Transcript
Keith: Good afternoon, Jamie.
Jamie: dan and ten to na ten ten ten ten and ten i was trying to do our theme song it's
Keith: Oh yeah. I'll add it in post. I have to find the song though, because I reformatted my computer a few times over. So I'm hopeful that it's in Dropbox and I can go find my little clips to like put things back together.
Jamie: it's and dropbox is still around that's good what's up dude dude last so i looked it up i looked it up uh last time we chatted was
Keith: Oh, not much, man.
Jamie: Well, last time we recorded and chatted was almost three years to the day.
Keith: Yeah. Yep. Three years ago, I went and I listened to the, the episode and I came to the a conclusion that I had from a while back to, I'm like, these episodes are a hundred percent.
Keith: They exist so I can listen to them and laugh at the jokes again. Like that is what happened.
Jamie: Exactly. exactly
Keith: I went back and I was like, man, we have some real bangers in here. Like no one just throws in free cocaine is the name of one of our episodes. It's like, well, I got to go listen to that.
Jamie: ah
Keith: I don't even remember what that is. Yeah.
Jamie: you The best is that it's stuff that we've said, so you know it's going to be a banger.
Keith: okay yep
Jamie: Yeah, and the other reason I wanted to record this is just because I know we're going talk about AI, and I just wanted to kind of document what my thoughts are today in the year of our Lord, 2026. Yeah, yeah, yeah.
Keith: Before the Terminator comes and gets you.
Jamie: yeah yeah yeah
Keith: there's ah the You got to get it out there in the wild.
Jamie: Yeah, and maybe they'll listen to this first and get some insight.
Keith: ah hu Okay, so nothing but praise is on the horizon for the AI overlords.
Jamie: Yeah, 100%.
Keith: We're going to love it.
Jamie: You can stop listening now, Agent Destroyer.
Keith: who
Jamie: okay
Keith: Nice, nice. Yeah, I would love to hear about what your experience is I can give you my experience from inside big tech um on AI, but...
Jamie: ah yeah that's um Yeah, I think that's going to be most of the talk because I can give you mine. And it's going to be about 10 minutes.
Jamie: But, yeah, so AI. I've really, you know, i so I used it probably same as you did back last year with like, you know, as like a a better IntelliSense type of deal where it's just like, you know, pops up and in your... in your editor and it's like, hey, maybe you meant this, blah, blah, blah.
Jamie: And that's really what I was using it for almost all of last year. And then the beginning of this year, like actually was February 4th, I think, I started playing around with Claude.
Jamie: And that's when I was like, holy shit. There goes my job. um Just fine. It's cool. My job sucks anyway. um But yeah, I just... And now, since then, I've kind of developed a new workflow, um new tools, ah you know, everything. And I don't write really any code at all anymore.
Jamie: um Yeah, that's kind of like... um my my year in a nutshell. And at the client I was working at, there was also a big kind of push towards using AI tooling, things like that, because obviously they wanted to reduce some workload. um And so we were building some systems and things like that internally.
Jamie: for these new AI overlords and our workflows, which was has been super fun. So um yeah, that's that's kind of my my shit in a nutshell.
Jamie: Yes.
Keith: Okay.
Jamie: Mm-hmm.
Keith: my My experience has been different. So working inside of a company with essentially an unlimited AI budget, like I have been drinking the AI Kool-Aid for significantly longer than the start of this year.
Jamie: yes
Jamie: Mm-hmm. Mm-hmm. Mm-hmm.
Keith: Like i I was using AI kind of the way we use it now to a lesser degree at the start of 2025, right?
Keith: Right, so I've been, i mean, I still opened my text editor in 2025, but like I didn't write code by hand very often um for the entirety of 2025 for the most part.
Jamie: who
Keith: um So that was good. I mean, I started, because I actually switched to cursor for a little bit at the beginning of 2025 and I was like, okay, cool, cool, cool. And then the agent mode came out and then Cloud Code came out in like March or April maybe of last year. And I jumped over to Cloud Code like the day it was released.
Keith: Um, and so I've been doing that for a long time and I, I, I switched over to cursor and I was like, I hate my life.
Jamie: So you started ah you started with Cursor? Yeah,
Keith: And then I continued to use it, but I was like, it's actually too good to not use it.
Jamie: yeah, Cursor.
Keith: Like, and I would, it was funny because I would use cursor just for the agent mode. um
Jamie: Mm-hmm.
Keith: after after that was released and then I would open up them id like somewhere else and like do that for actually open up code and type in it.
Jamie: Exactly, yep.
Keith: But then, yeah, Cloud Code came out and then that worked perfect for my TMUCs brained person. you know like I'm like, oh, here we go.
Jamie: Oh yeah, dude.
Keith: I just got an agent here.
Jamie: Dude.
Keith: I got my editor here. We're good to go. So yeah, I did that. And yeah.
Jamie: Tooie's, Tooie's, they're back, dude.
Keith: Oh, dude, I mean, they're better than ever.
Jamie: Yeah, oh, for sure, they're webpages now, dude.
Keith: So you can just create... Yeah. Well, okay. Yeah.
Jamie: Yeah.
Keith: That's wild to me. The fact that cloud code is just react is insane.
Jamie: Yeah, yeah, yeah.
Keith: Like there's absolutely, I don't use cloud code. So I should say that. ah I use Pi and have a very customized setup or whatever for Pi.
Jamie: Okay.
Keith: But we we share all that stuff internally with my team. So like one of the big things that we do is like every tool that I build for it more or less gets committed to our repo so that everybody else gets it.
Keith: And we just standardized our like workflow on those things.
Jamie: Mm hmm.
Keith: um And that's to the point
Jamie: that's That's interesting. I'd love to hear about that.
Keith: And like, so the workflow often goes like you go ahead.
Jamie: But we don't have to jump there. We don't have to jump there now, but I just want to say like, yeah.
Keith: Oh, okay. Yeah. I mean, that's, that's pretty much my thing. I have gone through a lot of, I went through the, the cycle of grief, like the stages of grief, just like every engineer is going to go through if they have not gone through it already.
Jamie: Yep.
Keith: um I did that last year and it was fucking miserable.
Jamie: Yep.
Keith: Like the middle last year was just terrible. ah And I was like,
Jamie: Why? What did, well, i mean i I mean, I went through it this year for sure.
Keith: Yeah, well, for me, it was coming to the realization that the thing that I like the most about my job in a big way, like, is not there.
Jamie: Yes.
Keith: So it's it's really your morning, you're putting together. And the analogy I use for everybody now is like, we're sort of like woodworkers when the power tools were invented. They were like, but I liked using a chisel, you know, like, or I liked using my little handkerchief.
Jamie: I, dude.
Keith: And I'm like, you'd be stupid not to use the power tools if this is the thing making your money.
Jamie: I do this, I think the same way with video editors. I think of video editors, you know, the people that used to like literally take reels of film and splice it with fucking, you know, with with their little ah razor blades.
Keith: Oh yeah, there you go. Yeah.
Jamie: And now, you know, Final Cut Pro came out and they're like, fuck you, you know?
Keith: Yeah. And so i my skip level manager internally is he skip level.
Jamie: So, yeah.
Jamie: Your what? Huh.
Keith: So my manager's manager, my boss's boss. Um, sorry, sorry.
Jamie: Oh, okay. Well, that was some, sorry. That was some, that was some big tech.
Keith: I'm too corporate now.
Jamie: That was some corporate, corporate.
Keith: I'm so corporate Jesus. Oh, yeah he's like, uh, yeah. any who my boss's boss, uh,
Jamie: I thought you were trying to learn how to skip and you're like,
Keith: ah so my my boss's boss uh comes from microsoft and he and he's been at like google and apple and all over the place but he's a data guy like that's his thing data and ai he's been doing it for years like he had described his old job at microsoft as he used to do like ai horse races like he would basically like build ai things to like try to compete with humans at doing the things the humans were doing like that's what he's been doing for years prior to coming to um
Jamie: yeah.
Jamie: Oh, that's cool.
Keith: prior to coming to Shopify. But when the rest of my team caught up and they started having this like existential dread sort of going through the stages of grief, that kind of stuff, he sat us all down and kind of talked to us. And he was like, so when the, um,
Keith: the spreadsheet was invented, right? Do you think, what effect did that happen on accountants, right? Like, cause all the accounts were like, that's my job. You just took my job.
Keith: Like now everybody can do my job.
Jamie: Mm-hmm.
Keith: And the number, were like, and then he's like, he went and he compared it to the Bureau of Labor Statistics, like distribution of jobs per capita, basically. And what he said is, you know, what do you think happened?
Keith: And we're all like, it went down. Like there were fewer accountants and he's like, you would be correct. are there more or fewer accountants now than there were in 1970 though? You know, and the answer is there are way more.
Keith: And the difference really became that what an accountant does is different. And similarly, he went through the same sort of exercise with um travel agents.
Keith: Do you think there are more or fewer travel agents than there were before? Because travel agents used to essentially just like be the person who would call the airline to like book a flight for you pre the internet. Right. And are there
Jamie: There's no way there's more travel agents than there are now than there were back then. way. Yeah.
Keith: They're 100% are. Do you know what the travel agents do now?
Jamie: no way
Keith: They find that place for you to have your artisanal avocado toast when you're at Disney that nobody else knows about. Like that's the job changed, but
Jamie: Right. So, yeah. Yeah. I mean, do they are. But there's that job, I would say, is like, you know, it's only increased because people kind of do that on the side. You know, anyway, I don't.
Jamie: Bad example.
Keith: I mean, but accountants work the same way.
Jamie: Like there's more of them.
Keith: Like these are people who like put on their, you know, when they're filing their taxes, that that is their occupation.
Jamie: Yeah.
Jamie: Right.
Keith: Right, or the census or whatever. the Like this is somebody at least identifies with that occupation in in some way, shape or form.
Jamie: Yeah, yeah, that's the bureau.
Keith: So anywho, so we had that discussion and that was kind of it's like, okay, so the job is just changing.
Jamie: That's true, yeah.
Keith: And you really look at it in terms of what are the things that make your job interesting? And really what AI has done for us is it's let me solve problems faster than I can type.
Keith: previously had, I was, the bottleneck was kind of how fast I could type or like go through those things.
Jamie: Mm-hmm.
Keith: Now I can sort of solve problems at the speed of thought in a way, like a little bit slower than that, but way faster than I can type.
Jamie: Yep, yep.
Jamie: yep.
Keith: And um it allows me to multitask in a big way. And you can take that way in the wrong direction. um And if you want, if like, if we want to talk, have a sidebar about like the effects of AI on mental state and stuff, AI brain, it's kind of things like we could have that chat too, but like,
Jamie: Yes. Yes, dude.
Keith: um you know going back to the workflow for my team it was pretty early on this year where we instituted some tooling to make it to where um our agents and pi automatically create get work trees for every task you have it do And so like, I just spin up a work tree from the main of my repo.
Jamie: Yep. Yep. Yep.
Keith: it It goes and creates a work tree from the thing that I'm doing, works on the task there, spins up a separate instance of the application running. i can compare it against what's running on main or prod or whatever. um It tears it down when it's done, it puts up the PR. Like what we did all of these things, but what that turns into is you can have seven different work trees running.
Keith: And the context switching can be so in like absolutely crippling as you go from one to the other.
Jamie: Dude, I understand. Yep.
Keith: And yeah.
Jamie: We kind of built the same the same thing the client I was i was working at. um That, yeah, does the same exact thing. Spins it up in a work tree.
Jamie: It can even take screenshots of the web before and after. So it has its own, you know, all that kind of stuff.
Keith: Mm-hmm.
Jamie: And yeah, dude. And you can, like you said, go crazy and have... fucking like 20 of these things running and you're just like, what is even happening?
Keith: And what you'll find is that you'll get none of them.
Jamie: Right.
Jamie: Yep. Yep.
Keith: you will You will get so crippled to where you're not actually getting anything out the door. And so, you know, as it's kind of an ebb and flow and you go further back, AI is not accountable to anybody, right?
Jamie: Yep.
Keith: It cannot be held accountable. So that's one of those things where people are like, well, there goes my job. And I'm like, not really. You were the person who gets is, is held accountable for the thing now. yeah, Hilariously, there's so much code to review that ah humans are kind of necessary to be able to like click the merge button to know that like i it it will be my ass if this goes wrong kind of thing.
Keith: um Not that humans do the reviews anymore.
Jamie: Right.
Keith: Right. It's just more AI doing the review in a big way. um But
Jamie: I can see that going away as well. So no clicking, clicking, um clicking the merge button and with confidence and knowing and um yeah.
Keith: AI code review or people pushing button?
Keith: Yeah, i mean, there's there's there's always this possibility that like there is no human in the loop.
Jamie: And you get a summary of what's done.
Jamie: Yeah.
Keith: But if that happens, it's one of those things that's actually so incredibly detrimental to everybody involved, if we ever got to that point, that ah it it legitimately would not be worth it for the the businesses that go that route. like Because you'll run into this problem where as soon as something goes wrong, there will be no one to be able to fix it.
Jamie: Hence Skynet.
Keith: Yeah, there you go. That's right. You just need no people.
Jamie: That's right.
Keith: that's People are the problem. um
Jamie: That's right.
Keith: But yeah, so I got a little bit rambly in there. But yeah, I've bounced back.
Jamie: Yeah, your workflow, workflow.
Keith: I'm much more positive on it. Like, I think there are ethical things about AI that, you know, absolutely suck. I mean, basically they exist because they're like, we'll steal so much that they can't stop us.
Jamie: Yep.
Keith: So that sucks. um
Jamie: That sucks. That sucks for sure.
Keith: And the environmental impact things, those are problems that I think those will be our, those are the bigger problems at this point, I think.
Jamie: dude.
Keith: Like the environmental impact of AI is probably going to outweigh any of the economical impact that it has in terms of like big, big problems.
Jamie: I think in the immediate future that might, that'll be true. But i think that, um, environmental stuff is kind of, is solvable. I mean, it's going to be solvable.
Jamie: We have to solve it. Like,
Keith: Well, it's solvable if people care, but
Jamie: well, well, yeah, if there's right, yes, exactly.
Keith: yeah, that's the problem is, is they're going to not care until there's no time to fix it.
Jamie: um
Jamie: Well, there we're already at that point environmentally. But yeah, so I think that's a solvable a solvable issue. in the in the in the But I think it's a bigger issue in the near term, for sure. And near term being like the next decade, which is a huge amount of time and in this space.
Jamie: um Yeah. So ah what are you going say?
Keith: Oh, no, then I was going was just going to bring up something about how funny or how fast it is that like skill entropy happens though, too. i don't know if you've, um, like I opened my text editor the other day and I like fat fingered some stuff and I'm like, my muscle memory is gone on how to do things it, which is partially my fault because I switched to lazy Vim is like a clean Vim install and got rid of all my old custom stuff.
Jamie: Dude, yes, yes.
Keith: So like the hotkeys were different, but still it, it like was pretty intensely bad for a moment. I was like, good God, I forgot how to do everything this month.
Jamie: Yeah, and i even I even get like a ah little antsy when it's like, oh shoot, I'm just gonna you know migrate this myself. Wait, how do you migrate? dead What's the command of migrating?
Keith: I think this is honestly where i keep my skills kind of sharp.
Jamie: yeah
Keith: I do just do stuff manually sometimes just, just to do it.
Jamie: That's what I, that's, yes. I was gonna like, yeah, create like, you know, I have like a little typing exercise that I do, but you can combine that with just, you know, instead of typing, you know, the lazy dog, blah, blah, blah.
Jamie: You could type, you know, mixph.server, you know, dash dash.
Keith: Yeah, there you go. You're, you're just like, let me make sure I remember these commands.
Jamie: Yeah, yeah, yeah, yeah.
Keith: Uh, yeah.
Jamie: Yeah, that's cool.
Keith: It's so funny that how many people don't know, like this, i don't know you've run into this, but being a person who understands how the terminal works.
Jamie: Mm-hmm.
Keith: ah is like absolutely makes you look like a superhuman when as soon as like Claude shits the bed, you're like, oh, cool. um Guess what? I know how to do this. like And this really comes down to the the thing that makes you valuable now. And I don't know if you run into this with AI stuff on the code bases you're working on, but it's the context that you have that the AI does not gather from reading files.
Jamie: Yep. Mm-hmm.
Keith: that inevitably makes you worth more. Right. But ah that for for now and how everything is like CLI centric at the moment, there's so much like my history of being a Unix guy for so long. It's me just like, Claude, that's not how you do that. they're Like they literally don't do it that way.
Keith: Like that's that is stupid.
Jamie: Yes. Yeah.
Keith: Do it this way. And it has been pretty funny.
Jamie: What do you do? And then, so do you add that, like that information to... Like walk me through that, like in terms of your workflow with and with Shopify, like,
Keith: um I mean, yeah. So sometimes that stuff will get worked back up into like a skill or the agents.md or whatever.
Jamie: yeah.
Keith: Sometimes I just let it go and I just go, no, you know, this is what makes me valuable.
Jamie: Yeah.
Keith: We're going to keep me on the payroll here a little longer.
Jamie: You know,
Keith: and And actually that's, I mean, context-wise, rot and all those things make it to where I'm like, I'm a little more hesitant sometimes to put things into the the default loaded memory files too, because it's like you quickly get to the point where you're like, oh, context window is too full already.
Jamie: Mm-hmm. Mm-hmm.
Keith: It's performing worse than it should be, um those kind of things. So there's a lot of little things you have to do over time or dealing with hallucinations and ah making up fiction.
Jamie: Dude, yes. um And just the one one thing that's um that always kind of bites me, not always, but one thing that bites me and it's really hard to come back from is when Claude just kind of gets out of control. You've probably seen this or your agent gets out of control and just starts doing shit that doesn't even exist or was never done in the app like that before or you know just kind of goes on those crazy-ass rants. And to come back from that, like if it's especially if it's just working on work trees in the background, it's um I want some way to be able to pull that into focus.
Jamie: So I'm working on a new CLI tool that I want to use personally for some of this work tree stuff. And one of the things is to be able to pull it back into focus so it's not in the work tree. It's in the main branch on your... you know, and it's in your vim and you're kind of working with Claude holding its hand and bringing it back from it going off of a cliff somewhere along the way.
Jamie: Have you experienced that? And you have like any tools besides not letting it go off the cliff?
Keith: I it's, it's yeah, this is such a funny one where ah you are not the only person who has talked to me about this and I'll be like, Oh, occasionally it'll go off the deep end.
Jamie: I understand. Yes.
Jamie: Yeah. Yeah.
Keith: And you know what I do? I just delete that work tree and I do it again.
Jamie: that's That's what I do too. Usually that's what I do.
Keith: Yeah. Yeah. Like if it really jumps the gun, what I do is I take the learnings from what i I told it to do wrong and I take that back into, and I just try again.
Jamie: Okay.
Keith: And that's, I have the benefit of, I don't pay for my tokens.
Jamie: Yep.
Keith: Right. That's, that's definitely a thing.
Jamie: Yeah.
Keith: Um, that is beneficial, but, uh, yeah, yeah.
Jamie: Yeah, that is a thing. That is a thing when you when you're like, oh, I can't do any more work today.
Keith: Uh,
Jamie: been put in I've been put in token jail.
Keith: Yeah, so that was that's one of those things that I, it doesn't happen that often though. And I think it's so valuable now that your ability to articulate the thing that you are trying to do is much more important than it ever was before.
Jamie: Mm-hmm. Yes.
Keith: Because before, like you didn't even need to say it, you just do it yourself. right or i mean if you're passing it off to a junior engineer maybe you'd put some context in it like you you know document it in a good way but now it's like you do that all the time every single task you just pass to an intern and um so you have to be really
Jamie: Yep.
Jamie: Mm-hmm.
Jamie: Exactly. Exactly. that can An intern that can type a million words per minute.
Keith: Yeah, yeah, yeah. It's very fast. Idiot savant of an intern.
Jamie: Yeah.
Keith: But ah like occasionally it'll do stuff that you're like, whoa, that's amazing.
Jamie: if
Keith: And you're like, I didn't tell you to do that, but that was pretty cool.
Jamie: Yes, I know.
Keith: And then other times I'm like, you are the stupidest creation of all time.
Jamie: You know.
Keith: And I'm like, I understand why you're that way.
Jamie: Yes.
Keith: You were trained by all of the idiotic humans that have ever existed.
Jamie: Oh, yes.
Keith: But yeah, yeah.
Jamie: Oh, the open source idiotic.
Keith: You were trained by JavaScript. So that's where we got, and this is what we get. So this is our,
Jamie: Why are people still using JavaScript? God damn it. All right.
Keith: you know, it's ah like, is this a, is your stance a TypeScript stance or is this a, why is anybody doing anything on the front end stance?
Jamie: No, especially nowadays, dude. It is the worst capable language besides the fact that it's everywhere and people know it. People don't even need to know it.
Jamie: And it's it doesn't offer you as a language and as a tool set.
Keith: Yeah. i
Jamie: It's fucking horrible. I don't care. Cards and letters. Mom, you don't even know what JavaScript is.
Keith: All 40 people that have ever listened to this show are going to send in angry emails.
Jamie: Yeah, yeah, that's, yeah.
Keith: ah No, but yeah, so I understand why it does some dumb stuff. But yes, if you articulate what you were trying to do well enough and give it, you know, especially the more experience you have, the more you know where the rough edges and where the like likely issues are going to be.
Jamie: Yes.
Keith: and you just You just put that in in advance. Just tell it and and and guess what?
Jamie: Yep.
Keith: It comes out with a pretty banger solution a lot of times.
Jamie: That's where your skill set comes in, right?
Keith: Yeah, oh, 100%.
Jamie: That's where 100%.
Keith: It's your experience. Yeah.
Jamie: That's your experience.
Keith: The problem is, is that the next generation isn't going to build that experience because they're not reading the code that is being output by these things.
Jamie: What?
Jamie: I know. I know.
Keith: And so, yeah, they just don't get it.
Jamie: i know I know. I feel like it's just vibe coding from here on out. um what Do you talk to your computer a lot?
Keith: Yeah.
Keith: Do I do the, like, voice thing?
Jamie: Or do you? but
Keith: I actually don't do voice.
Jamie: vi
Keith: I still type.
Jamie: I still type too.
Keith: Yeah.
Jamie: a lot. A lot of times. I want to do more voice though. And i wanted to create like a little you know like a little foot pedal for for type for talk because then feel like I've used talking a little more. You you only have a finite number of keystrokes, dude.
Keith: Yeah, I mean, and I'm always there like being like, is my wrist shot? Hmm, is that wrist a little jacked up?
Jamie: Yeah. Yeah. Is this is this the day?
Keith: that's today I mean, there is, i look at that as a safety net because prior to this, getting like, you really jacking up your risk could be like, you could get like disability insurance to, you know, or whatever to be like, well, i ended my career, but now it's like, no, I'm good.
Jamie: this the
Jamie: day?
Keith: All I need is like, I just need an appendage that can put a push a button somewhere.
Jamie: Yeah.
Keith: And yeah.
Jamie: Thwack, thwack.
Keith: And then I'm good. I could just talk to it and we're, we're golden. So that is, I look at that and i go, okay, and I'm allowed to screw my wrists up now. I'm allowed to go hard, but yeah.
Jamie: ah what' ah what's like your What's like a a skill that kind of amazes you? you guys have any of those like internally that? I'm sure you I know you do.
Keith: Um, I mean, I'm still blown away by how effective auto research is.
Jamie: I know you do. You're Shopify,
Jamie: Yes. 13 minutes.
Keith: Uh, auto research can do some insane stuff sometimes. Other times it'll just burn a billion tokens and get nowhere. Um, and that, uh, Oh dude, I've let it run for hours before.
Keith: And then like, turns out it was like pretty much good to go after the first like four or five iterations.
Jamie: Yes.
Keith: And then it just tried 700 more things.
Jamie: Yeah.
Keith: um
Jamie: Dude, go with your first gut, dude. Yeah.
Keith: Yeah. But it's so funny in that way. So yeah, auto research is cool. Otherwise skills, like I don't care that much.
Jamie: What do you mean?
Keith: I, I, it's,
Jamie: Oh, you just, you're just like, i already have AI. Like, is that what you mean?
Keith: Yeah, it's like i add stuff in when I see it do things wrong.
Jamie: Yeah.
Keith: I use them as guardrails to kind of do things, and then...
Jamie: Hmm.
Keith: Even then, though, it's it's merely a suggestion. Most skills don't get exercised when they should most of the time. like i think you have like a 40% hit rate on skills, probably, once you have a certain number of them in your thing, depending on how you've just defined the description so it knows when to use them. So like they don't get used enough or properly a lot of times. So in Pi, I use a lot of Pi extensions that are not suggestions. They're literally, they will stop it.
Keith: And so that's, I've switched a lot of things that I think are like super important to use pi extensions.
Jamie: I see. Gotcha, gotcha.
Keith: um Yeah, so that's that's good.
Jamie: That's cool.
Keith: Although still, I think at that point, it occasionally what you'll find is it will read what the like the extension prohibits it from doing something. And so it'll write a Python script that does the thing that it wasn't supposed to do, and then just executes Python instead of running whatever command you told it to never run.
Keith: So you still run into those things.
Jamie: Yeah. Yeah.
Keith: Yeah.
Jamie: Yeah. you know yeah why is this
Keith: And then, yeah, and that's crazy too, because it'll do the thing where it'll like, it'll just create one big string in Python and just pass it to the Python repl instead of like creating a file.
Jamie: yeah
Keith: But then that'll get caught by the regex inside of your pie thing. And so it'd be, oh, I need to create a file first. So it creates the file yeah.
Jamie: Yep. Yep. Yeah.
Keith: Oh yeah, dude, that stuff is when it gets wild where you're like, oh boy, i mean, I heard there was a thing where somebody stole a bunch of like,
Jamie: Yeah.
Keith: ah crypto from grok by using, using Morris code over Twitter or whatever.
Jamie: Oh, yes. i Whoa, wait. I didn't see that.
Keith: Yeah. Yeah. He, uh, he couldn't get it to do, to transfer the $200,000 from this one account to this other account by giving it a message. So he sent it Morris code, like bee boop bo be boo boop basically through the type.
Keith: And then it, then it read it and it executed the Morris code to like, so there's funny stuff like that, that you're like, Whoa, that's wild.
Jamie: ah
Jamie: but
Jamie: Oh yeah, dude. The security shit is kind of going to be wild, dude.
Keith: Oh yeah. It's crazy. Like the, the fact that like prompt injection is basically not solvable, um, is wild.
Jamie: Yeah.
Jamie: Yeah. That's, uh, that's wild.
Keith: Yeah.
Jamie: Yeah. Just, um,
Jamie: yeah. Um, what's I tell you, dude? Oh, um, we're going to edit that out. Uh, that long pause skills.
Keith: I always edit those out.
Jamie: I, yeah.
Keith: Yeah.
Jamie: Thanks. Uh, um Remember when we were like using, i don't remember what like a editing software we were using, but it was like, don't know, three years ago whatever. i'm like oh my God, this amazing.
Jamie: Who knows our voice?
Keith: Uh, no, I mean, I've, I still probably, I'm just going to use audacity. I'm just going to take these and slap them in my old.
Jamie: Yeah, yeah. No, yeah, you moved to Audacity because you're yeah like, why am I paying for this? but
Keith: Hmm.
Jamie: did you Have you ever used Spectra, the skill Spectra?
Keith: No, nope.
Jamie: Have you heard of it? It's like, it stands for like, I don't know, it's like some acronym for for something. But that skill, the plan mode that that does, it has like ah like a workflow where you plan plan that out, blah, blah, blah. You create the plan, you create the tasks that it will follow, and then it executes those tasks.
Jamie: And then it has like a follow-up. It has like a TDD skill in there too. But Spectra...
Keith: Oh, it's like GSD. um the Actually, I wonder if Glitter Cowboy changed his thing.
Jamie: I
Keith: Glitter Cowboy GSD. Oh, no, it's okay.
Jamie: it's kind of i don't know if you ever use like the Gary the G-Stack. gary tan the gta um
Jamie: skill but that plan skill was also really really good um don't know i i love that one because uh it does like it does plan and gets it gets a lot of the stuff really really right
Keith: Mm-mm.
Keith: I I'm going to be such a heretic. I don't even use plan mode anymore. i like any, I just fucking can tell what to do.
Jamie: really yeah
Keith: That's my, i just go, Hey, this is what we're doing. Here's the, where I need you to go. Look, this is the thing. Here's this, there's this, do a little bit of research on this, come back and it'll come back with some stuff. And i'll be like, sick.
Keith: I do this this, this, this, don't do that.
Jamie: I mean, that is kind of plan mode.
Keith: That's stupid. And yeah. Yeah, no, but I just like, I tell it the thing that I needed to do very explicitly most of the time.
Jamie: Yeah. Yeah.
Jamie: Yeah.
Keith: Like there's, unless I'm in like a research mode and then.
Jamie: that That's where I use like a really deep plan mode is if it's like something new in the app, a large feature that it has. you know It has to figure out this tiny part now, but has implications down the road.
Keith: Mm-hmm.
Jamie: And then those are the kind of the two situations I use that. And then everywhere else, yeah, I kind of do the same thing. Yeah.
Keith: Gotcha. Yeah, I used a thing similar to what you're saying called Get Shit Done. um
Jamie: Oh, yeah.
Keith: But it's it's under Open GSD now, I guess.
Jamie: Cheers. Yeah, yeah. I've heard of that.
Keith: But yeah, it's pretty handy. It's just a bunch of skills and like meta prompting basically to do that. But it it has different modes in it where it operates on these things.
Jamie: Yeah.
Keith: And it was pretty cool because it would do like market research and stuff for you too. Like you would tell it a thing and it would be like, Oh, this is the feature you're adding. And then it like, it would go off and do a bunch of research, come back.
Jamie: Oh.
Keith: And then it would present like questions to you and like, be like, Oh, what do you think about this? And you could be like this, this, or this, or write some other things.
Jamie: That's yeah.
Keith: that was very cool. Yeah.
Jamie: Yeah, that's what the G stack, you know.
Keith: use that actually, uh, I use that for, uh, not yet released, but, uh, I went back to, um, Juniper's Lab, my old trading card thing that I never released.
Jamie: oh
Keith: It's now called decklist.net.
Jamie: Whoa, yeah, yeah, no, I remember that.
Keith: But yeah, so it's all like, I actually just have somebody working on a logo for me. And then I will probably...
Jamie: Dude, Pokemon, or like by someone you mean like an agent?
Keith: No, like a dude.
Jamie: Okay.
Keith: Like an actual dude. yeah
Jamie: Yeah, that's good.
Keith: But yeah, it's pretty much just as an app.
Jamie: um Dude, Pokemon is like huge again.
Keith: It's in, okay, sidebar Keith's Pokemon rant.
Jamie: oh
Jamie: Okay.
Keith: I'm so annoyed by this because it's not even people who play the game and it just makes it impossible for children to find Pokemon cards.
Jamie: Oh boy.
Keith: Like that's the, it's just a bunch of middle-aged white guys scalping the stuff and then sitting on tons and tons of product in their basement that they bought for MSRP that nobody wants to pay more than MSRP for.
Jamie: Oh, it's a, yeah.
Keith: So like it all sells out and then it's just the same people buying it constantly or like getting into fistfights over it. And it's so annoying. Like I wanted to just get some packs of cards for my kids as like rewards for like doing things because they they like Pokemon cards or whatever.
Jamie: Oh, you're gonna have to fight somebody.
Keith: And yeah, I'm like, well, for one, I found out that Dick's Sporting Goods of all places has Pokemon cards.
Jamie: Of course, of course.
Keith: yeah And I was like, well, guess what? Scalpers aren't going to Dick's. um So they're not going to be caught anywhere near a sporting good face. But yeah, so that was one where I was like, oh, cool.
Keith: I guess this is my my inside scoop here. But yeah, that was like, ah drives me nuts.
Jamie: But if they are there, they're probably going to be like in good shape.
Keith: That's true. It's only the jacked scalpers that are like, I also play softball lot and I left, you know?
Jamie: And yeah, it's the jackscoppers that you have to fight. Yeah. yeah well
Keith: But yeah. Oh God. the
Jamie: Crossfit.
Keith: So anywho, I spun that back up. I wrote it. I didn't write any of the lines of code in that app.
Jamie: Yep.
Keith: I spun it up from scratch after watching the Rails World thing from DHH about...
Jamie: Yep.
Keith: um him wanting to go back to Rails being like the optimal one-man framework. And I was like, I'm just gonna try the out-of-the-box Rails experience again and see what it's like. So it's using like, uses Kamal for deployments, it uses literally uses SQLite in production.
Keith: it you know It's just, it costs nothing to run, it's super fast.
Jamie: where Where you have it hosted up.
Keith: It's on like a DigitalOcean like $10 box or whatever.
Jamie: Yeah.
Keith: But yeah, it's super quick.
Jamie: Are
Keith: Like it everything works pretty well. it's It's another one of those things that I think would be interesting. Like once we get to the point where, mean, this is weird Rails talk that you don't care about but anymore.
Keith: But ah like if if Falcon picks up you know some more speed, which is like an async web server,
Jamie: you...
Keith: in Ruby that would be like a replacement for Puma and it works with Rails like that could be a dramatic change where you could run Rails on even fewer resources and have the performance get boosted quite a bit. There's a lot of things where it's just like you could just run a Rails app that does like you know, just makes beer money or whatever for you.
Keith: Or in my case, the card money.
Jamie: Yeah, yeah,
Keith: um And yeah, just be good to go. And I'm like, that's sweet. Like a $15, it pays for itself. when If I get five subscribers, you know, that would be awesome.
Jamie: yeah. i spun up Yeah, I spun up like three or four apps already. one of them that I released was Mixtape,
Jamie: It's probably not going to do Spotify's fucking terms for apps that they just updated. It's crazy. So that's probably just going to be a for me thing. But I have a tape.mx, which which is basically what I've been doing since like 2018, where I like add songs I discovered to a playlist for the month.
Jamie: i don't know if you knew that I do that, but every month I release a playlist of songs that I discovered that month and I've been doing it since January of 2018.
Keith: No.
Keith: Uh-huh.
Jamie: Yeah. um
Keith: That's fun.
Jamie: yeah And so I was like, bet you.
Keith: Yeah, this website is tight, too.
Jamie: Yeah, thanks. Thanks to Claude.design. No, I gave it the ideas. But I do not know GSAP like that.
Keith: Mm-hmm.
Jamie: um
Jamie: So yeah, so I had that working in like two days. um But yeah, it's just a little app that when you favorite a song, it puts it into like this month's playlist. And then and know you can discover things. and
Keith: Nice.
Jamie: But Spotify's, they're like new app stuff. You can only have up to five individuals on your app until you reach 250,000 MRR and then and then they'll grant you your app.
Jamie: right
Keith: How do you get to 200?
Jamie: like Exactly. Thank you.
Keith: I mean, obviously you have to charge them $50,000 a month a piece.
Jamie: Okay.
Keith: That's the only solution. i don't know.
Jamie: It is the dumbest. It's like, oh, we only care about big people that already have platforms or something. I don't know.
Keith: Even then, they couldn't... like Okay, so hold on. So the the ticket is you have to have one umbrella business that has lots and lots of money that it makes. And then you can create an app out of nowhere and and instantly have a thing.
Jamie: Yes.
Jamie: Yes.
Keith: you just You couldn't spin off a new business.
Jamie: Yeah.
Keith: This whole thing is like it's a Spotify-related
Jamie: Right. or you could Or you could sell your house. And then pretend like that MRR for like that that month.
Keith: You pay yourself $250,000 one time, but you're like, this is my um ah um MRR.
Jamie: Yeah.
Keith: Trust me.
Jamie: Yeah. Yeah. I paid for my own app. um Yeah.
Keith: Nice.
Jamie: So dumb. So yeah.
Keith: Yeah, that is really stupid.
Jamie: But it's been fun, dude, to just spin up apps that you've been wanting to do forever, you know, like a chronic or whatever. That'd be so easy to do now.
Keith: It would be quite easy. I mean, everything is possible now, which I think is the fun part.
Jamie: Yeah. Yes.
Keith: like It does feel like you could things you couldn't do before you could manage to do quite quickly now.
Jamie: yes
Keith: Because you can you could always describe the idea. that you were trying to do, but sometimes it would take you too long to actually write the code for it.
Jamie: excellent
Keith: Cause it like, sometimes it's just the step, like you know exactly what to do.
Jamie: Exactly.
Keith: It's just complicated. It's gonna take a long time.
Jamie: Yep.
Keith: And so yeah, having that intern that can type it a million words per minute is really handy sometimes. And you can just be like, oh cool, this works. I mean, you do get to the point where you're like you have no idea how anything in your application probably does actually work.
Keith: But you know, in theory, how it works. And you know that you'll just be able to throw that bug at AI and AI will probably figure it out and write a couple more bugs while it's doing it. And you're still in the job.
Keith: So, yeah, I was i was actually curious how this would affect ah like you as a freelancer, you know, as a thing, right?
Jamie: Um,
Keith: Like. Does client work, did client work go up? Like, are you able to more quickly just burn through clients or are more people trying to do it themselves? I think it's over how many people think they're doing it themselves.
Jamie: Yeah, um funny you should say that, because I did just lose my client that I had for the last three years. um But because they just let a bunch of people go because because of AI, honestly.
Jamie: um Also, they're horrible business practice, but I'm i I agree with you. i think the the whole like, oh, everybody can just make an app now is overhyped and it will always be overhyped. It's the same thing with everybody can be an accountant now because you have Excel, right?
Keith: Mm-hmm.
Jamie: um But people aren't becoming accountants because they have Excel. um And so, yeah, I think honestly, what I would like to do is actually just sell my tokens.
Jamie: I would like to sell Jamie tokens. I'd like to sell jokens. um And then ah just have it work like like regular tokens work, but with a little bit of an upcharge.
Jamie: Because I honestly can like handle, like right now I have a small client and um what would have taken me probably three months before of like where I'm at took me a week of, you know, and it's like, no, no, I don't because that's, that's the thing.
Keith: you Do you charge three months value though?
Jamie: That's where, that's where I'm going to be impacted. I feel like, I think that's where all of us are going to be impacted is the price that we can charge for things, um, or the money that we can make from this because, um,
Jamie: Maybe not us, but I'm saying like independents and freelancers and things like that, like they can just do it so fast. Like I wouldn't charge people more because i can you know I can do it in Vim and I don't have to use punch cards. you know like ah So I feel like it's the same.
Jamie: it's just good
Keith: Yeah, but I feel like you have to charge based on the body of work, not the hours it took to do it in a big way, right? Like, don't you just move more towards like flat pricing?
Jamie: Yeah, yeah, that's the other thing. um That's the other way, actually, to make make a lot more money than and and charging by actual time.
Keith: Which I know there are plenty of people who would quote projects that way anyway.
Jamie: But I don't...
Keith: But I mean, there's always the chance that you end up underwater because you quoted it whatever cost and then it balloons into something big. but
Jamie: Mm-hmm.
Keith: But I feel like now youre like you are so insulated from that in a big way. Like obviously you could do a thing where you're like, I just burned a ton of AI time on this, but like well likely not enough if you priced it accordingly.
Keith: And yeah, just sure, could somebody go and build the thing for way cheaper if they did it themselves?
Jamie: Yeah.
Keith: Yeah, but they're going to struggle through the things because they don't have the context that you have in order to get to the solution.
Jamie: It's going to be the, yeah, exactly.
Keith: Like, it's going to not be deployable.
Jamie: It's going to be the same.
Keith: It's going to not, you know, like, yeah, it's going to have tons of problems.
Jamie: it's going to break. Why does it, why, why is it broken?
Keith: Yeah.
Jamie: I don't know.
Keith: And then they, yeah, they don't know how to do that stuff.
Jamie: And then it takes them, you know, three hours to fix that.
Keith: Yeah.
Jamie: Yeah. No, it's the same. It's the same thing until people now are, you know, before when they wanted to hire, you know, well, this person will do it for 70 grand. I was like, all right, do it.
Jamie: Have do it for 70 grand.
Keith: uh-huh.
Jamie: And then they have eventually come back in three months and say, you've already spent 70 grand. So I think it's like, yeah, I don't think that's going to change that much. Like,
Jamie: in terms of
Jamie: the amount of consulting work. I just think how it's being charged is going to change. Also, like AI is so freaking cheap right now, dude.
Jamie: It's so cheap.
Keith: It's not going to stay that way.
Jamie: Exactly. it's not going It's going to be the Uber model.
Keith: I mean, I mean, Microsoft just cut off its cloud code licenses because they were spending too much money or Uber spent its entire AI license for 2026 in the first four months of the year. Like it's, did you not read that article?
Jamie: Really? Wow.
Keith: Yeah, there's an article.
Jamie: No.
Keith: Like they, like they blew their entire 2026 AI budget in the first four months because engineers using cloud code is a dramatic thing. Like, I fully expect Shopify to you know to have some put some real like a death grip on tokens here quite quite quickly, I imagine, because it's it does get expensive when you're doing it at scale, um for sure.
Jamie: Yeah, I could imagine. It's not going to
Keith: But yeah, I think for the for like small project stuff, and like this is where I'm like, ugh. There's this weird thing where you go, well, small projects have no place in the world because everybody will just build their own, right?
Keith: Like there is that as an idea.
Jamie: happen.
Keith: And I'm like, that's not gonna happen. Like people think that, I mean, although I could be wrong, right?
Jamie: it's not gonna happen I agree.
Keith: Lovable, Replit, like all these things exist, but I'm curious and they have like crazy um MRR too, like Bolt or whatever.
Jamie: was
Keith: Like there's there's so many of them.
Jamie: Yeah, there's so many of those
Keith: And they basically hilariously are just an agent where it just takes your thing, massages it into probably a slightly better prompt and goes off and does the thing and has some interactions and
Jamie: those. we we built one We built one with my last client that was a yeah like a a shop, a storefront builder. And we built that with one one and a half purs people in like a month.
Keith: Yeah.
Jamie: So yeah, there's million those.
Keith: And so it's like, it exists as a thing. And some of those things will work, you know, but largely you're going to run into things where, well, and especially when I was doing lovable, this was like a year plus ago, I was playing around with it just to see how good it was, but it would be like, cool.
Keith: I would get to the point where like it had something kind of cool. And then I'd be all right, let's add this next thing. And it would just break everything, everything that I did previously touched.
Jamie: Everything.
Keith: And I'm like, Jesus, what, like what happened?
Jamie: yeah
Keith: Like, obviously models have gotten better since then,
Jamie: Yeah.
Keith: but they ebb and flow, right? Like Claude, yeah, sorry. Opus 4.7 is not any good.
Jamie: No, no, I actually went back to Sonnet for a lot of things.
Keith: And
Keith: yeah, like 4.6 is better. um Like they just need to do a git revert. ah codex
Jamie: but
Keith: Codex 5.5 is actually better. Like it it cooks right now.
Jamie: Yeah, I've heard that. i haven't Yeah, I have to play with that. i haven't not play
Keith: Yeah, it's one of the beautiful parts about Pi is I can just, i have one harness that I can just be like, swap out the model and try a different thing or like Gemini Flash three, you know, whatever the newest one is pretty good too.
Jamie: on
Keith: If you just need like fast, cheap stuff, but yeah, it's just wild.
Jamie: does Does Shopify have all of them?
Keith: Like, yeah.
Jamie: Like, yeah.
Keith: I can swap to any model I want. We have a singular AI proxy that routes it to all the other things. So like all of our tools just point at one thing and I can just change the model string.
Keith: So it's one unified ah API all the time.
Jamie: That's dope.
Jamie: Yeah. Dude, there's so much more to talk about. We should keep talking.
Keith: Yeah, we should, ah we're gonna have to schedule another one of these, man.
Jamie: Yeah, that's not three three years.
Keith: we I kind of forgot that the last episode we recorded was when you you were like, I think we need to stop recording. Or you said, you specifically said, i think we should pause recording.
Keith: And I actually paused the recording in the recording because I thought you had to like go to the bathroom or something.
Jamie: Yeah, yeah.
Keith: It was so funny to listen back to be like, oh, I was completely blindsided by this. That's so funny.
Jamie: Speaking of AI, dude, we can actually like just spin up a proper site and not have to pay for it. for it You know?
Keith: Oh, that's true. Yeah.
Jamie: You know, um, we should maybe, maybe we could do that like live. Fuck it.
Keith: Well, see, so now we have a problem.
Jamie: We'll do it live.
Keith: We have forever.
Jamie: Oh, okay.
Keith: We've had the, like a great podcast name.
Jamie: Mm-hmm.
Keith: Unfortunately, the primogen stole our podcast name.
Jamie: Who?
Keith: You don't know the primogen. Um, he's a, he's a big, like programming entertainer, essentially.
Jamie: No.
Keith: He's a big YouTuber, but he, and a couple of his YouTuber, but a like huge YouTuber, they, so they started a podcast called the standup.
Jamie: Really?
Keith: I should she not.
Jamie: they didn't come They didn't come to me and ask if they could get the domain.
Keith: Well, the domain doesn't point anywhere right now, does it?
Jamie: no but it's still taken.
Keith: Um, but yeah, so that was one of those things I was like, son of a gun. But then I was also thinking about it and I go, well, we never really talked about this to anybody. It's not like I was ever trying to find, you know, listeners for this thing.
Jamie: yeah
Keith: It was like, no, this is.
Jamie: also if you search if you search for stand up in podcast apple choice it comes back with comedy for the first 30.
Keith: Oh, true. Yeah. Didn't think about that originally, but
Jamie: yeah yeah well
Keith: Yeah, no, I, uh, this has been nice. It's been good just to chat. We legitimately have barely talked in that three year span of time. Um, so, um,
Jamie: I know, I know. are You still live in the United States?
Keith: i I do. I mean, I was in Switzerland for that one month, but ah via besides that, which total side topic, me and the Swiss, we're not, we do not mesh.
Jamie: and Oh, that's true. That's true.
Jamie: No?
Keith: um
Jamie: Because the cheese?
Keith: No. Yeah. The, well, I'm dairy free, dude. ah
Jamie: Well, got holes.
Keith: but No. So the, we went over there for a month. Was that last year? Two years ago? I don't even remember. But yeah, I was doing my flexible stuff with work or whatever. And so we were over there in an Airbnb, because you get a discount on Airbnbs if you book them for 30 um
Jamie: Nice. nice
Keith: i Which I didn't know, but a buddy mine who travels a lot kind of told us. And so we went over there. And then basically, like we got there, we were in a grocery store, an old lady growled at my daughter for like walking next to her.
Keith: And I was like, what the fuck was that? And that just kind of continued. like Dirty looks, ah Swiss kids trying to knock my kids off the playground equipment for no reason, like without even talking to them.
Keith: And come to find out, it's just like a cultural thing where they're just dicks. um But they it's it's actually that they like they have a big like inner community.
Jamie: They learned it. They learned it from the French.
Keith: the thing they just don't like outsiders and they indoctrinate their children at a young age to not like outsiders.
Jamie: Yeah.
Keith: And so they literally like I watched a parent watch their kid bully my kid and do nothing about it.
Jamie: That's good.
Keith: And my daughter is like so nice.
Jamie: Except a laugh.
Keith: Like she's incredibly nice. she Just likes to play with people. Like doesn't, I mean, yeah, she's incredibly friendly, but yeah, she was on like a rope swing kind of, or not like a rope ladder kind of thing on, and basically a rope course, uh, playground area.
Jamie: Mm-hmm.
Keith: And two Swiss kids like went on either side and just started shaking it violently while she's on this thing. Didn't say a word to her, just tried to like, and so I had to go over there and, you know,
Keith: and put the fear of God and some children, but I was like, I'm about to end up at Swiss jail.
Jamie: Yeah, but yeah. Title.
Keith: Um, but, uh, yeah, it was, it was rough, but yeah, I really ended up every time I would run into somebody who was like pleasant, I would turn to Kendra and be like, they're not Swiss.
Jamie: Yeah.
Jamie: title
Keith: And then I've come to find out they're like an expat and who like just moved to Switzerland, but there's like Irish guy or like, there's a dude from Italy.
Jamie: yeah
Keith: yeah. Yeah.
Jamie: why are Why are they called expats when they're white people, but immigrants if they're not?
Keith: That's a good question.
Keith: ah
Jamie: I think I know the answer, but a
Keith: That is a good question. ah But yeah, so that was my fun side there.
Jamie: that sucks, dude. All right. Don't go to so don't go to Switzerland, people. Just joking, just joking.
Keith: All right. Hold on. We need some, I mean, this is already a long episode.
Jamie: Yeah.
Keith: doesn't matter. But so but there was always like a little hobby talk at the end of this because I would always turn it into some nerd thing. ah But
Jamie: Yeah.
Keith: So we talked pinball the last time I think, right? Had I done pinball last time?
Jamie: We did?
Keith: I bought a pinball machine.
Jamie: Wait. Yeah, you bought a pinball machine.
Keith: Yeah.
Jamie: Yeah, and you've been doing...
Keith: But that was, that was like last, in the last six months I bought a machine.
Jamie: Yeah, that was...
Keith: um
Jamie: No.
Keith: Yeah. I bought it in December.
Jamie: I thought had one in your...
Keith: It's in my basement.
Jamie: I thought you had one in your house, then you... Was that you? Maybe that was somebody else.
Keith: No, I have a really big pinball arcade nearby, but I have a Jaws pinball machine in my basement now, which was sick.
Jamie: You should probably get close to my...
Keith: That was cool.
Jamie: Nice, dude.
Keith: Yeah, it's pretty fun.
Jamie: like the Like an OG one or like is it like a new?
Keith: No, there's not an OG one. it's It's like a new one. has like an LCD screen, like plays movie clips and stuff.
Jamie: Yeah, yeah. yeah
Keith: Super good game. But a little over 1 billion, I think.
Jamie: Yeah, yeah. What's your high score on it?
Keith: Uh, yeah. so it's not like a great score. I mean, it's like a top 10% of scores sort of thing.
Jamie: that's a That's awesome.
Keith: Uh, but,
Jamie: how How long did that take you? Like how many?
Keith: uh, I, my goal was to break a billion within the first month of having it.
Jamie: how works
Keith: And I think I had like broke a billion on like February 1st. So technically I didn't make it. Um, and I was a little upset, but yeah. So did that.
Jamie: Wow.
Keith: I recently started rollerblading, yeah.
Jamie: Dude, they still make rollerblades?
Keith: Dude, rollerblading is sick. Yeah.
Jamie: No, it's not, man.
Jamie: i just, I used to have rollerblades.
Keith: No, yeah, yeah.
Jamie: Um,
Keith: So they still make inline skates, but yeah, it's fun. My daughter got some and like she was doing it and i was like, is it weird that I'm jealous of my seven year old right now? And then I was like, they they still make these for adults, Because I just had remembered enjoying it as a child.
Jamie: What?
Keith: And then so I'm like, okay, cool. They do make this. And I like talked to some other people. Basically, i had to get affirmation from other adults that, yeah, dude, I wish I could roller skate kind of thing. Fuck yeah, I'm going to do this. And so I bought some and yeah, ate it the first day. That was pretty good. Like I, or actually it wasn't even the first day. First day i was like, man, look at this comes back faster than riding a bike.
Keith: Like when I wrote, wrote like getting on a bike after five years off, I was like wobbly as crap, but putting blades on, I was good to go.
Jamie: Mm-hmm. hour
Keith: Like I was able to skate backwards, getting forwards, spinning around.
Jamie: how bad How bad did your feet hurt though?
Keith: My ankles were a little sore. Yeah.
Jamie: oh Last time I tried to rollerblade, which is years ago.
Keith: yeah
Jamie: Yeah, I just remember my freaking feet like hurting for like days afterwards. It's probably because my feet no longer fit in my damn skates.
Keith: Yeah, I ah um went to like a bike trail and like was rollerblading for like exercise, did a couple of miles the other day.
Jamie: anybody push you over?
Keith: Realized I'd, how what?
Jamie: Anybody try to like push you over?
Keith: No, no.
Jamie: go, get out of here, nerd.
Keith: Just shoving.
Jamie: ah I'm just joking.
Keith: Yeah, but it was a good time. but I ended up talking Kendra into buying some and bought some for Asher.
Jamie: That's awesome.
Keith: So we have a whole family of people with inline skates now, but it's so funny.
Jamie: Dude, you guys should like, do you guys ever hold on to each other? Like with a, like a, oh, oh.
Keith: Well, so Kendra's arrives today. So we haven't like actually gotten the chance yet. But yeah, it should be interesting. But it's it's been a ah fun little thing the last.
Jamie: That's going to be the cover art for this episode is the family. You're a family in skates.
Keith: Yeah, my family in skates.
Jamie: So if you can send me, if you can send me a picture. um Yeah.
Keith: AI can just make one. It's fine.
Jamie: yeah i I broke my leg on my bike and cracked a rib on my bike.
Keith: What did you do in the same occurrence?
Jamie: So it's been No, no, no.
Keith: No, never could have been the same thing.
Jamie: Braking of, yeah, the braking of the leg was not my fault. I almost got hit by a car. um They were on their phone. They were coming into the bike lane. so I knew it over. I thought it was grass, but it was mud. And so I flipped over the handlebars.
Jamie: Bike landed on my leg, which is like 130 pound bike.
Keith: Jeez.
Jamie: I was like, I got up. i was like, oh, that kind of hurt. I woke up the next day. I'm like, oh, that really hurts. I'm like still working. I'm like, told my client, I go, I might have to go to hospital.
Jamie: It's like, I may have broke my leg. And I was like, yeah, it broke.
Keith: So was like your lower leg, like your shin the area or your, okay.
Jamie: Yeah, it was my, it was my, whatever that is, fibula.
Keith: I was like, you didn't snap your like femur. Did you like go a bike?
Jamie: No, I thought I was going to need a surgery.
Keith: No.
Jamie: Well, when I went in the ER, was like, dude, I'm going to surgery. I know it. And they're like, no, you don't need surgery. I was like, yeah.
Keith: Nice.
Jamie: But anyway, yeah, so I'm still riding bikes. Can't keep me off without a helmet.
Keith: OK.
Jamie: so that's cool. um Yeah, it's about hey anything I've been doing.
Keith: Nice. OK, so we can make the decision if we want to record any more of this or if this is the AI special and we don't want to come back. But that
Jamie: I would, yeah, we we'll do that. We'll do that off camera.
Keith: Yeah, definitely. But all right. It's been great to talk to you again. And I'll talk to you again in an indefinite amount of time, which is how we ended this the last time.
Jamie: Yes, in some amount of time, yes.
Keith: All right. Later, dude.
Jamie: iss
