that does something similar. Oh, [snorts] yeah. There we go. It's I'm the very first result. >> [snorts] >> Yeah, this is the uh the way to just take really big numbers and and use for your transforms to multiply them. >> [clears throat] [snorts] >> Guess we figured out the fast for because we wanted to know if Russia was was nuclear nuclear testing weapons underground. He wanted to see if we could tell tell the difference between regular seismic data nuclear tests. [snorts] Yeah, I remembered that in the uh in the lore. Turns out you can break any complicated waveform data into into its most basic functions. Sine square saw and triangle. Yeah, I think any any there's a restrict there's like I think any periodic function or like any some qualified periodic function you can the crazy thing with FIA transforms is um um let me think so that I think there's um there's a fast vari transform and then there's a I think another variant of the fast fiery transform that's uh let me think I can't can't remember the implementation of the fast f transform now but like there's another version of the fast furry transform it's like the even faster fa transform that's like some prime prime based division of it and then it's the fact that uh you can do fier transform forms on like rings and then there's the um there's the number theoretic transform where you you can uh because with the FIA transform you have to do um floatingoint calculations so you can end up with imprecision but with the number theoretic transform you can avoid uh I think you you I I haven't gone through the math, but I'm pretty sure you can do it with pure integers, so you can completely avoid floatingoint rounding errors. So there's that. And then and then the FIA transform is a special case of the lelass transform. So you can generalize the whole thing even more. >> [snorts] >> Oh, we got a super chat. Thank you very much for your f first super on the live stream to fix [snorts] fix your quiz. Just for curiosity, they say stuckset code is now public. How easy is it for my for my friend to f up my PC using that code? Uh I'm not sure. I guess if it's public then hopefully one would presume that the attack vector has been patched. Um I can't I can't remember. I think uh Stockstead itself I think was designed to attack um programmable lodger controllers in the uh centrifuges of a Iranian nuclear facility. So uh it probably doesn't actually do too much to your machine. I think it it'll just like lie dormant. But uh just don't plug if you're if you have any uh uranium purification centrifuges. Don't plug them into your computer. Should I'll break them. Turn. Uh, I like the toilet command. It's It's a pretty good command. Sucks. Net wasn't designed to attack normal PCs. It's made as a sabotage for nuclear centrifuges. Yep. I need my centrifuges up and running to provide fuel for the reactor runs my BTC miners. Yeah. And if you're running all these like powerful AI models and stuff, you do you're basically need your own nuclear reactor now. Is there another database or can we can I just Google? How do I learn more about fast for circles? Uh I'm I'm not sure what that that part means. Is is there another database? Uh do you mean uh are you talking about the blog post? Um for me I just Googled um multiplying numbers using fast for that'll I I actually have two articles about um uh a couple a couple of math articles. So this this is the one about multiplying numbers to do for your transforms. Uh, and then if you want to just look at the overlap add, overlap save. Um, I I got I dug into this when I was doing like audio stuff and you can actually look into this project if you want the fast fast meme transform. So, it was a way to um transform uh it's a tool to take arbitrary audio and transform it into a Linux command. So, you can actually run this in the terminal and it'll play audio. It basically [snorts] just takes the waveform and uh gets the fiery components. And then I there's a bunch of uh I found you can actually there's a few tricks you can do to to like decrease uh the size. So we just kind of like clip some of the audio. Um, but yeah, if if you're super if you are a huge fan of FIA transforms, you can uh check out check out this. I think this didn't actually get much traction, but and then I yeah, this is where I talk I think this this is not my best quality writing here. I think that some of this I tried to like reason through why uh it was okay to um this is my reasoning on on being able to clip uh some of the information. [clears throat] Just use solar power. It's green and green is good. That's true. 20 weeks to store enough energy to run the ls lshw command. It could be faster than that. [snorts] I'm confused. I just wanted to know more. That's effing amazing. Thank you. You're welcome. LV.exe is interesting. I gota I gotta go off that. It's hurting my eyes. Croian Yeti is decent and scary if you want classic one that was considered malware. Oh, viruses. Ruin systems. Yeah, I don't I don't really want to play with viruses. They'll accidentally escape. [snorts] Um, all right. I put it put in my order for tonight. I'll probably leave things there. Uh, I don't think I I don't think I actually coded anything tonight. Um, let's see. I I guess I got to at least do some planning. Oh, wait. I had a compile error, I think. Oh, I fixed it. So, that's all I accomplished. Then I I broke it and then I fixed it. Yeah, I'm at the I'm at the stage of a working [snorts] on something where it's it's hard to get started. It's hard to visualize how to how to structure this, right? Because I' I have the testing thread and then the thread that's actually doing the work and to actually I'd like to test things deterministically. I wanted to kind of like fuzz it but also be deterministic. And because it's in a separate thread, it's going to have to synchron synchronize with the main thread. >> [snorts] >> Thank you. Good night. Have a good night. Robert, do you know the up you you power command? >> [snorts] >> I do not. Great stuff. Could you use a thread pool with a count of one to help make it more deterministic? Um well the the test thread itself. So basically I'm testing a component of the uh so I have this in inmemory chunks class and it runs its own thread and this is a unit test. So I for the purpose of the test I'm not I can't modify anything in this or in the thread because I'm testing the thread because it actually runs in another application another part of the application. But I guess really what I should do is just like um inject these region updates and then um so send this message to the other thread in memory trunks thread and then this this method will have to like synchronize. It'll have to just like sit there and wait. It'll have to wait until some condition is true. Uh but I guess then I have to if due to a bug, one of the the failure cases could be a deadlock, but I'd like to have it actually crash then. Real question is now yourself from just streaming. The real question is now yourself from just streaming coding for eight hours. Uh for fortunately I'm only doing one one hour per day of of the coding. There's this Instagram musician I ran into [snorts] the only code to write music. You power-l e this command you show your battery percentage terminal. H I can even output text into the toilet if you want to look cool. That's a That's a great use for the toilet command. All right, we'll uh we'll leave it on that for tonight and we'll be back tomorrow night for day 38. [snorts] All right, have a good night, guys.

My Plans For World Domination (Part 1): https://youtu.be/Rk7pOEF0GFQ My Plans For World Domination (Part 2): https://youtu.be/FvMSjbXhfXw The Terminal Block Mining Simulation Game: https://youtube.com/shorts/RQiNQfpacco Why Did I Write A Terminal Based Video Game? https://youtube.com/shorts/Txew9PDnKu8 Why Did I Choose Java For Terminal Block Game? https://youtube.com/shorts/P4JWs73T6nw My Shopify Store: https://shopify.robertelder.org Become A Channel Member: https://www.youtube.com/channel/UCOmCxjmeQrkB5GmCEssbvxg/join Terminal Block Mining Simulation Game: https://github.com/RobertElderSoftware/robert-elder-software-java-modules