--- title: "S01 E17 - Back to Basics & Off the Cloud: 3D-Printed Moles, a DIY Pocket, and Ditching Third Parties" shortTitle: "Back to Basics & Off the Cloud: 3D-Printed Moles, a DIY Pocket, and Ditching Third Parties" season: 1 episode: 17 slug: "back-to-basics-off-the-cloud-3d-printed-moles-a-diy-pocket-and-ditching-third-parties" publishedAt: "2026-08-28T13:00:00.000Z" duration: "00:53:39" canonicalUrl: "https://bloodsweatandtokens.xyz/episodes/back-to-basics-off-the-cloud-3d-printed-moles-a-diy-pocket-and-ditching-third-parties/" audioUrl: "https://api.riverside.com/hosting-analytics/media/9587234fa95ebcde49ed7ed22b47517a8881834ed629bc519b32e90f510a110c/eyJlcGlzb2RlSWQiOiI5ODg5ZGYzMi1mNTE4LTQ4M2MtOWE3NC0xZGM2Njc2NDdkODUiLCJwb2RjYXN0SWQiOiI3ZmE0NDhkNi0xMzhiLTRkOTktOTFhMi1lYzBkNGExMWI2MDYiLCJhY2NvdW50SWQiOiI2OWE1ZmJhOTRkZmVkYWY3NDViZTk1YTMiLCJwYXRoIjoibWVkaWEvY2xpcHMvNmE5MDU2NDM3NTc0YmE5MWNhMTVmMTI5L3RheWxvci1tYWNkb25hbGRzLXN0dWRpby00VE9qRS1jb21wb3Nlci5tcDMifQ==.mp3" youtubeUrl: "https://www.youtube.com/watch?v=h9RpWePaydE" --- # Back to Basics & Off the Cloud: 3D-Printed Moles, a DIY Pocket, and Ditching Third Parties ## Summary Is AI actually pushing us back to basics? ## Show Notes Is AI actually pushing us *back* to basics? This week Taylor demos two builds that argue yes: a 3D-printed "don't whack the lemon" exhibit for Crimson's Emotional Endurance Carnival (servos, heat-set inserts, and an Arduino Mega — spec'd almost entirely by Claude for pennies on the dollar), and a homemade Pocket clone — a Raspberry Pi Zero + microphone rig that records, rsyncs over Tailscale, transcribes with local Whisper, and pipes everything into a Chroma RAG database without a single byte leaving his network. Sean — now "Chauncey" to every speech-to-text engine in his life — brings the voice-transcription infrastructure he's been building and the old-school backup renaissance (Arq + Backblaze are back, baby). Then the pair wrestle with the bigger picture: Bill Gates' warning that AI will be "the greatest equalizer ever invented or the worst source of injustice," notification-induced variable-stimulus madness, cloud bills versus idle local compute, and what Stripe's $7B OpenRouter acquisition says about the death of the one-lane developer tool. Plus: NVIDIA's record earnings, 2M more AWS GPUs, and OpenAI testing ads in ChatGPT ("the beginning of the enshittification of AI"). **Key Discussion Points** - **AI-assisted physical making** – Taylor goes from zero CAD/electronics experience to a working lemon-popping carnival game in a month, with Claude generating the bill of materials - **The $5 on-air box** – a 3D-printed recording light (~$1 of filament, 20 minutes) vs. a $20 Amazon equivalent — customization over commodity - **DIY local-first audio RAG** – Raspberry Pi → rsync → Tailscale → local Whisper → Chroma, with nothing ever leaving the network - **Back to basics** – rsync, Time Machine, and local compute as the pendulum swings away from runaway cloud costs and privacy exposure - **Bill Gates' AI essay** – "the greatest equalizer ever invented or the worst source of injustice" - **Notification overload** – Duolingo pings getting the same stimulus weight as a family emergency - **The one-lane era ends** – Stripe buys OpenRouter for $7B; the 2020 "do one thing well" dev-tools thesis is dead as everyone absorbs everyone else ## Transcript Sean C Davis: Probably putting an offer in t a today. Again. Round round three. Third house. Taylor MacDonald: Hey man, they say third times the charm, that's my understanding. Anyway. Sean C Davis: That's true. It's true. Taylor MacDonald: Welcome back to Blood, Sweat and Tokens, the podcast where Sean Davis and I are talking about product development in the age of AI. We're talking about tooling, we're talking about philosophy, the impact of humanity, the impact of the environment, all that kind of stuff. So we're we're no no no bounds left uncovered. How are you doing today, Sean? Sean C Davis: I'm good. Can I tell you something, Taylor? talking about AI. Taylor MacDonald: yeah, please. It is a podcast. Sean C Davis: I I so you you introduced me as Sean Davis and I was like, I was gonna be super snarky and I was like, my name is Sean C. Davis. But since we're talking about AI, I have a short story, which is and and one of the things I want to show today is about voice to text. And so I'm doing this all the time, recording videos, yada yada yada. And then taking transcripts out of it. And it is constantly I think I I think I kinda like slur the Sean and the C. And so more often than not, it's chauncy. Chauntsy. Taylor MacDonald: Shaughnessy. Shaughnessy. Chauncey. Sean C Davis: And so I've started Taylor MacDonald: so AI is like transcribing that back to you? Like, hey Chauncey, has that happened? Sean C Davis: It thinks I'm Chaun Chauncey Davis. And here's the the the benefit of that though is that at Netlify there are three different Chan's all spelled the same and I'm kind of leaning into this one and I think I'm just gonna change my name at Netlify to be Chauncey. Taylor MacDonald: Okay, how many can you quantify the number of Seans on staff at at Netlify right now? Like, can you go to Slack and search the company directory and count the number of Seans? Because you have Seans that have different I mean, you've got every spelling under the book. I've probably met five different Seans just through my interactions over there. Sean C Davis: yeah, if you include all the spellings. Taylor MacDonald: Yeah, it's crazy. You know, I wonder what the Venn overlay of people named Sean and technology is. Sean C Davis: I yeah, I don't I don't I don't know. I don't know. we probably have a handful though. Yeah, I dn yeah, yeah. Taylor MacDonald: Well, all right. Well, so today so today I know that we're getting close to winding down the concept. Is this is episode 17? Wow, that's nuts, man. Yeah. Sean C Davis: I don't even know. Is that right? Taylor MacDonald: Yeah. This is like a network TV show. Should we should go to like 22 episodes before we call it quits, you know? Like, I don't know. Anyway, Sean C Davis: That's right. Yeah. Yeah. Taylor MacDonald: I'm just kidding. We're not gonna call it quits, but we are gonna hit the brakes on the concept of this product that Sean has been developing because I think it's answered some questions, it's asked some different questions or or caused us to ask some different questions. but regardless. we're gonna kind of spend a little bit today doing demo. Unfortunately I'm not sure we're gonna be doing demo on the product. is has anything happened on the product, sh Sean in the last well, it's been like three weeks since we demoed anything. You were out for a couple of weeks and then last week we were just catching up on the latest gossip. Sean C Davis: Yeah, what we were talking about before winding it down was that we were going to we were gonna show voice transcription within the application. And because we started to think, okay, this isn't right, I wanted to do that anyways, but I what I decided to do was to build it in a more agnostic way and to but to get that experience and try to set up some of that infrastructure that also plays into some of the other projects. that I have that I think are a little more suited for long term value. So I have some voice demo to show, but it's not gonna be within the the UI of the product. Taylor MacDonald: Got it. Okay. that's cool. And I have some demos to share as well today. So as you know, I've been working pretty heavily on a few things. one of those things is a well it all kind of evolves around this concept of 3D printing. Okay, so my wife started a company last year with a a business partner called Sprouting Change, and Sprouting Change is kind of this organization that's designed to encourage coping skills or you know, educate their clientele, users and so forth on how to negotiate some of the trials and tribulations of life. So whether it's the economy, the coming gas crisis, I don't know if you've heard the United States only has about forty days worth of gas available in reserve ahead of us, which is the l Sean C Davis: What? I did not read that. Taylor MacDonald: yeah, yeah. It's the lowest the reserves have ever been in the history of strategic reserves or something. I don't know. It's it's bananas. So fill up your car, bro, because it's about to get rocky. Yeah. anyway, whether it's that, Sean C Davis: Okay. Glad I bought that bicycle. Taylor MacDonald: whether it's that, whether it's your kid, you know, barfing on your new car seat or something like I don't know. Like at the end of the day, all the stuff in life that gets at you, it is, as Brene Brown would say, it's it's robbing you of the marbles in your marble jar, right? And eventually you hit the bottom of the jar and you kind of break down or you lash out, you get angry, you get short, you get frustrated. So the the concept here and what my wife has been working on is this immersive physical experience that she's calling the emotional endurance carnival. It's got about 18 different exhibits in it, all of which have varying degrees of uniqueness, complexity, and fun. The idea is every one of the exhibits or every one of the stations throughout this experience Are you going to teach you and reinforce some sort of like critical skill and negotiating challenge and tribulation in your life, you know, giving you emotional endurance, the ability to put a little bit of space between the stimulus and the reaction. And it's really interesting, it's super creative. They come up with all sorts of stuff. They've been doing a podcast for over the last year where they're talking about the emotional endurance movement and they're generating some real interest and activity. And it's super fun. So one of the things that she came to me about a month ago and said, hey. I am interested in building a whack-a-mole exhibit for the carnival. Well, that concept has evolved quite a bit. It's no longer whack-a-mole. Now it is the idea is not to whack the mole, the idea is to avoid whacking the mole, right? Because if the mole represents stimulus in your life, like something pops up unexpectedly and you can smack it. Like that's the exact opposite. That's that is the the antithesis of of of being emotionally endurant, I think. So anyway, so the idea is to just observe the lemons, or sorry, the moles as they pop up in your life, but instead of moles, we're making lemons, right? So it's more like don't Sean C Davis: Love that. Taylor MacDonald: whack a lemon game. Anyway, it's kind of weird. So there's a whole like narrative, storyline, and companion app that goes along with this. But I was just asked to build the physical infrastructure, the actual box that has these moles popping up and out. And so what I did was I I bought a 3D printer as as I've mentioned a while ago, and I spent about a month putting it all together. And I finally completed the build last night. Or at least the the actual working mechanicals on it. Would you like to see the latest Sean C Davis: Absolutely. I'm Taylor MacDonald: progress? Let me see if I can get it off my phone and onto my computer. That's probably what I should. Sean C Davis: I'm I'm fascinated with the well well two thing. I'm just one, I've wanted a 3D printer for a long time just to screw around with and and solve random problems. But I also I think that you're hitting on something that I've been fascinated with and really interested in over the last few months, which is kind of you know, how do we use AI in the real world, get us out off of the off the screen a little bit and and that's kind of cool. Yeah. Taylor MacDonald: So I think that's really interesting and and it relates to a topic that I want to talk about after we get through the demos here. But Bill Gates wrote an essay. It was like a six thousand word essay yesterday came out. Did you read it? Sean C Davis: No. Taylor MacDonald: Okay, well it's you can imagine it's you know, pretty magnanimous. It's it's it's pretty heavy. One of the things he talks about is like absent of a plan here, we are hurtling towards. A major crisis in humanity, in that most of the experts, the chief executives amongst these AI providers, they don't want to say that quiet part out loud. But Bill Gates, like that's what he's effectively saying is like we have a major problem and we need to start vigilantly figuring out how to slow this down today. I mean, I'm paraphrasing or whatever. but one of the things that he said towards the end of the article, and I should find the quote because I don't want to mangle it. but this resonated with me. He's like it's either AI can either be the greatest catalyst for either the greatest equalizer or something, or it can be it's gonna set us up for soup certain doom. you know, which is I I think not surprising. Sean C Davis: AI will either be the greatest equalizer ever invented or the worst source of injustice. Taylor MacDonald: That's the one. Yep. So Sean C Davis: Yeah. Taylor MacDonald: I and I think that's interesting, right? Because as you talk about bringing stuff up out of the computer, bringing it into the physical world, in a lot of ways, and some of the arguments I think that I have made here is like if you look at this as with an optimistic view, maybe what AI is doing is it's forcing us to get back in front of each other and have a human conversation, right? Like over the years, all the notifications, all the the static, all the noise has has influenced us in a way. I don't know if we were talking about it on the podcast, but I recently had a conversation with somebody about that like phantom vibration effect or the the inconsistent stimulus. You know what I mean? I've was it Skinner, one of the psychologists back in the 70s or 80s or whatever, ran an experiment you know, where you press a button and you either get like a food pellet. It was for a dog or a or a monkey or a rat or something, they get a food pellet or they Sean C Davis: yeah, yeah, yeah. I know this experiment. Yep. Mm-hmm. Taylor MacDonald: get shocked, right? But it was an inconsistent stimulus. So you either press the button and risk getting shocked, or it might work out and you might actually get, you know, something like a positive stimulus, like food. But then it creates this inconsistency where you're like, well, if I hit the button, I know I've got this risk of something negative towards me, but I have the the potential for something positive. That is what these phones are doing, right? So like you get a notification, let's say you're walking the dog, you feel it buzz in your pocket. Maybe you're in the middle of a conversation with your wife or your friend or something. Do you stop and look at your phone? Sometimes I do. So I try to be like mindful, but then it's annoying because I can feel it in my pocket and then it'll buzz again in like a couple of minutes. And so it's just like becomes this Sean C Davis: Yeah. Taylor MacDonald: nagging, like, fuck, I need to check that thing. I need to like lean in and see what that is. Okay, so you pull your phone out and it's like Duolingo nudged you to like do something. Not helpful. Not Sean C Davis: Right. Yeah. Taylor MacDonald: helpful, Duolingo. And yet the stimulus attached to that notification is equal. to the stimulus of my wife saying, Holy shit, I can't find my daughter after school. Like that it gets the same Sean C Davis: Right. Yeah. Taylor MacDonald: weight in my life. And there is no way, at least I haven't figured out a way, to like distill away the things I don't care about. Because the only re I mean, all like you care about like one of those notifications at one moment in time, but then it lives on in perpetuity 'cause you never go back and like clean that up or There's no easy way. You know, how about this? You get a notification that's important. You look at it and then you go to the notification center to find it, and it's not there. And then now you're like, my gosh, I know there's this important, Sean C Davis: yeah. Yeah. Taylor MacDonald: like, valuable thing. All of these, these ups and downs, like all of this inconsistent behavior, I think is driving us insane. Right? And so if AI exists, it destroys the environment. Well, that's a loss. But, but it reduces. Sean C Davis: But Taylor MacDonald: The the signal Sean C Davis: but Taylor MacDonald: to noise ratio for me personally. So, you know, that seems like a fairly self centered argument, but I don't know. Anyway, sorry, I went down a rabbit hole. Sean C Davis: Yeah. You know, I will say, okay, total yeah, total aside, but I I got the I got the Apple Watch for for fitness about two years ago. But one of the things that has actually come in handy is that I I get a ton of notifications on the screen and on the computer, but I only send a very select few to the to the watch. And so Taylor MacDonald: To the watch. It's good. Sean C Davis: I will occasionally get distracted when I'm talking with someone. by the watch, but if my so if my phone buzzes, I'm like, I don't care about that. If my wrist buzzes, I usually can take kind of like quick little glance, but stay focused in the conversation. Taylor MacDonald: Pretty sure I have undiagnosed like A D something or other. Like I'm on the one, we're all on the spectrum. I just want to point that out. If somebody says you're on the spectrum, that's not inherently a bad thing, is it? I mean, if you're on the spectrum, literally everybody could be on the spectrum. We just don't watch life is that's what I'm saying. All right, anyway. Sean C Davis: Life is a spectrum, Taylor. Everything's a spectrum. Taylor MacDonald: I think my problem is that it's like becomes this nagging kind of tick, you know? So if my phone buzzes in my pocket, Sean C Davis: Yeah, yes, yeah. Taylor MacDonald: as I'm driving a car, for example, this happens all the time. I'll get like and you know. And then maybe if I get two or three at the same time, like you know, and then I'm I'm like I'm like I have to get it out. But then Sean C Davis: god, that's the worst, yes. Taylor MacDonald: if you don't check it, or at least mine's configured, so that if you don't check it a couple minutes later, it's gonna buzz again. And so then it's like v by about the second time that those buzzes go off, I'm like about to wreck the car. Like it's just driving me absolutely insane. So I gotta pull over and Sean C Davis: Yep. Yeah. Taylor MacDonald: check it. I don't actually pull over, I just pull it out and do what everybody else does, which is multitask behind the wheel. That's totally great and I feel Sean C Davis: It's safe. Taylor MacDonald: To make that admission, but yeah, safe. Do it at home, kids. All right, back to the demo. Okay, so again, the idea of this was a whack-a-mole exhibit. And so what I ended up doing was I found this 3D model on MakerWorld. So since you're gonna buy a 3D printer now, Sean C Davis: Obviously, yes. Taylor MacDonald: once you so I got the Bamboo Lab X2D, and it has been extremely easy to use. It's got a mobile application that links up to a site called MakerWorld. Makerworld has thousands upon thousands of different 3D models and they're all open. Most, I think all the ones on MakerWorld are open source or free, like you can just download them and print them. They support different types of filaments and you know, different printers can do different things and that kind of stuff. But for the most part, you I've had a pretty good success rate of just going through there and be like, I like that. I'm gonna try that. And then in a couple of scenarios, I have modified them. This lemon right here that I'm holding in my hand, if you're listening to the audio, you won't be able to. See this obviously, but this is a I don't know if you can see the texture on this thing. it's not really I'm struggling to get the camera to focus. But I Sean C Davis: I can kinda see that it is Taylor MacDonald: mean this thing looks like a lemon. I printed this off and stuck it Sean C Davis: It looks like a lemon. Taylor MacDonald: in the we have a bowl of like lemons and limes in our kitchen, Sean C Davis: Yep, yep. Taylor MacDonald: and I I stuck this in there for a while and everyone was fooled. It was amazing. So really wonderful, Sean C Davis: That's awesome. Taylor MacDonald: incredible amount of detail, and I've been super impressed at how easy it is to get going. So once I started using Bamboo Lab, which is the software that comes along with the printer, it's really handy. It you know, there's an iOS version of this thing, and you know, you could do most of the work just printing and managing the queue directly from your phone, but for more complex stuff, like slicing the models into you know different filament groups and stuff like that, that would happen on the desktop version of the Bamboo lab or bamboo studio, I think is what they call it. Sean C Davis: Okay. Okay. Taylor MacDonald: And then I got into using this product called Fusion 360, which I guess is a fairly contemporary modern version of a CAD software. I've never done anything in a 3D space. I mean, anything of any value. I found that to be extremely easy to get going too with the advent of AI. You know, when I ran into an issue, I just screencap it and hand it to Claude and be like, tell what I'm doing wrong. And after a couple of like trials, a couple of you know hiccups along the way. I was able to get I'm I'm now fairly efficient inside Fusion. And so I started creating my own models. So let me show you what I got. So Sean C Davis: Well that's so quick quick question for you. Like, does that mean that you're starting from s is is AI good enough to start from a concept that's in your brain and bring it to reality? Or do you have to start with something that someone else has created? Taylor MacDonald: Well, that's an excellent segue, sir. Because when I started this, I had no idea what I was building or what I wanted. All I knew is that I wanted a table that would have six holes distributed across the top of them, and randomly lemons would pop out of each of the holes. Like that's where I started. Sean C Davis: Okay. Taylor MacDonald: And so I have about this much mechanical engineering or sorry, electrical engineering knowledge and even less mechanical engineering knowledge. And so this was like a big risk. And it wasn't an insignificant cost to get started up, but I'm pretty pleased with where it landed for all it's worth. So what I ended up doing was I opened up Claude from the get-go and I was like, here's what I want to do. Where do I even go? You know, I decided to take Sean C Davis: Mm-hmm. Taylor MacDonald: the plunge and I bought the printer. And so I think one of the very first models I printed out was this lemon. and it's beautiful, it's great. But then I realized pretty quickly, like, wait, how do I connect this lemon to one of the poles or the you know, assembly arms that will be raising and lowering the lens. and it encouraged me to get into Fusion. I ended up opening the file, which was an STL file. I imported that into Fusion and I was able to make, I don't know if you could see in the picture, and I'll play this video here in just a second, but see these little black, little black like pedestals at the bottom of each linen. Those were additions that I made to the model that I got from Maker World. So I just opened it up and And what, you know, was able to like attach a little cylinder shape to the bottom of it and like join that in as a body. And then I when I was printing, I I used two different filament types. So you can it's like black and yellow right next to each other. so obviously you can see this thing in motion now that we're watching the video here. these are little servos down at the bottom that are powering all the different devices. getting back to like Claude's support of this project, what I ended up doing was asking it to just walk me through. Like it I know this is possible. How can I do it? And it did a number of both visualizations. Like it would put together like here's a drawing of what this should look like. And then it was just kind of like one step at a time. Build the base assembly. We need somewhere to mount the servo. Sean C Davis: Mm. Taylor MacDonald: I ended up buying a pair of digital calipers so I could like measure down to the millimeter all the different little corners. So when I bought the servos, for example, I just popped them out of the box, and you can see I like that fat part, I literally just measured around the body element, and then I put a hole in fusion into my object that I can mount it. I ended up learning about all sorts of stuff. I learned that you can buy these things called heat set inserts. This is now my note, my most favorite activity is to install heat set inserts. It's the most satisfying thing I've ever done. When you print So if you think about 3D printing, Sean C Davis: I need to know more. Taylor MacDonald: I didn't know this. And I I mean it totally makes sense, but if you stop and think about it, it's like, well, how does this work? Like how does it actually build something like this? And all it's doing is it's basically making layers, right? So this linen might have like 300 layers in it, right? So if you stop and you like chop it into 300 little slices, conceptually, Sean C Davis: Mm-hmm. Taylor MacDonald: you can imagine that the extruder on the printer, all it's doing is like going over each one of those layers. And all that it means is that it's Taking in filament, which is like plat hard plastic on a spool, you run that into the extruder, the extruder gets worn, melts the plastic, and then the little extruder nozzle is just being moved by a little motor over the position that is that layer. And when it gets done, it Sean C Davis: Mm-hmm. Taylor MacDonald: inches up and does the next layer. When it gets done, it inches up and does the next layer. And that's how it works. And I this is probably Sean C Davis: And it's it's hardening s like one layer hardens by the time it does the next layer so it can stack it. Taylor MacDonald: Well, what I'm learning is is that you know, the more yeah, there's like some I don't know all the physics about it, but yeah, there's some sort of connection between the layer that was just laid down and the layer that comes on next, and that these need to bond well. And some of the Sean C Davis: Mm-hmm. Taylor MacDonald: things that can influence how well that bonds is the temperature of the prior layer. And so some of the issues that I've run into and what I think are are amazing, I still don't know how this works. But like these little servo mounts have like a little window in them. And I don't know how in the heck it prints a like a like a gap. You know, somehow it strings this like hot plastic across this little window. Now, when I print the servo mount by itself, that works perfectly. It's beautiful. It's always clean, nice, and nice and tight. If I print Three servo mounts together or something, where it's having to like bounce between objects to lay the layer down. Because if you imagine you have three of the same object, those Sean C Davis: Mm-hmm. Taylor MacDonald: are all layers, right? And so it's going to print object number one, layer one first, and object number two, layer one, and object number three, layer one. So you've increased the Sean C Davis: Right. Taylor MacDonald: round trip. What I have learned is that the more that you increase that round trip, the more flaky the layer would be, particularly in those moments where you're spanning a gap. And what happens is it gets real stringy. Sean C Davis: Mm, mm. Taylor MacDonald: It's almost like it lays down a a string and it's good for a minute, but it starts sagging as it moves onto the other objects. By the time it comes back, it doesn't have that flat set foundation anymore. So I'm learning all sorts of stuff, and I I highly encourage you to do it. as far as the internals of this. The way it's working, I bought an Arduino Mega, which also Claude was like, dude, go buy it. Claude basically put together a bill of materials for me and even gave me links over to Amazon, like buy that servo, buy this ball bearing, buy this, you know, T nut, buy these extrusion rails. And so I kind of just yellowed it on a leap of faith Sean C Davis: Mm. Taylor MacDonald: that Claude was sending me down the right path. And I can confidently tell you that it did. It was not without trial and error. All of the pieces that I was responsible for. Like it could it was a little bit painful, like it would be like you should build a yoke at some point. So I originally started with a servo that was just pushing the rod up and down. And I was thinking like it had a it had like a pivot arm. So the idea was that on the servo you've got a little motor head and you attach what is called the crankshaft or something to that, and that's what gets moved when the motor twists. From there, I had another angle or like another arm pivoting off the end and that was my pivot arm. And I attached the metal rod to the pivot arm and the idea is in the idea of my head, or at least the architectural design, was that, you know, when this crank shaft goes up, it will just push the pivot arm straight up as long as, you know, but pretty quickly what I ran into was lateral force. So I'm learning all this stuff, so I never knew anything about this. But the lateral force that was being applied to the vertical rod as a consequence of the way that assembly worked. was causing it to like move too far left and right. So then it Sean C Davis: Mm. Taylor MacDonald: came along and it was like, dude, you should build a Scotch yoke. And I'm like, the hell? I don't even know what that is. And so Sean C Davis: yeah. Taylor MacDonald: it took me a while when it would say that stuff. And I think one of the things that bothered me or that got really persistently annoying was that Claude kept assuming I knew all these terms. It was like, just connect that to to Sean C Davis: Yes, yes. Taylor MacDonald: the GPIO pin D15. And I'm like, yo, slow down. And I mean, I've learned a lot and it's been great. And that's one of the aspects that I have really appreciated about AI is you can just be like, stop. Just stop. Like, explain the most basic concept to me because you've moved five steps too quickly, and I'm I'm not I'm not following you. So anyway, this has got the brain that's powering it all is an Arduino Mega. And then I bought what's called Arduino Mega Power Shield, which is this little thing that connects on top and gives me all the header pins, and then I'm using those header pins, like basically I was I had Claude help me write the programming instructions for the device. And and through that experience it would be like connect servo number one to header pin, you know, digital pin number, you know, eight. And so then I would go Sean C Davis: Mm-hmm. Taylor MacDonald: wire that up and and it so it walked me through every single step. Now there were a few things, adjustments that I made to the build, like using Dow rods. Instead of I I originally tried to print columns and that became really brittle. And so they a couple of them broke on me. And so I was like, that's not gonna work. And so I ended up moving to like three eighths inch poplar dowel rods because I can I could Sean C Davis: Mm-hmm. Okay. Taylor MacDonald: like shave them down myself and manage it. And then if they get a little bit loose, I can pin them in to the various you know, components that they're attached to. But anyway, it's been an amazing journey. I have done Sean C Davis: Yeah. Taylor MacDonald: I'm I'm loving the check this out. This is a another 3D object that I printed. For me and Crimson, because the kids continue to Sean C Davis: nice said. Taylor MacDonald: interrupt us. Check it out. It even has a little LED. I've I've installed a little switch on top. If you're listening to the podcast, what you're seeing here is a 3D printed on air, like a traditional on-air box that you would see in like a recording studio. Sean C Davis: Yeah. Taylor MacDonald: and it's got a little switch, it's got like an LED like running around the inner inside of that, a little battery compartment on the inside. I mean, this is the coolest thing I've ever Sean C Davis: That is so cool. So instead Taylor MacDonald: Ever been able to do? Sean C Davis: it's like okay, so I'm I'm guessing it costs more probably to make that than it would to buy some like piece of crap off of Amazon. But also you get to do exactly what you want to do. Yeah. Taylor MacDonald: You're referring to this? How much do you okay? So you're referring to the on air box that I'm sh that I'm showing here. Why let's guess how much you think this might cost. Okay, so Sean C Davis: Yeah, I think if you I here, let me here, let me look right now. Taylor MacDonald: again, if you're not listening, and this is pretty solid construction, right? So this is probably what, like maybe six inches by four, three inches or something? Rectangle. It's got translucent lettering that says on air against a red backdrop, and then there's a little rocker switch on the top that illuminates the entire box. And it looks kind of funky on my screen, but in person it looks frickin' cool. Anyway, you would be surprised at how much this cost. if I were to itemize this. Actually, what what is your guess? How much do you think the total round trip on this? I mean not excluding the Sean C Davis: All right. let's see here. Taylor MacDonald: cost of the printer, right? But if you want to include the cost of material like the the filament that's fine. 20 bucks. Yeah. Yeah. Sean C Davis: Alright, this this is this is similar. Twenty dollars, right? It's similar, it's similar. It's got an it's got a USB charge rechargeable, but you know, whatever. Taylor MacDonald: That's almost yeah, okay. So on Amazon, Sean's showing us a a similar product, a twenty dollar on air box with a little rocker switch. Yep. Yep. I mean I could do that too, but this actually has four double A's, and I think the lifetime on this is gonna be like there like a I think I'm gonna get a lot of life out of it. Anyway. Sean C Davis: Yeah, I yeah, yeah. Taylor MacDonald: Okay, so the material itself, there's probably about fifty cents worth of printed material in this. Maybe a dollar's worth of printed material. Yeah. Sean C Davis: That's it? Holy smokes. Taylor MacDonald: It's so you basically meter it based on weight and meaning the spool, I don't know how much but you get like a pound of filament, we'll say, and every product in it that you're printing is going to consume some of that filament. And so you could basically quantify the the actual real value, which I have not done, but you could quantify that based on you know how many how many grams of filament you're consuming in your printed product. Sean C Davis: Okay. Taylor MacDonald: On the inside here, and I might be able to And show you. on the inside there is a battery shield that cost me about 50 cents. So that's holding the battery. All four double A's. The double A's themselves were what? I don't know, a couple bucks probably. The Sean C Davis: Mm-hmm. Taylor MacDonald: rocker switch, I think that cost like a dollar. The LED that is inside here cost about a do I probably spent two, two or three bucks on the LEDs all in just to get them shipped, and then I cut it down to size and wired it up, and then I think that's it. because there's no resistors. An LED, the LED strip that I had already had resistance built into it. So normally when you use like a traditional LED, you need to have a resistor to prevent it from frying the LED because it'll send too much power. And so the resistor will like slow down the amount of power Sean C Davis: okay. Taylor MacDonald: that gets to the LED. in this case I didn't have to do that. So I think probably all in this probably cost me about five bucks and maybe like 20 minutes worth of time once I actually had all the materials in front of me. Yeah. I Sean C Davis: That's amazing. That's amazing. Okay. Taylor MacDonald: know, it's incredible. And it blows my mind, dude. I wish I could be my son in this day and age. I mean, granted I know that's a lot of statement, but like yeah, Sean C Davis: There's there's trade there's trade offs, yeah, yeah. Taylor MacDonald: you know, the environment, like we used to sit by a stream, but now my son can build crappy plastic products in the you know confines of his own bedroom. the truth is Sean C Davis: Yeah, 'cause all the pools are empty. Taylor MacDonald: the truth is Sean C Davis: Sorry, sorry, continue. Taylor MacDonald: no you're good. The truth is is that I Have always been interested in manufacturing. I've always been interested in how people make things. How do you put these things together? And this is such a circuitous route to getting here, but like this is the coolest thing ever. Now I have all these Sean C Davis: Mm-hmm. Taylor MacDonald: ideas. I'm gonna like do this. I'm gonna fix that problem. I so I'm really excited about the possibility and to think that our children are growing up in a world where you can go from idea to fully functional output. You know, at a fraction of the cost of what you would spend to buy this stuff. I mean, sure, you knew you could always buy that rocker switch for twenty cents a pop, but who's got the time, energy to sit down and learn an entire industry's worth of stuff simply for a font DIY project? Yeah, well now AI does it all for you. It's it's freaking nuts. Sean C Davis: Yeah, 'cause you just don't you don't have to go to that low level. I I I don't remember who it was. It was a I think a friend of a friend had started had kind of started a business or at least an Etsy shop or something, because they had I can't remember exactly what it was, but it was some outdoor recreation, you know, like car top carrier or kayak rack or something like that. And there was this part that was really weak on it, whereas the rest of it worked really well. And what he found was that it like it kept breaking, but that everybody was having this problem. And the manufacturer was kind of like, yeah, send it back. We'll give you a new one. And he's like, well, I can just make this piece. And then you buy this, you know, five dollar piece or whatever. And you don't have to go through all of that hassle and kind of like built a whole business around that. And now that you say this, I'm like, man, I've got all these problems where I'm like. Buying something that's too big or not right or whatever just because I'm trying to work around some very specific problem where I could I could probably just print a part that would fix it. Taylor MacDonald: Yeah. I mean it's the same argument that we're making in defense of like vibe coding software solutions, you know? I Sean C Davis: Yep. Mm-hmm. Taylor MacDonald: don't want to pay. I mean, listen, the value proposition has changed. It used to be like, yeah, I'm willing to tolerate that really exorbitant license fee for your software because you've got everything figured out and all this stuff, even if it doesn't necessarily like comport exactly with what I need out of that particular software, or it's got its rough edges, you have an opinionated product roadmap. I can't influence that. This is a constant problem with all things. And now, whether it is a physical printed product or a piece of software, you have a ton of latitude to be like, you know what? I'm just gonna do it my way. And I'm willing to accept those risks because it's me. You know, because I have made these I have made the choices. So that is a great segue to the next demo that I want to share. sorry to steal the thunder here today, Sean. I've got so many things that are all coming online here. Okay. Sean C Davis: No, this is great. I I like this theme of 'cause this is also gonna be in in the real world, right? Is this another Yeah, yeah. Let's keep it going. Taylor MacDonald: It is. Okay, so as you guys know, and I don't have it in front of me, but as you know, I have been a big fan of this Pocket device. So Pocket app, I think it's like heypocket.ai or something. They're not a sponsor of the show, unfortunately. I wish they were, but like these things are incredible. In fact, they're so cool that we're we're gonna be giving out a bunch of these at Christmas on behalf of the company, and I am real excited about it. So I probably should bleep that since it is a Christmas gift. Surprise and delight. Sean C Davis: Yeah. Taylor MacDonald: Just ruined Christmas. Sorry. anyway. Sean C Davis: It just Taylor MacDonald: Okay, so what this device has encouraged me to do is for every conversation that I have in person, which has always been the gap, right? Like since we've been remote at AMPL in 2020, pretty much every every meeting is over Zoom. You know, most of our staff, well, let's say about half of our staff live outside the tri-state area where we're headquartered. And so we just naturally have to negotiate like external. communication or remote communication. And what comes with that now in the age of AI, as everybody knows, are these annoying note takers that will transcribe the whole meeting. It will appropriate what Sean said, what Taylor said, and now you have a record of literally everything that's been said. And this has completely changed the way that we are doing business. for good and for bad, for what it's worth, but like if I forget the context of a meeting, I can just go back and rewatch it. I can even watch it on ta like double time. So I don't even have to burn that much time. Or nowadays with tools like Fathom, what you know, and granola and all these other things, there's like entire rag systems where you could say, What did Sean say about that one thing? you know? And so the idea is, wouldn't it be amazing if I could just pipe my entire brain into a into a similar rag system? And then I would have immediate access to all this stuff. And this is valuable because you and I will have conversations around the water cooler. and I will forget what you said, but I remember the premise of it. I'm like, man, Sean told me that cool story about that one thing, but I cannot remember what it was. If there only there was some way to just query for that information. Well, the pocket closed the loop in a big way. So if you and I went and had a sandwich, as we might do later today, if we have a conversation that is particularly inspiring and enlightening, that is now tied to my ability to recall that information. Sean said the most prolific shit. Over that bag of chips. And I cannot remember what that was. But Sean C Davis: Mm. Taylor MacDonald: the pocket gives me this real handy in my hand physical device. I press a button and all of a sudden I don't have to worry about it. I am using this constantly on the Forest Conservancy board where I'm the secretary. It's almost eliminated my job as secretary. Basically, we record the meetings and it will within minutes transcribe and derive outputs, key takeaways, largely pretty accurate. It gets a little confused on who's who because it's trying to figure out you know the same sound of somebody's voice. So I think maybe sometimes if it gets similar voices or I don't know how it's doing it, but anyway. So Pocket's great. Pocket is a device that costs a little over a hundred bucks. And to my knowledge, you can get the free version with no subscription, but you still get all the faculties of rag and and you know shareability of your conversations. And so I love it. I think it's been really impressive. However, As we've talked more and more about data privacy, as we've talked more and more about some of these other things, I have asked myself the question. and this is now quickly becoming the topic of a a couple of conference talks I'm gonna be doing this fall. I've asked myself the question of how can I do this offline? How can I do it without paying a third party at all? Still I still Sean C Davis: Mm. Taylor MacDonald: have this problem where if I have a conversation with you and you dump some trade secrets on me, well, now it's on a transcript and that's gonna live somewhere. And if it going out to the cloud, you know, And I'm not suggesting that I mean, companies need to have a very well defined use policy for these tools, like a clear framework, a clear understanding relative to the terms of engagement with their clients, confidential information, proprietary information, all that kind stuff. That all with is still very much an important consideration. And so some of the questions I've asked myself is like how if if I needed to keep this completely private or a closed system or whatever, how could I do that and then in doing so kind of keep my costs down? So I built one. Thanks to AI, what I ended up doing, and here I I'll let me see. I don't know how to like I want you to see what's on my desk in front of me, but I'm not sure how to do that. Maybe if I join from my phone, I can share my screen on my phone. Is that a thing you can do? Let me see. Sean C Davis: You can if you've if you've done that what what's that that continuity mode then you can use your phone as your camera and you can do the desk view thing. But I don't I think it takes a minute to get going. Taylor MacDonald: that's crazy. I didn't know you could do that. I just wanna join Sean C Davis: So you can use your iPhone as your webcam, which is really cool. And then when you go into the iPhone settings on the computer, you can switch it to desk mode and it finds where your desk is and shoots down instead of at you. It's it's kind of crazy. Taylor MacDonald: Can I not flip this around? Sean C Davis: look, you're here twice. Taylor MacDonald: Hey, here I am. Well, that's weird. I can't figure out how to turn the camera around. Maybe I can't do that. Anyway, I'll just do it like this. See if I can. Okay, see that? Sean C Davis: Mm-hmm. Taylor MacDonald: That is that is my pocket that I have created. Uh-oh, my video froze, there it goes. Okay, so check this out. So this over here is a Raspberry Pi Zero. When you press this button right here, it kicks on that LED. Now it is actually actively recording my voice. So what I'm gonna do is I'm gonna stop sharing on my phone real quick. can you see my screen? Okay, very good. All right. So this directory that we're looking at is a directory that's sitting on my Mac Mini. Okay, so on my network I have obviously a bunch of different computers. One of them is a Mac Mini that I bought back in January. It's really beefy. It's got a bunch of RAM in it and a and a a processor that's like really well positioned to do local inference models. and so basically the way this works is when I hit the button to turn off the recording. on my makeshift pocket, it is now going to R sync. See that directory that just popped up? So it's now Sean C Davis: Mm-hmm. Taylor MacDonald: going to R-sync the copy of this video or this audio recording off of the pocket device off the Arduino. Or sorry, this is a what is this? it is a it this is a Raspberry Pi. see these two WAV files? These are the audio that it recorded. and I I've got a little, obviously like a little microphone module on my breadboard there, and that's what was picking up the audio. And then it invokes a local model to do the transcription. So you can see this is the actual transcript of everything we recorded during that particular recording session. Now the Raspberry Pi is connected to my Tail Scale network. So it this never actually leaves my my network, what it's doing is it's R syncing the contents of the WAV files over to the Mac Mini. And then it's using, you know, an open whisper model to transcribe the text from the audio file on the Mac Mini. And so nothing ever leaves. And then what it does is it takes the contents of that transcript and it pipes it directly into a RAG database. And now once it's in the RAG database, I can go to any of my various agents And I can query for information relative to what was just recorded. It's amazing. It's the coolest thing in the world. And it never leaves. And because it's Sean C Davis: Awesome. Taylor MacDonald: on Tailscale, which Tailscale is like a private VPN that you're creating where you can kind of stitch together Sean C Davis: Mm. Taylor MacDonald: all your different devices on completely inaccessible addresses. I mean, you it gives you an IP address that's not even a real IP address, I think. It's it's it's like a local address just local to your Tailscale network. So this gives me more confidence that I can interact with what I've got on my Mac Mini, i.e., the RAG database, from all the various services that are connected, because it's going through this completely closed up VPN. I feel more comfortable about that at least, and I can confirm that none of the data that I just recorded, i.e., the confidential or proprietary conversation I just had with a client, never leaves my network and lives exclusively on my Mac Mini. What do you think about that? Sean C Davis: Yeah, yeah. This is man, I I mean I've been I've been tinkering around with something similar and I I kind of came at it with the same the same problem statement basically being these things are changing so fast, they're expensive, they they store the data in the cloud and their their SDKs are gonna and their APIs are like they're changing weekly. I want the control. I want the control. And then there's the security and the privacy aspect to it as well. And so I've started more with a kind of a cloud based scenario, but and I'm s and I'm starting to tinker with the duplex portion of this too, which is which is really fascinating. Taylor MacDonald: I hate dupe. I hate this crap. Can I tell you, I hate it. Like when you talk to OpenAI and it's like, hmm, okay, hmm, let me look that up. It's it's just it's like I just want just answer the question. Like, I Sean C Davis: Yeah. Taylor MacDonald: don't need you to like pretend that you're cognating the way I do. You're not a human. Stop trying stop stop cosplaying as a frickin' human. Anyway. Sean C Davis: Yeah, I'm my I'm wondering if I can get more control over that harness with duplex then so we don't have some of that annoyance, but we'll we'll see. That's that's early days. But I've I've a couple of man, I have so many questions I want to ask you, but I'm gonna start with two. So the first one is do you so the the database is on one of the machines locally, just like local SQLite database or something? Taylor MacDonald: postgrows. Yep. Sean C Davis: Postgres. Okay. Do you Taylor MacDonald: Yeah, Postgres with the vector extension. actually, no, I'm sorry. It's Chroma. This is a Chroma database, which is a little bit different. Sean C Davis: Mm. Taylor MacDonald: specific to vector, it's like a vector product. and that just happened to be whatever when I was vibe coding the solution, it was like, pick whatever you want. And I was like, I'm going with Chroma. Sean C Davis: okay. So are you are you backing up the machine and or and or database? All right. Taylor MacDonald: I am. Yeah. I am. So I'm I use I use time machine to back everything up locally to an external. Sean C Davis: gotcha. Okay. Because I I actually I had used this this product ARC for years, ARQ. And and gets and it backs up to a cloud storage provider and I'm just using Backblaze B2, but it I just I just I I had canceled my membership like four years ago because I'm like, everything I do is in Git and the and I don't care about anything that's on the machine. And it has completely changed because now there's so much more of a benefit of working locally. I'm like, backups are back, man. Taylor MacDonald: Yeah. Yeah. Well, and in so many ways, okay, so like back to that original statement. Does AI is AI getting us back to basics? Like, you know, cloud is great. It's been amazing, but if you've ever gotten a five thousand dollar unexpected bill because something bad happened or whatever, that's a real stressful experience to go through. I don't I don't love that. It was like the complexity of all these systems got so you know, overwhelming to the point that you're like, I'm willing to just pay somebody to deal with this complexity. But now we're kind of pulling it away. Like R Sync has been around since frickin' sliced bread. Yeah, I like i Sean C Davis: Mm. Taylor MacDonald: at the end of the day, the technologies that we're reverting to are tried and true. I don't need and even more, like even the non-technical user doesn't necessarily need, you know, an engineering degree to understand how to keep these things running. All you gotta do is be patient. And kinda understand the real obvious stuff, like don't leak keys, don't leave a gaping port to your network open and stuff like that. But beyond that, why fight it? You know? I think I think the Sean C Davis: Yeah, I think Taylor MacDonald: service providers have overplayed their hand. We're getting to the point where it's just like that cost value proposition is gone. Sean C Davis: Yeah, I mean in some in some ways, yes. I think this is this is a this is a question I think about a lot working at Netlify and not not trying to chill Netlify or anything like that, but that I think that for me there's some stuff that I don't mind managing and some stuff that I just don't I don't want the the stress of or the responsibility for. You know, like I was I was as I was tinkering with I think it was the voice thing and planning it out. One of the things that Claude said was like, Hey, you have this massive point of failure, which is your Mac Mini has to be on. And I'm like, Yeah, I know. I don't care. Right. Like, I'll just turn it on if it goes down or it needs to reboot or whatever. It's not a problem. But for my web application, I I want somebody else to worry about all that stuff. Taylor MacDonald: I completely agree with you. I am not suggesting you host a production website from your Mac Mini under your desk. That would be bad. But Sean C Davis: Ha ha ha. Taylor MacDonald: for many reasons. That said, you know, if you want to do it, theoretically you could. Like at that's the Sean C Davis: You could. You could. Taylor MacDonald: irony of all of this. You can shoot yourself in the foot if you so choose. I I don't know. At the end of the day, I keep asking myself the question of the amount of money, you know, when you go look at your expense light item, and ours are way up now. Because of all this AI adoption, you do sometimes need to ask yourself the question of like, and I guess this gets back to the AI addiction episode that we talked about this a couple years ago or a couple episodes ago where it was like, this is amazing and the promise is is unbelievable. Now that we know how much it costs, it's less amazing. you know, and Sean C Davis: Yes, Taylor MacDonald: it like at the end of the day, I don't know, we all have to kind of reconcile that reality. And I I do think that there's a whole lot of compute on these machines. That we're not using. And if we ship everything to a third party, it's Sean C Davis: Mm-hmm. Taylor MacDonald: not even just the cost. It's now the privacy. What happens to that data? Are they training on it? All this kind stuff. So to be continued, but it's definitely an interesting of evolution of the way that I personally have thought about adopting some of these tools and workflows and things. All right. Before we go, and we're Sean C Davis: Yeah, I love that. I love that. Taylor MacDonald: getting close we're getting sorry to cut you off, but we're getting close on time here. One other thing I wanted to mention is did you see that Stripe bought Open router for seven billion dollars this week. Sean C Davis: Man, I yeah, that was I I have been I've been I have been bringing that up. Taylor MacDonald: Sucks. I'm so bombed about this one, man. Yeah. Sean C Davis: I so I my my first two my first two reactions to that were what what world are we living in where where a router is worth seven billion dollars? This is what what are they doing? What are they doing in San Francisco, man? and the second thing is that I Stripe is up to something and because, you know, they came out with that projects API a couple of months ago and it just felt like a weird move. So I feel like Stripe is on path for like this hard pivot. But if you back up, you know, maybe something we could talk about next week is that In the in the 2020-ish era, there were all of these developer tools that what they were really trying to do was establish their one lane and do that thing really well. And now it seems like everyone is like, Well, I've got my lane, and now I'm trying to absorb all of these other lanes. And I don't think it's not necessarily a bad thing, and it's not necessarily and and I don't think we know. who's going to win. And I, you know, Netlify's trying to do something similar, but you know, in maybe a little bit of a different way than other folks. So it's it's just interesting. I think we could we could dig into that too. Taylor MacDonald: It's absolutely industry interesting and I I would love to dig into it. I can't help but think of the correlation between the way Netlify has evolved, its public facing persona in light of AI. You know, the products that you're developing Sean C Davis: Mm-hmm, mm-hmm. Taylor MacDonald: today are different than that Coralane. You had that very clear direction, like this is we host static websites. I mean if you go all the way back, you host static websites. You guys coined the term for the Jamstack movement and it made a ton of sense. It still makes a ton of sense. I mean, I would argue that a lot of those core principles around the Jamstack still hold true. It's all the other stuff that you it's you guys have to compete with the other competitors in the world that may have their own framework that are pushing the boundaries and like pushing you. They I don't know if this is true or not. This very well and forgive me, I d I again don't have a lens into the decisions on any of these companies, but I do think that you guys pivoted around twenty twenty out of that lane of static focus towards the hybrid model. It was driven by Gatsby, it was driven by the next JS adoption, it was driven by what Brussel was doing. And so I think that's I mean, I think it that's kind of just a necessity. It's an imperative for a business to evolve, especially in technology. But I do agree Stripe did something extremely cool. When it did it and it was so good. And it can that product has continued to be rock solid as far as I I'm aware. Sean C Davis: Yep. Yep. Taylor MacDonald: And I can only hope that this acquisition doesn't spell doom and gloom. But you're right, if they're growing out of that lane, who knows, man? Anything could happen. This could be the next sales force if if they want it almost feels like that's where they're going with all these adjustments. They want to compete at a much bigger Sean C Davis: Yeah. Taylor MacDonald: footprint than the enterprise. Which you know, I guess Sean C Davis: Yeah. Taylor MacDonald: makes sense from a business standpoint, but it sucks if you're the developer that has come to rely on their very easy to use product. Sean C Davis: Yes, open router is amazing. I love that. Taylor MacDonald: I I agree. I'm I'm a bit disappointed. Okay. Other headlines here are NVIDIA post record earnings again. Let's see. surprise, surprise. there was also some announcement between AWS and NVIDIA where they announced two million additional GPUs for agentic and physical AI. So there you go, Sean. Like this is becoming more and more a thing. Sean C Davis: Hm. Okay. Taylor MacDonald: Let's see. And then the last one I saw that was really interesting is OpenAI is going to start testing ads at Chat GPT. As of this month, they have started testing them. so it looks like we're gonna be seeing that coming to a agent response near you. Sean C Davis: Man. Yeah. It's the beginning of the inchidification Taylor MacDonald: Speechless. There's nothing to say. All right. Well listen. Sean C Davis: of AI. Taylor MacDonald: I know. I love that that's an actual word now. Okay. Sean, thank you. I appreciate your time. As always, it's been a pleasure talking with you. Thanks to everybody who's listening. please feel free to subscribe on any of the podqu podcast platforms of choice or come over to YouTube, drop some comments, go to our website, bloodsweat and tokens.xyz, and leave us some some feedback. We'd love to hear what everybody's thinking and where we might be able to take some of these conversations in the future. Until next time, this has been episode seventeen. It's a wrap. Thanks so much, Sean. I look forward to seeing you later.