Coding Blocks - Water Cooler GPT

Episode Date: April 16, 2023

We’re doing a water cooler talk today. Also, Allen can tell you how not to leak secrets, Michael knows how to work a spreadsheet, and Joe has been replaced by an AGI. The full show notes for this episode are available at https://www.codingblocks.net/episode208. Topics Resources we like

You're listening to Coding Blocks, episode 208! Wa-wa-wa-8! Wa-8! I think that's a big deal, right? Was that the melting voice? I think that was the melting voice. That was the melting voice. No, no. So I went back and I was trying to find the melting voice.

Like, what did the melting voice sound like? And I listened to the time code that was given in Slack. And I guess I just didn't hear it. So maybe that was like, maybe what I think of as like the melting, I always think of like, you know, like the witch from The Wizard of Oz. I'm melting!

You know, kind of thing. Well, that was the eight, eight, eight, just a second ago. I thought that was like me just like being cheery and screaming. Okay. So maybe that was, maybe that was, I'm melting and i didn't even realize it that's right so hot it's somebody left me out in the sun and i'm melting i'm melting so um while i'm before i completely melt though uh let me tell you this in case if you didn't know, you can subscribe to us on iTunes, uh,

stitchify or, you know, wherever you like to find your, uh, your podcast, man. Whew. And skibetser,

if you're hearing us, uh, let us know if that was the melting point you're talking about. Yeah, we need to know. Um, so on the spot, you find all that kind of stuff.

If you, if you want to, you can send your feedback questions, because I got to read this to comments and blocks. kind of stuff if you if you want to you can send your feedback questions anywhere it's because i gotta read this to comments and coding blocks.net and make sure you follow us on twitter at coding blocks and uh i think it's my turn to talk uh about coding blocks.net which has social links at the top of the page uh and i'm joe gbt i was gonna make a similar joke. Oh, that's so good.

Okay, well, I guess I won't now. Thanks. Thanks. And I'm Michael Outlaw. I'm Alan Underwood. Yeah. Yeah. Geez.

Yeah. Way to take away my thunder there. The phrase, it's important a lot, and people will think you're actually gpt right now oh right well it is important so it makes sense there you go hey so this particular episode apparently we're just talking about topics i found because we were all supposed to come with topics to the table and and nobody else did so you did such a good job we didn't want like we weren't you know like what kind of ego thing is this that you had to like call that out we were fine with it we were like you know what he's done such a good job but you know you wanted the credit so here you go yeah so if this is a bad one blame those guys is all i'm saying right well you know

now that you bring it up now that you bring i'm glad that you did share that these are your topics because should anybody like hate these topics oh man all right so we had no review sadly so i guess uh i guess we'll just go do we have any news or anything you did orlando code camp um anything else coming up or are we good did you see somebody uh go ahead we We did. Hold on. One of our Slack people. Awesome. Awesome. I don't know if it's Simon or Simone Cuomo. He's speaking or she is speaking at Vue.js in London and has some free tickets to give away. So, please, if you're in the London area and you would like to attend the Vue.js conference,

drop us a message on this particular episode. Go ahead. What if they tweeted using at coding blocks and also the hashtag for the conference? Oh, I like that. Yeah, yeah, yeah. So we'll get that up in our show notes if you don't know what the hashtag and all that for the conference will be. But do that. Tweet us at coding blocks with this hashtag that we'll provide.

And, you know, again, if you're in the London area, please do that. If you plan on attending that or you would like to attend that, do that. And then that way we can get somebody some free tickets. So thank you very much for doing that for the folks out there. Well, I'm surprised that like that was the choice of, uh, you know, handles to use because there was a, uh,

a hashtag that I thought we should definitely make trend, which was Alan heard it, right? So, yeah, that was going to take off. Wasn't it? Viral to the core. Oh, it sounds it sounds good i don't know why you're so

down on yourself is it because you picked the topics yeah it's been a long week man all right so um all right so that out of the way did you did we have any other news jay-z you started to say something uh oh i was just gonna bring up something kind of funny that's kind of kind of news okay um did y'all see the story about uh the person that leaked some information some um confidential documents like pentagon documents might have heard something about it uh in a discord server yep might have heard something for minecraft was it a minecraft one yeah it was like it's just a small community minecraft as it sounded funny there's some sort of funny.

There was some sort of dispute, some sort of argument. Someone's like, look, I got the proof. Boom. Yeah, whoops. I heard that they may have found who it was and have taken him into custody. Yeah, I saw the picture. They're 21 years old. It's crazy.

That's brutal. Yeah. Yeah. So what are you in for yeah look at all the crap yeah well i was playing minecraft right and somebody called me out and i had to had to throw it down there yeah not a good idea all right so getting into topic number one. So I got to give a little bit of background here only because I got roped into coaching baseball again, which is hilarious because the first time like I walked up with my son on skills day to go out there and hit the ball and do stuff and and the guy sitting at the table is like there's by the way there's like 10 or 15 dads standing around and dude looks straight at me is like we need a coach like what why why me like what the bald-headed dude like i look like a baseball coach or something that doesn't make

any sense so anyway i got roped into it last. This time I go to the skills challenge thing or the skills day. And, and my wife's like, look, do not volunteer to be a coach. And I'm like, all right, that's fine. I'm going to be quiet. So I'll stay there. And my son's telling everybody that, oh yeah, my dad coached last season. And, and so, and so that got the thing out there. My son leaves his bat at the thing.

And so the coordinator of the league calls me up and was like, hey, your son left his bat at the thing. You want to come by and get it? So I go over there and he's like, hey, we need a coach. I'm like, really, man? You didn't volunteer. I didn't volunteer. I got voluntold.

So, yeah, I did it it again but here's the thing so all that it was the funny backstory but truly the most miserable part about being a coach of a baseball team is making the lineups like who's going to play in what position each inning and if you've got more than nine people, who are you rotating into other positions? Who's sitting on the bench, all this, and there's all these rules around it, right? Like, um, the same boy can't sit, uh, more than one inning in a row and all this other kind of stuff. So there's like, there's all these rules and it's, it's embarrassing how much time I've

actually spent making these lineups before every game, right? Like when I was bad at it, it was two hours. Now that I'm good at it, it's at least 45 minutes, right? And in my engineering brain, I'm like, it's a rules engine. It's truly like we talked about this way back in the day when if you're going in for, I think it was when we were talking about interviewing, um, if somebody gives you a problem, write it out in like pseudo code. Right. And, and the whole time that I'm making these lineups, I'm thinking the same thing. I'm like, there are nine positions. Each inning has to have one. You have certain boys that are going to play these positions.

So I'm like slotting all this stuff out in pseudoudocode in my head right and i'm like man i should i should just write this and if i write it it'll be a super useful utility for me but in in in my way of it needs to scale to a billion users right comes way for it now i'm like well man we need some state right like this has to be saved because you need to go from game to game like you don't want the same lineup every game and all that kind of stuff and based on like sick or availability like that could change on the fly so we're going to need a kafka stream that's right you know like that's right take that in as it happens i like how you're thinking right so so so i'm sitting here and i'm like okay well well they obviously need to be able to put their players in right and so this thing's growing right

like at first all i need is the thing that has like six innings with slots for players in each inning but then it's like well wait wait wait wait what if we could put the players that had store them and then you could see how they like progress over time and you could even, Hey, you can even rate their skills. Like how good are they? Are they at fielding a ball on the infield? How good are they at feeling the ball on the outfield? Then you can have all these kinds of algorithms that would like place people in the baseball. Let's collect statistics about him. All right. Well, Johnny's at the bat and he's a, he's over three on a Sunday, but it is a rainy overcast day. So that improves his chances. He's seven out of eight on a rainy day. Right. So all this stuff happens and I bring this up for a couple of reasons. One, I could totally do the MVP of this thing, right?

And I could probably do it in a couple days and it would be reasonably good and and this is where my brain starts going crazy and this is what I'm wondering if you guys ever hit this in your project so Jay-Z I know does a lot of like little side projects all the time and I think he actually finishes them a small minority which is still impressive outlaw I don't know do you do you write any side projects anymore anything like that i mean everything that i write would just be like to help my own life be better i'm not trying to like put it out anywhere so so basically what this started as right for me was i need i need to be able to do this quicker without so much um mental anguish the problem is I got to thinking about I'm like man this is a big problem there are a lot of teams

in just this local area like I don't know maybe 50 of different teams that we could all play against and I know all these other coaches are going through the same problems. I'm like, man, if there's that many in my one little spot here in Georgia, what if I sold this thing for like, I don't know, a buck a month, two bucks a month. I don't know. I would probably pay $5 a month not to have to deal with it. I'm going to go and register roster, my roster.com. See, I even like that.

I even like the domain, domain man how'd you come up with that so fast so honestly okay go ahead go ahead go ahead what are you gonna say no i mean just like as you were describing it though i'm like man this sounds like something like it at most if i weren't gonna like just write a shell script for this or something like that maybe just break out excel and you know do it there well that's what i do yeah that's what i do i put it in excel and then i create new tabs duplicate tabs and i'm back to 45 minutes then it'd be formula driven no like i said well it's not a formula it's not that it's not that complex i i list out the innings and then the players and then then I start filling in positions. Right. And then,

and then I go back and I make sure that I don't have boys benched back to back in. It is a pain. I'm saying I would have spent the one time writing it and then that would be it. No, no, no,

no. I hate Excel for that reason. Like I don't want to deal with it that much, but, but also the reality is too, you want to track it from game to game, right? Like do I?

Yeah. Like you don't want to bench the same kid in the first inning, every game, right? Like that's demoralizing. At least he really sucks. He does, but no,

but like, I mean, you know, I mean, that's just the way the random numbers worked out. If that's the way it worked out, that was the way it worked out. that was the way it worked out. That's what it is.

Yeah, you get used to it, little Johnny. You're on a rainy day, so you're not as good. It's so funny, man. But, I mean, I have gone as far as benching my son in the first quarter just to show I'm not partial, right? Which is kind of stupid, I guess, because I've been dedicating all my time to doing this. At any rate. So, I have so many thoughts around this, and I've paralyzed myself, which I do on occasion when it comes to side projects every time. So, I could.

Go ahead. No, finish. You just. I could do this and make it very simple right and and maybe have a three game flow to where it will attract three games then i'm like you know what this would be a great project for me to finally sit down and do the youtube series that i think people would actually get a lot of valuable stuff out of, right? You've got a backend, you've got a front end,

you've got all this kind of stuff. I think that's like legit. I just want to know, like, have you already like evaluated the different options and decided on the authentication mechanism? And, you know, cause there there's gonna be a billion concurrent users i know man i know this is where it gets hard so i actually thought about hey what if i had

jett gpt like create me a form and it's really good with html yeah you see because that's another thing right like let's let's be realistic here So I'm somewhat joking about this outlaw, somewhat joking about this, but the reality is this is really how my brain works, right? Like I'm going to, I need to wire it all up and there's going to be 20 tears and 85 services or something.

I don't know. What if he just went to the chat? GPT gave it the rules and said, create, create a list for me that follows these rules. That might work. And that might really upset me a little bit if it was that easy. Oh,

but, but in all seriousness. So if I, if I decided to do this, I'd have to decide on a backend, right? Um, is it a SQL database?

Is it a no SQL database? What, what is it? Well, this is where you need to understand your, your data use case and like how you plan to query it and how often do you plan to write to it? Or maybe you don't plan to write to it except for that one time, initial time and everything else is reads. So it's going to be really understand and important to understand,

uh, you know, those, those access patterns. I like where you're going with this, right? Designing data intensive applications. I mean, this're going with this, right? You're welcome. Designing data-intensive applications. I mean, this is baseball. It's stats-heavy. But then, if you're going to do a front-end, right, what are you doing it in? You doing React? You doing Angular?

Does it need to be on a mobile device? I could go the route that jay-z started off years ago which was um what was it progressive web apps pwas oh yeah man i mean like there's all these options so i don't know man like i've seriously this is the one project that i've thought man this is actually sort of complex enough to, to make something. See, this is where I would,

I would just focus on getting the rules engine right first, just however you want to write it. You know, shell script, Pearl, Python, Kotlin, C sharp,

whatever. I don't care. Write it and get that right first because all the other stuff that you're talking about is irrelevant to that and that's the need that you have it is yeah it totally is you know what i'm not even joking i think this is a really good case so because kind of like the thing is it's hard to even kind of organize the rules so this is a case where justify attempted, can you generate an algorithm that lists kids and lists their position preferences

and assigns them to positions on the baseball team. And it gave me a little algorithm in Python. But then you can go in there and say, oh, and also the kid can't play the same position twice and the pitcher can't do more than three innings. And you can kind of like add on these things because it's stateful. So you can kind of keep building up and building

up. And once you're comfortable that it's got an understanding of it then you say not doing a kotlin oh that's really interesting it's wonderful because you don't you don't have to that's the whole thing that i like about it's like you don't have to be well organized like you can just kind of talk with this thing like in english until it gets it i may have to try that out like and just see where it takes me because the interesting thing is what's even a bigger pain is when you get everything set you spent 45 minutes if you're good at it you print the thing out you take it out there and then what you said was actually really funny the pitcher can only go three innings actually it's more complicated than that um without going too crazy on it there are pitch counts and the number of days a kid has to rest before he can play again. Right.

So if you go over 30 pitches, the kid has to rest a full day. So that means that he cannot play tomorrow night. Right. If he goes over 45 pitches, then he has to rest two days and it keeps going like that. So it might be that you had this kid in there for three innings, but he threw 30 pitches and you need them tomorrow night. So you got to change it all right then anyways. So that entire thing that you spent 45 minutes on creating, you got to now just go destroy it on the fly while you're standing there,

trying to get everything set back up. That's why you as the coach need to walk out there, you pull up your pants, get a big wad of chew in your mouth, go up there and start kicking some dust on the umpire's shoes. Just stupid shibber shibber. And you point to the pitcher and you're like, go to the dugout. I've seen a thing or two.

I know how this works. Outlaw coach last year. So, don't know man it seems like one of those things that is actually complex enough that would be an interesting series that you could do to show a lot of different pieces i think it's complex enough but small enough to do it as long as i didn't try and scale it to a billion users but what about anything any of the generators that are already out there? I have not found generators that take all the rules into consideration. That's the problem. So there are things out there to where you can like plug players into

different positions and then even print it out. Like they'll even have a nice little baseball diamond and all that kind of stuff, but nothing that gives you like good inning to inning game to game type flows. And I've looked cause I was going to buy them. I was like, man, I don't want to do this anymore.

I just want to push a button. I don't know. Now you got me all excited about just getting a big wad of shoe and walking out there, pulling my pants up and like kicking somebody's like dust all over. That's what I want to do now. So, so I guess the,

the other part of this though was like anytime that you guys ever create something do you ever think about like maybe i should sell this on on like some sort of subscription basis or never or hey you buy the software or any of that sometimes sometimes but i usually try not to think about businessy type stuff because i just think like so like i don't want to maintain it i don't want any projects that are gonna last like months and months and months so a lot of times like when i'll have an idea for a business or something i think like do i want to do this 40 hours a week for next uh 10 years and the answer is almost always no so you so i actually you were probably very

qualified to talk about this because you actually did one and sold it like did it become like a chore that you just didn't even want to deal with at some point yeah absolutely yeah it was pretty much abandoned by the time i sold it worked out really nicely for me yeah i was gonna say so even even knowing that like how much time do you think you put into that particular site at the time oh i i don't even know in the beginning like probably more than 100 hours okay and then maintaining it over time an additional yeah i definitely built in you know new features and stuff so you could probably guess that like another 50 60 hours something like that over time maybe up to another 100 okay so 200 hours total and then you were able to sell the thing

and you actually ran ads on it do you make any money with that at the time like i think it was like 100 to 200 bucks a month something like that okay so you were all right so how long were you making that like a year oh a couple years like three four years okay so for so for three years you're making 100 to 200 a month yeah call a thousand back in the 80s too so like he was rolling in it yeah you know this is actually what i had like a thriller jacket you know like i might went out and got a michael jackson thriller jacket with it and the pants to match you know actually it was driving a delorean it was right around the time we started this podcast i think is about when you had sold that thing somewhere around there. So, so that was about nine years ago. So you made between three and $4,000 off of it in ad revenue. And then we don't need to talk about how much you sold it for, but do you think

in the end it was worth the, the 200 hours you put into it? Oh yeah. I mean, I learned a lot doing it and, uh, it was cool to, you know, to you know get money committing and like you know it's just inspiring and fun and whatever it was not cool doing taxes you know just kind of complicated things for like a thousand bucks in a year it's almost like yeah um but it was fantastic to get to sell it when i didn't get out of it and uh that was nice i didn't feel bad um it's kind of like what it's one thing when you're excited about a project and your passion and you're going for and people are writing in and asking questions or having corrections or just want to know stuff about it and you're excited about it. And it's another when you're four years in and people are asking questions or writing stuff. I would get sometimes people would write and say, hey, I think this calculation is wrong.

And then I would spend an hour trying to figure out if they were right or wrong and whatever. And I had a bunch of tests. I got a ton of test data from other tools and brought it in so i had actually this is the only project i probably ever had 100 test coverage on and i could do it because there were like new third parties you know there's like there was no database it was just code and uh so you know if someone said something was wrong i would like take their example and try to prove it or not but uh three four years in in, it was like, oh, geez. I think I set up a rule in my email.

I was just like, sorry, I can't look at these anymore. I just don't care. And it's not to be rude or anything, but it's like, you know, I just didn't want to spend an hour or something just to find out that it's someone else who, you know, was a little bit confused about the formula or whatever. Right. Right.

So that's interesting. I mean, the approach you took was not a subscription basis. It was basically, Hey, here's this free utility. And, and the best way I can probably put it to anybody out there, the tool that he had was sort of similar to, um, the epoch converter. If you, if you Google epoch converter, you can plug in a number and it'll tell you what day of the year it was, right? And what time and all that kind of stuff. The tool that Jay-Z did was similar to that. And I'm assuming, actually, let me go there.

Epoch converter. The site's not up anymore. Yeah. Oh, yours isn't? Mine isn't. Yeah, it's gone. Oh, wow. So Epoch converter is actually nicer.

It doesn't look like they have any ads at all yeah that's a pretty great site i use that x64 encode and decode all the time um that's one of my favorites well i could buy the domain back for 1500 yeah that's crazy so so the interesting thing about this is though you got you ran ad revenue you didn't provide a subscription it was just a utility that people could use and you still ended up making a decent amount of not a ton of money i think it is still up though am i looking at the wrong one hold on a second color mine.com it was dot org wasn't it yeah it's dot org according to the github yeah it's not org dude oh yeah it's still up awesome i don't even remember okay so we don't we don't have to uh you don't have to get around it but yeah so it's a it's a

a color converter right and so that's interesting i mean you did it but you haven't really done one since that that was any type of thing that you're trying to make money off of right because all this coding box is taking up so much time it does take a time by the way it did used to be colormind.com and you can actually see um the about page still says yay email me at admin at colormind.com oh that's interesting yep you're funny but uh yeah it was uh it was related to some stuff i was doing kind of tangentially for work i wanted to do like a color kind of search engine so you could like pick a color and kind of search. And so I'd like learn all this stuff about color. And like there's a – it's really – I thought it was really interesting.

There's a lot of different stuff about how – you kind of think naively like RGB to RGB values. There's a red, green, and blue value. If the numbers are similar like 255, 255, 255,5 255 and 255 255 254 probably pretty close right uh and it is kind of most most of the time but there's cases where it's really far off or like numbers that are really close look totally different to your eye and that's because rgb is designed for pixels in your monitor right it's not that's not how your eye sees colors that's just how we happen to map it so it's surprisingly off in some ways and so i got all into figuring out like well how do our eyes look and you know what are these other color spaces and there's a bunch of them cmyk which

is great for color like printing and it's because it's additive you actually add the cyan magenta and the yellow and the or whatever the k was i forget, together in order to make those colors. There's just a bunch of them, and some are designed to kind of map closer to your perception. So all you have to do, really, is just convert one color space to another, and then you can easily do that distance formula and see if two are similar.

But then there's all sorts of parameters, and this one lets you specify the white balance number or whatever, or the color of the light in the room. And it gets kind of fancy. So I had some fun with that. But, but after you did that, you've never really done anything else that you were trying to,

to publish, to make money on, right? No. And I found what happened to, so part of it was I stopped working at the company and I didn't have a need for a color search. And also I found that people had built, and when when i started they didn't exist but um by the

time i was done with it people had built color search engines that were actually pretty good and they were doing things in a smarter kind of just different way um where they were um like the part that i was kind of missing is like uh i wasn't really rounding the images very well and so i would like literally look say you know you the images very well and so i would like literally look say you know you pick a color orange and say i want things that look similar to this and so i would do that uh by taking all the colors in an image and then seeing you know how close you were to the closest colors and i say okay these are the top 100 items that are closest to your color but it didn't really do a good job of like taking into account like uh the amount of color in the image and stuff and like if you had image just like a big gradient

like what you really need to do is kind of like round that image down to like a smaller palette of colors and figure out the percentage there and there's just more work to be done and i was like uh i'm kind of done with this project i don't really want to attack it from that angle and especially when there's other competitors that are now doing this and doing it better. So, so when you finished it though, like any other ideas that you ever had afterwards?

Like, um, I mean, I know you've done some that were associated with the show, like, uh, the, the podcast player thing and all that.

Oh yeah. Like, was there any reason why you never went that route where it was like hey let's try and make some money off of it or any of that yeah i still think about qit all the time and i want it like i was just trying to like figure out some stuff about um playwright tool or pino apache pino and i was like man i wish i could search for all the podcasts where people were interviewed or talking about like pino or playwright because those are things i need to learn now and there's just no way to do it uh The thing is, it was kind of a lot of work

to maintain and curate a list of podcasts. So, a lot of podcasts, they only have seven episodes and they disappear. So, there's new ones starting up every year. So, there's all this content but you really have to spend a lot of time curating it. In the meantime, you're paying Elasticsearch, you're paying

for a database, you're paying for hosting. So, you're wondering it's like, you know, what's that? How much time am i really going to put into this for the next 10 years yeah it makes sense i mean it's hard that's that's honestly one of the reasons why i have a hard time starting side projects because i'll get all gung-ho on it for a week and then it's like life takes over and i don't have time and then by the time i come back around to it i can't even remember what i did. Yeah. It's hard. I love game development,

but for me to do even a kind of like a simple version of a game that I would like really like to do and make, I'm probably looking at like at least a year of like five to 10 hour weeks, you know, to have something that I would feel comfortable publishing. Yeah. I don't want to commit to that. There's other stuff I want to do. Yeah. And five to ten hour weeks in a year you're talking 250 plus hours like that

yeah that's not a small amount of work there is a apino podcast so i'm trying to remember what it was called yeah they're just uh one just started um they're on like episode two now they're deep it had well it's tim berglund uh okay from star tree yeah yep uh it's the one i'm thinking of i don't remember the name though real-time analytics there should be a link in the show there you get it there you go yep and tim berglund's the guy he used to do the streaming analytics podcast from kafka uh so i listened to all those episodes that he did, and so now he moved over to the Pino people. That's excellent. He actually did a really good... Here's a free tip for you. The first episode of Real-Time Analytics is the only one I've listened to, but he does

a really good job of explaining what real-time analytics are, because obviously real-time and computers doesn't really exist. So it's just kind of talking about there's two kind of real-times. And I'd never heard this explained, and that exist so it's just kind of talking about um there's two kind of real times and i'd never heard this explained that's why i wanted to kind of bring it up is like when people say real-time analytics there's kind of two things that people might have in mind one is i have the data as soon as it happened so i have the analytics about what just happened you know half a second ago or three seconds ago or whatever your definition of real time or near real time is but then there's also a different definition of real-time analytics which means that it's fast enough that i can query and see my results within a couple seconds so it's

less about how stale the data is and more about how fast i can query it and i that was a surprise to me because i had always thought of real-time as being low latency in terms of the data that you're looking at, how fresh it is. Yeah, that's the way I'd always thought of it too. It's like however fast you're able to get the data in, as soon as the data's in, it's available to do whatever the thing is you're trying to do.

Display it, query it, whatever. Going back to your thing, though, Alan, seriously, from my point of view, I never have thought, oh, I should make this available for sale or whatever. So in your rules engine example right if i were in the in your shoes in that situation what would end up happening is i would focus on that rules engine i would get that and then be like okay and i would i would yeah i would be

like okay i i got what i needed and i'm gonna move on and it would never it would i see people that put out apps on the app store for example or whatnot and i'm like yeah i guess yeah i guess you could do that it never occurred to me it's a lot of extra work or it's a lot of work to take it from that script that does what you need to something polished that you can that you can actually put out there whether it be for free or for sale or what but this is a lot of extra work and that's i guess the thing that i've noticed and we we see it as developers too right like i mean man the three of us have worked on so many data projects out over the years. There's always like this disjointed cash song.

I've dated everything, man. It's pretty close. But like if you just try to move data, it's like everything's so disjointed. You got to do sort of the same thing over and over in slightly different ways. And it's always sort of bespoke code. That's what I've noticed with the baseball thing that like every I've been in two different leagues and they use like four and five apps to get stuff done. And it's like, man, if somebody would just create a thing because they all need the same information. Right.

And they're all sort of providing one slice of a feature and it's like why is this so disjointed and so broken up like if and that's where my brain comes in is i'm like i i know how to glue this stuff together i know how i could make a good one but but that's not a small amount of work on what seems like a a decently small problem i just know behind the scenes from all the stuff that we've worked on over the years that that is a ton of work to make all that stuff work together. And keep it working for three years down the road. I love NetFly and Vercel and there's a couple like

Firestore. Firebase is a great example of things where you can do stuff for free. So even like database hosting, you can get some limited database hosting for free. But then you also run the risk of things like a lot of people do stuff on that free tier for Roku. Remember that? And then a couple of years later, they pull the plug. And so there's people who haven't done side projects in years, have been running for free, and they just don't want to migrate it. They don't want to get back into Headspace because they moved on to other things. So yeah, it's tough yeah i think uh there was somebody it might have been from

amazon that said that um software engineering is more about it's the maintenance and the keeping things running that is more difficult than anything else right adding features is cool and all but keeping stuff alive is way harder which you know there's tools out there that help with that nowadays like kubernetes and whatnot but that's still not a free lunch either so yeah you know all right how many projects i've started where like just on the couch put on a movie i don't really want to watch and just like download some new framework or some new library like see if i can make something cool happen and never touch it again yeah totally i've did it i remember meteor js when i was on vacation years and years ago. I was like, oh, this

thing's the best thing since sliced bread, right? And then, yeah, I don't know. I haven't touched it in eight years. Is it gone? It must be. I don't know, but I mean, when's the last time you talked about it? And that's why you killed it. You're the responsible person

for it. My bad. They're still around. Are they? I'm sorry. Hey, see? It wasn't me. All right, so the next topic, and this one's pretty interesting because this is almost like a PSA for all the developers out there using these tools like ChatGPT and Copilot and all that stuff. So there's an article on Mashable,

and I'm sure there's a million other articles on it now. Samsung employees were using chat GPT and had a few sensitive data leaks, a few that they've let us know about. Right. There's probably more that we don't know about. But the gist of it is this like developers were sending things to chat chat gpt to do things like hey um make this more efficient right optimize this code or whatever well when you do that you're sending that data out to the cloud to an ai thing that's going to ingest that and use it right so that's one the record the audio recording that was sent yeah they sent somebody decided they wanted to take an audio recording and turn it into a presentation so they set a sensitive audio recording up there which of course chat gpt is going to take it i would just love to meet the person that thought

this is a good idea. Oh, I mean, this is totally something I would do. Not real work stuff, but personal stuff. Oh my gosh, absolutely. My wife always complains about YouTube videos. She doesn't want to watch a video. She would rather just read it. And so something I would do

for her is you can use to paste a video link and i'll generate a transcript she much prefers that for some crazy reason i don't understand well some youtube videos already have the closed captioning on it yeah but you'd have to watch it she wants to read it yeah she wants to read it um and yeah as fast as she can read and sometimes they have transcripts but um chat gTs or OpenAI's Whisper, what do you call it? Their Whisper API is really great. Does your wife understand what YouTube is?

Yeah. She wants to read it? Yeah. And I totally get it too. Sometimes there's a video that's like an hour and a half long and she wants to like get. Well, you don't watch the hour and a half long ones. Yeah, you have to scrub through, you know, skim through whatever.im through whatever but you can you know yeah you can blast through f through there pretty quickly yeah totally yeah so so i think two of the ones that they mentioned here were

sensitive source code or or proprietary source code that was pushed out there to chat gpt so that's now in in an open space that shouldn't't be the one that was optimized some code and then the audio recording. So, you know, again, this is more of a PSA. If you're using tools like chat GPT or copilot or any of these other ones out there, whisper a high pay attention to what you're pushing out there, right? If your company finds out that you're copying source code out to one of these engines to see if you can make it faster, they're probably going to be pretty ticked off if one of their proprietary algorithms was let out into the wild, right? Yeah, totally. And I think it's a good thing to remind people, especially larger companies. A lot of them use tools for digital loss prevention that do things like go out

and they scan their network traffic. They scan their basically everything that's leaving their network or being uploaded, being emailed for basically artifacts that could be recognizable. So things like source code, for example, would be something that would make sense.

Samsung is watching their network traffic to see if any of the source code leaks. And so I assume that's how they caught it. Maybe they caught one thing or caught wind of one thing and they went back and looked at their logs or something and looked for anything else in there or figured out the people that were involved, which is kind of scary. And these people, I presume they weren't trying to do anything malicious. They were just trying to get their work done. You can imagine someone taking an hour-long video and being like, hey, step one, transcribe.

Step two, summarize. And that's pretty appealing. And they may not realize that they're literally uploading all that stuff to the cloud. It's kind of scary, too, when you think about, like, we've talked about Copilot before. Who knows how many plugins you have installed for VS Code? I don't even remember you know you can obviously go look but i've installed a ton of uh extensions over the years and vs code how do i know that one of them hasn't gone malicious and the latest update uh is now sending whatever's in the file context uh out to the internet to some something doing

something bad with it i don't even know i mean that's why I'm, let's take this to a different platform, like Chrome extensions or Firefox extensions. I am super cautious about what I install and if I install, what's even enabled versus I might keep it but not enable it until I need it. I'm the same way. I absolutely think about what Joe just said a second ago with the, what are they doing on an update? Because when you first get it, it's probably fine, right? Like you look at it, you check it out. Hey, everything's good. But if you're not, if you're not paying attention to every time you update one of those things,

you don't know what gets slipped in there. Right. And, and with the way that we've all sort of been programmed to update everything, every time something comes out, I mean, it's almost a, people don't pay attention anymore. Right. Like that's, that's kind of what it's getting to. So yeah, it's, it's dangerous. Like be careful with these tools i mean co-pilot i think you've worked with it the most joe um i think it's got things in there to where it can say like hey don't scan this um this code or whatever for for information right like there's settings in there for that kind of stuff yep and it's up to them to enforce those settings and deal them correctly you know and hope that it's right yeah yeah by the way one of my tips i'll go ahead and bust it early uh have

you seen it um aws just announced a competitor to um to uh what's called to co-pilot i haven't tried it yet um so i probably shouldn't give it as tip but code whisperer is the name of it and uh what's kind of interesting about it is that it looks like copilot except that it's only a few languages so it's really trimmed down on what they what they use to train it but here's what i think is cool well two things one's free uh at least for now second it seems to have a lot of really good tips specific to aws so they have like i don't know what they did to make it do this, but it looks like it's got a lot of information about best practices specifically for AWS.

So if you're working with AWS heavily, it might be worth checking out that one in addition to or instead of Copilot. So it's kind of interesting. When bots compete, can you imagine like Copilot saysilot says oh you should do this and

whisper says no no you should do this and come on it's like no you shut up it should be like this and then you know before you know you know you can't even park in the the the red zone the white zone white zone is for immediate loading and unloading i think and then the red no the red zone is for immediate i don't remember yeah well you i'm sure you guys have seen all the stuff about like all the competitors of chat gpt at this point right like google put out bard ai and and i'd heard like they rushed it out the door like it wasn't nearly as polished and yeah yeah well so here's something funny so i know i said i was gonna stop talking about this but i got one more for you uh so chat gpt v3 i forget i heard some stats around like how much it costs just to train it um so like kind of like i don't fully understand it but the part of the

the deal with the gpt stuff is like it's super expensive to train it takes a lot of processing power just to bring in all the information that makes this large language model work and so every it was like hundreds of millions of dollars or something to change to to train gp3 and 4 um i'll try to find the exact numbers but um alpaca which is a project from mit is similar large language model but they trained it in a fraction of the time so it cost significantly less and i'm going to try and find those numbers will probably take me a while um however the way they trained it is they used gpt to train it oh so they so they cheated yeah they cheated so there's a little bit of controversy there um but i mean if it works it works right but you can imagine uh you know if you're barred or you're one of these other tools like you can get a leg up just by using the existing tools to train the new ones yeah but then you're gonna

have this situation where like the bots are training the bot like yeah feedback loop right yeah yeah exactly like you're gonna you're gonna start it's like if you created – if you automatically – how do I say this? So you want to generate test data, so you create the test data to model your real-world data. But then the next instance is going to build off of that test data to create, quote, mock, something that looks like real data. And you keep going back and forth forth and before you know it, you just got like this smooth, better data. That's, you know, I don't know. I, I mean, you actually raised this question back a few episodes back when you asked

like, you know, are we going to, are we going to live in a time where it's like, okay, effort, you know, 20, 2022 is like, okay, from now on, we can't trust any data to train any of these bots because we don't know, like, is that data generated by a bot? Or how do we know what it came from? So instead, you can only train the bots off of data prior to 2022. Yeah, and the feedback loop is scary well i mean the interesting thing about what you started with there though is it took 100 million dollars or whatever it did to train chat gpt and mit did it for a lot less by piggybacking but did they really like if they had gotten to

the same end point without using chat gpt would they have had to spend 100 million dollars to do it? Would it have been possible? I guess this is a question like did did they find anything new or were they just able to get where they were because they were able to leverage it? Right. Like, I mean, that's always been the argument with like things like cryptocurrency or machine learning or any of that kind of stuff. It uses a ton of electricity, right? Like a lot of it. So all these technologies that are out there aren't cheap and they put a real hit on the energy grid and on heating and cooling type stuff, right?

Like it heats up the areas where they are. I mean, I, it's been years since I've looked into any of this kind of stuff, but like a lot of the data centers were put in, in really cold climate areas. And I'm sure they're having an impact on those areas, just like,

you know, global warming or anything would from, from driving cars and stuff. Those areas aren't so cold anymore. Yeah, exactly. Yeah. So areas aren't so cold anymore. Yeah, exactly. Yeah. So I don't know.

It's, it's pretty interesting. Like if you were to take it even deeper than that and be like, Hey, what, what is the impact on the actual world from, from running this kind of stuff? 24 hours a day.

I bet there is. Yeah. Oh, by the way, I got those numbers. If you want them, we want them. It's not, I was way off on how much cost uh so they the estimate it's not officially public but people um kind of put

together based on how long they said to train it uh and then of course uh there we go gpt3 is estimated to have cost 4.6 million dollars just in like from the time they hit like train in the train button uh till it finished uh gpt4 took 34 days and i'm sorry no dang it this is gpt3 again i have a bad link i don't know how much gpt4 costs but uh it looks like it's about uh 100 trillion parameters whatever that means. It took longer to train GPT-4. It's a lot bigger. Alpaca, the MIT version of it,

$600. That's pretty significant. $4 million is not as much as what I thought it was going to be. Yeah, I didn't either. I thought it was way more. But if you're just talking about Yep. I mean, 4 million is not as much as what I thought it was going to be like. Yeah, I didn't either. I thought it was way more, but I mean, if you're just talking about compute time and it crunching data, maybe that's,

I mean, that's, that's a decent amount of processing. Yeah. Well, I imagine some of that cost is just for the storage of that data to train it on. Right.

Right. Or is it the articles I was looking at is just talking about compute. Yeah. Okay. Yeah. Okay. Yeah. So pretty interesting. But, but the,

the original thing of that is be careful what you're doing with some of these things, right? Like, yeah, you definitely don't want to be putting either your sensitive data or your company's sensitive data out there for these things to, to go through.

Right. And that's why I was like an extra bit of safety net, as a bit of a precaution, what I did was when I created my account, I just signed in as Alan Underwood. Yeah, that's a good way. There you go. I got nothing to hide. I got all my roster lineups out there.

It's fine. All right. Well, hey, let me jump into this before joe tries to sorry so uh if you haven't you should be sorry if you haven't already left us a review we would greatly appreciate it um we didn't get any for this episode so you know oh sad i'm melting um so So, yeah, we would greatly appreciate it if you would leave us a review. You can find some helpful links at chatgpt.com. No, www.codingbox.net. And, yeah, with that, we head into my favorite portion of this show, SurveySense.

Okay, so this is episode 208. So according to Tateco's trademark rules of engagement. What are those, by the way? We got a question about that. Thank you, James Scheider. Oh, really? I'm on Twitter. This goes back. this goes back this goes

back uh so i could be saying his handle wrong but it was that that's how at least i think it was pronounced supposed to be pronounced and he hasn't ever corrected me in all this time but that was his handle on slack and he uh he or she i don't know i'm i'm i don. I could be, I don't remember. But he or she hit me up on Slack because there was a period of time where we would, before we did the quote to Tuck Co. Trademark Rules of Engagement, where I'd be like, oh, who went first last time? Who went first this time? And so this person wrote in on Slack and was like, hey,

why don't you just go by even and odd numbers of the episode number so if it's an odd number episode well alan's name is a that's the first letter of the alphabet so he's an a so he's so he's the odd the one he's the odds and j is an even number in the alphabet so joe would be the even numbered episodes so all the even numbers are joe and all the odds are alan and that's what that's what to tucko's trademark rules of engagement are and it's been like a couple years running now i think right yeah but you know i i'm i am perfectly willing to you know give credit where credit's due it was it was a great idea i didn't come up with it and i'm willing to say like hey and he's been immortalized yeah yes or she he she don't know yeah um so

yeah to tuck those trademark rules of engagement jay-z you are first and um i think you got a winning streak if i remember right i am on a one-game winning streak, sir. I lost so bad last week. Oh, yeah. Yes. You got a winning streak to start. That's right. Okay. So let's start with, let me see here.

Okay. You ready for this, Jay-Z? Here you go. Let me find it here. find it here and get my answers pulled up so i have it at the ready name a liquid in your kitchen first of all i should say like we've surveyed a hundred participants name a liquid in your kitchen that you hope no one ever accidentally drinks. Oh, geez. I cannot stress enough kitchen. Oh, geez. There's two answers.

I don't know if people keep one of them in their kitchen. So I'm going to go with. Oh, this is so painful. Okay. I'm going to go with soap soap okay okay that was the only other one i had i want to say bleach because i can't think anything yeah that's why i didn't know if people kept it in there dang it they probably do they probably do maybe i don't know i don't okay um yeah i don know. That's kind of disturbing. Be careful if you go to Alan's house. It's the takeaway.

So in the kitchen? I mean, it's everywhere. Okay, so we're going to go bottom up. Number five answer on the board. Baking grease for two points. What? That might actually, yeah. Alan's like, I don't know.

That might actually be good. That might be the worst thing ever. This one I don't get. Number four answer on the board, soy sauce for four points. Will you pour it in a different bottle? I mean, I know you're not supposed to, quote, drink it, but you do consume it, right? Like, that one's okay. Cooking oil for number three, 16 points.

Vinegar. Number two answer for 30 points. Uh-oh. Soap is going to hurt. Number one answer on the board. Mr. Joe's that comes out with a commanding lead. 46 points.

Oh, that's 46. That's what I was going to say. When you took that, I was like, well, bleach is my only other option. I got nothing.

It's a tuck. Oh man. That's rough. Okay. So this. It's a tucko, man. Tucko. Man, that's rough. Okay, so this is your chance to redeem yourself, Alan. Ooh, that's rough. All right. Okay. It'd be some obscure question where there's like 20 answers and they all are two points.

That's what's happening here. Name something that might come out of your nose when you laugh. Really? Oh, you didn't like that question uh now i gotta think about it snot okay oh that's a good answer uh milk milk okay that's pretty good that's pretty good number five answer on the board breath for three points i mean your breath out of your nose i mean it happens but why would you think that yeah that's weird that's a weird angle number four answer on the board soda for four points good answer dr pepper specifically yeah that'll burn number three answer on the board water for five points okay okay number two answer on the board milk milk for 32 points. Why is that so many? Come on.

Wait, what? What's this controversy? Why would that be the number one answer? I said it was the number two answer. The number one answer on the board is snot for 48 points. So I need 30-something here on this next one.

I missed the numberlly. Okay, I missed the number two. Okay. So as is tradition, you get to pick the third question, the tie-break question. You ready? Here are your choices, Jay-Z. Choose wisely.

Name a smell people hate. Name a job you wouldn't want to have if you worked at the circus or name a character from game of Thrones. Number two. Okay, man, I was going to say,

I was going to say, if you said the game of Thrones, I was going to have to hurt you and kick you in the shins because there's no way I'd be able to pronounce all the names. Oh, yeah. I don't know. I don't know. Okay.

So here we go. Wait. Didn't you just – yeah. Yeah. Yeah. Okay. Okay. So, Jay-Z, name a job you wouldn't want to have if you worked at the circus. Lion tamer. Lion tamer. Okay. So Jay-Z, name a job you wouldn't want to have if you worked at the circus.

Lion tamer. Lion tamer. Okay. Alan? Being shot out of the cannon. Okay. Okay. All right.

Here we go. I like this one. This is awesome. This is so awesome. Number six answer on the board, Elephant Trainer, three points. Number five answer on the board for four points, Tightrope Walker. Okay, yeah. Now, here's where I'm really surprised with the both of you though

because the number one answer seems so obvious clown dang it janitor oh yeah see i thought that i actually thought that but i was wondering if anybody else would janitor was 55 points on the board you said clown just now jay-z yeah clown was the number two answer on the board. You said clown just now, Jay-Z? Yeah. Clown was the number two answer on the board for 13 points. Oh, man. Dude. Yeah. Number four answer on the board was human cannonball for six points. Nice.

Nice. Number three answer was lion tamer for 12 right that hot garbage there goes my street yeah I should have thought a clown though that's my bad I actually thought janitor I was like nobody's gonna think of that

they're gonna they're gonna think of other stuff yeah I can't I was like four people in the car I'm like I don't know i better drive i'll roll separate all right all right well there goes my winning streak all right you're so close so my last one have you guys ever heard of Twilio? Yeah. I never really understood what it did, like voice stuff.

All kinds of stuff that is related to telephones. So if you ever call up a place and you get put on one of those automated, you know, push one to do this and push 12 to do this. And then, oh, that's probably using something like Twilio. They may even have done it in Twilio. So it's an automation system that you can use. And it's also good for text messaging, people sending emails, doing all kinds of things, but it's usually related to the telephone lines. Right? So long story short, I was going to try and set up something that was some sort of automated system like that. Right.

And, and really it was going to be something fairly simple. Um, it was basically like, Hey, if you call this particular number, I want it routed. I want it to say something to you specific. And then I wanted it to forward it to somebody's cell phone, right? Dude, that is hard. I spent, I spent so many hours trying to get that working, right. And I got it real close. And it's one of those things where I got so frustrated, just walked away from it. It reminds me of the game Splinter Cell from the Xbox. Like, I got so far in that game, and I got so frustrated in this one point that I walked away from it.

I came back to it like five months later, and what I found out is I just needed to turn on my infrared goggles, and I could see the mines in the field, and I could shoot them and then walk through it. And it was literally the end of the game. That's where I am right now with Twilio is I'm right there at that minefield right before you open the door to the end of the game. Here's the interesting thing. There's a reason why I went this route. There is a, uh, a site, another service called call rail that does exactly what I just told you, right? If you go to call rail, you can set up something, you call this number and you can have it forward to another thing and you can have it do what's

called a whisper, which will be like, Hey, to be connected to your call dial, whatever. And, and it'll do that. Right. The part that drove me crazy is call rail. I want to say it was like 50 bucks a month. It gave you 10 phone numbers and, uh, you know, however much usage, but I was like, dude, Twilio is like two bucks a month. I'm going to code this man. Let me tell you. So it's one of those services that is absolutely amazing. They got tons of stuff, but it is 100% for developers to make automated systems. And they have their own online studio thing that you can drag pieces around and do things. But if you start getting outside, it reminds me of EXTJS, like all the little demos that they show you are really easy.

Hey man, you can do a little form here and you can do something like that. And as long as you stick to that formula, you're good. As soon as you want to do something special, oh, dude, now you've got to set up all these special bins, and you've got to set up all this special code stuff and these deploys. But, man. Can we back up a minute? Because I'm lost.

What was the ultimate thing that you want? Without talking about Twilio, just tell, just describe like, what was it that you wanted to accomplish? So that's what it was like. So you dial a phone number. So you get this phone number from Twilio. They have a bunch of numbers. You or you're calling phone number.

So, so think about this. No, no, no. Who's calling the phone number? You or somebody's calling? a customer,

a customer is going to call your service, right? Um, so let's say that going to call your service, right? Okay. So let's say that you have a lawn service, right? So you buy a phone number for your lawn service. You're going to put that on a website. Somebody's going to call that, all right? They're going to call that particular number.

And what you want to do is you want to forward that to the person who owns the lawn service thing, right? But you want to do it in a way to where they're not getting spammed and stuff. So what will happen is when they call that number, it'll pick up and it'll say, Hey, you have reached, you know, so-and-so's lawn service, press five to be connected or something like that. Right. And then the person will have to push five. Otherwise it'll never get connected. So that cuts down, believe it or not on a lot of spam. And then what would happen from there is that would then dial out to the real phone number, right? So the person doesn't have the cell phone or whatever of the person who owns the lawn service. They have the number of the company that they're calling, right? And so it's a nice way to not be giving out your cell phone number,

but to still be able to take in calls. And from that cell phone, you'll see that the call came through that one place. So, so I said the thing where, you know, press five to be connected. Then the person that's receiving that phone call, the owner of the lawn service will get something that says, Hey, you have an incoming call from, you know, your website, whatever it is. And then, you know, Hey, press five to be connected. And so both the caller has to do something to be connected. And then the person that's being called gets a notification of where the number was coming from, right? Where the call was coming from. So that was the flow. And it seems like a fairly simple type thing, man, I've probably spent 30 hours like collectively over time trying to get

that thing to work perfectly. And I'm close, but I never realized how difficult some of these systems could actually be because you have all kinds of weird things that can happen, right? Like if you think about what happens when a phone call is made, right? Did they not press the number after a certain amount of time? Okay, well, do you have it repeat the thing? Say, hey, press five again. Do you have it hang up? Do you do whatever, right? Do you send it to voicemail? On the other end, hey, did they hang up in the middle of it? What happens when that happens? Does it call back? Because that's one of the things I screwed up on is I would hang up and I'd get a call back right after it. And it was like, oh man,

I didn't close the loop somewhere. So for those who've never heard of Twilio, it's really interesting. It's kind of cool thing to go out there and look at. But I was truly surprised by how difficult it was to do what I thought was a seemingly straightforward task. And I don't think it's their fault i think it's just a very complex system when you start talking about people interacting um with phones people that are calling people that are receiving calls all the things that happen in between those those handoffs here and there and everywhere it was pretty crazy so yeah a friend of mine that uh this was years ago uh back in the mid-2000s he had created his own like switchboard for home for his home oh like did all like none of this was you know like off the shelf kind of stuff like he had uh yeah he he'd he coded everything himself

but when you would call into the number you could like who do you want it yeah you'd press a certain number to talk to whoever you want and otherwise you would go to like oh well this is just a spam you know uh bot kind of kind of call or something right Right. And, and he was, you know, to your point, he was like, yeah, a lot of, a lot of them would just get filtered out. It's, it's truly amazing. Like when I first started researching this stuff, one of the big things was where people were like, man, I'm getting so many spam calls from, from this number. And, and the answer was make them do something that is feedback, right? You can even have like, I will say like, if you ever want to go look at it, it's a pretty And the answer was make them do something that is feedback, right?

You can even have, like, I will say, like, if you ever want to go look at it, it's a pretty cool thing and it's pretty cheap to play around with. And actually a lot of it's free if you're trying to do development stuff. But they have an interface that's like, hey, you can talk, right? Hey, either push five or say, um, Johnny, right. And it was really good at the kind of stuff. And you could even take texts. Like I had programmatic texts come in there that would say things and it would convert it to a human voice. And it was really good. Um, but yeah, it was just one of those things where it's like, Oh, this seems like a really simple thing. And Oh man, Oh man, it was, it's like everything else developer wise. Like you look at stuff that like, you know, anytime that you're given a ticket to work

on, they're like, Hey, how long do you think this will take? Ah, two days, you know, two weeks later, you're like, dude, you're still in meetings talking about it. Right. Yeah. So I don't know. It was just kind of surprising to me, but also a very, very cool product. If you've never heard of it and it's T W I L I O.com.

If you're somewhat interested in it. Yeah. Very cool. Yeah. Cause one of the, like it's come up in other things though. Are they known for anything else? I see them at conferences all the time. They're developer advocates.

I can name a few of their advocates. They're definitely very active. I just never really understood what the product did. Well, like I said, a lot of it, it's text messaging and all kinds of things, right? So I was talking about the phone piece of it, but you can totally do things to where it's like, hey, text to subscribe to something, right? Like your best buy things, right? Like you go buy something the best buy and it's like, hey, do you want to receive text messages for updates on the shipment or whatever? They're probably using something like Twilio so that anytime something happens on their end, they call the

Twilio back in and it shoots a text message out to you. So, um, a lot of those type communications are what they handle there. It's, it's really cool stuff. And it's really cheap. If, if you have the ability to program it and do it, you'll save a lot of money over something using like a call rail, where I believe like call rail is actually using Twilio behind the scenes, but because they've made the service so easy to use, they're charging a premium for it. Right. Which is, is something similar to what we talked about back in the day where like Amazon has an API, but you can go pay for a nicer API where people have made things a lot easier for you, right? It's something similar. You call and make an appointment with your doctor. It's like, press 1 for this, press 2 for this, press 3 for hours.

And then you get reminders about your appointments and texts and stuff. Like, hey, press C to confirm or reschedule or whatever. So pretty soon you'll be able to call in the coding box and you can press 1 for Alan. Right, right. There you go. It hurt my brain a little bit. Yeah. Press the 13th digit of pi for Michael.

Press the babillion key for Alan. I don't know. What's yours, Joe? Number two. We won't dive into that one. Don't know why. That number's come up a lot this episode somehow. It has. Now it's time for my favorite part of the show. It's the tip of the week. Yeah. So a couple weeks ago,

a couple episodes ago, maybe the last one. I don't remember. I gave a tip about basically Docker mounts you could do using build kit. And part of that came or actually the whole thing came out of me doing a usability study with Docker about a product that they just released called DockerNet. So if you upgrade to the latest version of Docker, I don't know what it is, but if you go get it today,

then you'll get a tool called DockerNet that comes along with it. And you can run DockerNet. And what it'll do is, I think it only supports Go right now, but they're very quickly adding other ones. It'll walk you through creating a Docker file for your chosen language and framework. And what's super cool about it is that it does it with best practices in mind.

So by default, it's going to give you multi-stage build. It's going to mount your source code in to be really efficient and keep your image layer sizes small so not just the final image but like small along the way so it's a really nice uh learning tool in addition to just being like a slick command line interface which is really nice so uh give it a shot and see what you think very cool next one uh so i just found out about a utility called Screen, which lets you basically do kind of session management. So what I was doing for is I would SSH into a machine and then I can use Screen and like split my window in half, split it in quarters. I can create multiple sessions and switch through them with like control a and then I would hit

quote and it would list all my sessions and I can go into each one so basically it was like a kind of shell that would create other shells uh tmux is a similar program that I used to use years ago that would do it but I really like the usability of screen so it would basically be control or I guess it's control either way control a and then whatever that's like c to create or d to detach and what's really nice is like you can keep it running on that server and uh you can resume so if you lose your connection to that server uh then you can come back in and resume and it's gonna pick up it's gonna split your windows just like you had it and set everything back up because it's uh uh, keeps, keeps track of basically,

um, what you, what you did to set it up. So it's much nicer than having like a bunch of terminals that like once you shut down, you know, your item or something, you can save those two.

But, um, it's just a really convenient way of, of doing that. And so I thought that was really cool tool and it was surprisingly easy to get started, but I found a really nice article, uh,

which I think someone dropped a link in. Yeah, it's been so long since I thought about that. I saw that you had the word screen written in the notes. I'm like, where are we going with this? And you mentioned utility. And immediately my mind thought like, oh, this is going to be something like Mac OS thing. And then as you were describing it, I'm like, oh, oh my god it's probably been 15 years since i thought about this thing yeah it's really slick and simple you know um you can do control a and

then question mark and it's going to show you all your other commands uh so i really like just the internal usability of it and it's really nice to see it and if i lost my connection i was able to get back in and just pick right back up which is super nice see yeah i don't know i i i never got into it so it was never like one of my uh like go-to kind of tools you know like we all have our you know you get into any kind of like unix like environment you know linux or whatever and like there's your go-to tools and that and i knew of it and i had friends that use it, but I never, it never became part of my like, uh, daily, you know, tools that I use. And it seemed like at the time back then when I first saw it, I was like, it just seemed like it was a little, uh, like the, the, the, the barrier to entry was a little rough at the time and so that was like part of

the reason i was like eh whatever i don't care i'll just like you know re sshn like who cares yep but yeah and maybe it's kind of maybe the documentation is better or whatever so you know i i assume that tmux was uh like older and kind of out of favor but actually like googling tmux for a screen it seems like tmux is maybe even the more popular one as tmux so that's probably a good tip if you're considering this is to maybe check out tmux too people are saying it's more powerful and a little bit easier to use so i'm going to be checking out tmux again because it's been a long time i think when i was using tmux i was definitely not a power user but i mean it was probably 15 years ago so it's kind of funny to hear that it's still doing really well yeah that's uh all right well there's a

an oldie but goodie yeah thanks to all those in there as well in the show notes which will be codingblocks.net slash episode 208. Yep. And then the last one I had in there was CodeWhisperer, which is the AWS alternative to Copilot, which only supports a couple languages, but if you're doing stuff with AWS, then I would check it out. I'm not doing stuff with AWS. I can't vouch for it. I just thought it was interesting, and it just got

announced like today. Yeah, that is cool. So, okay. My tip of the week is rather bland and generic, but get up. Go for a walk. So I put a link in here to a thing on Amazon that I found where, like, if you already have a treadmill, right. I found an, a desktop attachment that you could put onto your existing treadmill and you can use it as a desk. Right. And this particular one has like a platform that can rise, you know, kind of like your standing

desk kind of thing. So depending on like what height you want the thing at your, your laptop or at the time, you know, you might want that if you don't want it to ever, if you don't ever want to change the height, there are less expensive versions then. But, uh, for me, I didn't know like what height that might want the laptop at. And so I wanted the ability to like raise and lower it. Cause I thought I would want it super high. It turns out I actually wanted it pretty low. Um, but, uh, so I bought this as a way to just test the waters of like, Hey, what is life like if, if I decide to walk while, while working and cause we have, um, you know, coworkers, there's people in our slack that swear by it right and i i was uh you know kind of pessimistic about like yeah i don't know like i want to try it but

you know some of those treadmill uh the the under treadmill um the under desk treadmill units that you'd find some of them are pretty pricey. I know Alan had one as a tip of the week years ago. And so like, this was my way to like baby step, you know, no pun intended into this to see like, you know, how would I like it?

Turns out I absolutely love it. I've had it now for a couple of weeks. You know, I'll, I only use it for like a couple hours a day. I'll, I only use it for like a couple hours a day. I like, I only, uh,

use the treadmill desk a couple hours a day. Otherwise I'll just go back to my regular desk, you know? And I figured like, I figured my thought was like, well, let's start, uh,

you know, you know, a couple hours of solid walking, you know, I'm averaging like five to six miles a day walking while working. And I'm like, that seems like a reasonable start for it, right? But the beauty of it is that's like however many calories or whatever that you're burning doing that while you were already doing something anyways. And I thought it would be a huge distraction working-wise.

I wasn't sure how well it would work. Turns out, it's just fine. I'm able to just focus just fine on it. It doesn't matter. That's pretty awesome. And yeah, this thing is, he already has a treadmill, and it looks like it's really well built. And actually, we had recommended, this has been years ago during one of the Black Friday episodes,

we had recommended one of these same companies, Vivo's desk ones, that would allow you to turn a regular desk into like sort of a stand-up desk. It's the same company, and this thing has great reviews. Yeah, and they have, like I said, they have a cheaper version that doesn't change height and, you know, if that's all you wanted, then I think that

one is like, I don't know. Let me see. Oh, I found it. Well, this one's only $100. Like, it's not crazy expensive for what it is. The other one, if you just wanted just the tabletop, it's $50. Okay. Yeah, that's a good price. Add one, if you just wanted just the tabletop, it's 50 bucks. Okay. Yeah.

That's, that's a good price. Add $5 if you want a cup holder. That's pretty awesome. So that's pretty good coming from you saying that like it ended up like, cause I'm sure we're all a little bit skeptical. Like, man, there's no way I can type and do this stuff. Right. Um, but here in the, the, I mean, you've been a developer for a long time and you just picked it right up.

That's kind of awesome. Honestly. Yeah. Like I, I acclimated to it like amazingly quick. I, it came in on a Sunday and I thought, okay, well, you know, let me just try it out. I'll, I'll, you know, just browse the internet, you know, on a Sunday afternoon while doing it and didn't think much of it. Right. It had in like, you just kind of like forget that you're even walking like whatever and then um but then i was like okay well the real test will be the next day when i like let's let's try to concentrate on an immediate you know and like immediately i had no problems i was like right

back into like dev life you know it's like you know slaying dragons because that's what i do right so at least one of us does yeah it was it was awesome i i've now i'm like oh man do do i go big so i you know i don't know where i don't know what i'm going to do now you know i think i'm at that 1500 one i think well i think for the time being uh you know i i don't know that i think for right now i'm going to continue on with this one you know um and if if my hours per day on it like start getting obscene you know that's when you know i would want to consider because the one difference is that like on a regular treadmill versus the ones that are made for the under desk you know usage those are

kind of purpose built to like not run at high speeds whereas that your typical treadmill is meant to be at a faster pace you know it's in it's engineered to be it can go at a lower pace but they typically think that you're going to like go on it for like you know some serious exercise so you're going to run or go to a walk at a fast pace um and the their your typical treadmill isn't designed to run for continuously for hours on end, right. Right.

Versus the, some of the under desk treadmills, you know, they're, they're kind of more purpose built for that sort of thing. So, um, yeah,

I'm going to put a link to the lifespan that I know that you've looked at in the past and the max speed on that one is four miles an hour. So yeah, it's, it's much lower than, than like your standard treadmill. Yeah.

I was walking at three miles per hour today while, while on the treadmill. Yeah. That's cool. That's cool. And, and like I said, the beauty of this,

like that's, that's some great, like, you know, great boost of energy to yourself, you know, because the problem is, you know, this is such a sedentary lifestyle, like as a software developer, right? Yeah. You know, I mean, yeah.

So whatever we can do, like cheat the system, like there, that's, that's what we do. I like it all right so for mine this one came out because i got we were working on something and we had this really odd thing so in kubernetes everything could communicate with each other except like there was this one pod that like anytime you try and hit it like there there was a port exposed, it was just like, nope, can't, can't get there, you know, rejected whatever. Ed tried several things, kept looking around, couldn't figure it out. So the behind the scenes, longer story of it, and I'll keep it fairly short is we're using the Strimsy operator. We've talked about it before.

It's like the, the Kafka, the Kafka operator that is sort of like, I guess, sort of the de facto one, if you're going to use the operator for Kafka. At any rate, anything that it deploys, it kind of wants to lock down communications so that the operator can communicate with those pods and those pods can communicate with each other, but it doesn't really want anything outside of those pods doing anything to them. And so what it did is Strimsy actually sets up what are called network policies in Kubernetes that basically disallow Ingress from anything except for pods that it controls. So if you ever find yourself in a situation where you're trying to hit some pod, you know the thing's available, you can shell into that pod, you can hit itself, you can do all kinds of things, but for some reason, hitting it from

another pod's not working, take a look at the Kubernetes resources called network policies, and there may be one in there that is cutting the ingress into that pod. And you could do all kinds of crazy stuff with it, right? Like you can, these network policies are pretty powerful. I've got a link in the show notes to them. You can set it up. You can do something crazy. Like, Hey, everything in my namespace can talk to each other, right? Like you can set a global thing for your entire namespace or you can lock it down by labels because you know, if you've done much in Kubernetes, they like you to use labels for a lot of targeting type things. And you could say, Hey, anything that has product, you know, e-commerce

on it can talk to this, but anything outside of that can't, right? So, um, you have a lot of control and, and it's not just for ingress. You can set egress rules as well. So, hey, this thing can't talk to anything over there. It's only allowed to talk to database server or whatever. So you have a lot of control and a lot of power over there. So if you didn't know about this, it can cause you a lot of problems. But now that you do know about it, it can also help lock down and secure your application. So definitely worth checking out. All right. Well, you know, as I asked earlier, if you haven't already, you can subscribe to us on I stitchify.

And yeah, you know, leave us a review to while you're at it. You can find some helpful links at www.codingblocks.net slash review. Yep. Hey, and while you're up there, make sure you check out. You let me get away with iStitchify. Okay, go ahead. I knew it was going to happen. That's the name.

Yeah, that is the name now. It's been dubbed. Yeah, check out our show notes. Sometimes have discussions up there, examples, all kinds of stuff up there. And hey, if you're not a member of our um sometimes have discussions up there examples all kinds of stuff up there and hey if you're not a member of our slight community go up there codingblocks.net slash slack and hit us up and and all the amazing people up there and i hate we got a twitter i just and don't forget if you want to make it to view js in lS in London then you need to use the hashtag that I still managed to forget

I'm not sure if they have a hashtag but we're going to have it in the show notes for comicbox.net slash 208 and make sure to just mention us in a tweet about the conference, use their hashtag and we'll search that and we'll find you and give you some tickets