10 August 2021 / Last updated: 10 Aug 2021

balenaPodcast episode 02: Bringing industrial-level 3D printing capability to the masses

For our second episode of the balenaPodcast, we spoke with Dival Banerjee (Vuecason) and Matt Parlmer (GenFab), two 3D printer entrepreneurs on their own missions to bring industrial-level manufacturing capability to everyone.
Check out the two-part episode. Here's Part 1:
Here's Part 2:
You can also listen to the episode on your favorite podcast app:
Please support our growing channels by sharing, commenting, and subscribing on your platform of choice. We appreciate it.
Check out the episode transcript below.

Full episode 02 transcript

ALEXANDROS MARINOS:
Let's go. All right. Yep. Hey guys. Great to have you on the stream. For everyone who's watching, uh, this is the balenaPodcast, where we're talking about new technologies coming down the pike, and what it means for every single person or community that is, uh, exposed to them.
Today I've got two of my favorite people on Twitter, uh, and outside of it, Dival and Matt, uh, who are working on some incredibly exciting stuff around the, uh, manufacturing, additive, and not, world. They're going to tell us a bit about what they do, and I'm going to let them introduce themselves. Uh, and then we're just going to talk about, you know, how, how the, the future that doesn't suffer from insanely long supply chains look like. Uh, so, uh, Dival, hello!
DIVAL BANERJEE:
Yeah, I'm Dival Banerjee. I’m working on metal 3D printers and making a combined metal 3D printer and machine tool. Uh, yeah. And it's called Vuecason (vuecason.com).
ALEX:
Cool. Uh, we're going to put links to all your stuff, uh, somewhere around the screen for everyone to, or audio, whatever, uh, for people to check out.
Hey, Matt!
MATT PARLMER:
Hey, how's it going?
ALEX:
All good. What are you working on?
MATT:
Lots of different things. Most of which have to do with, uh, taking open source, uh, robotics and commercializing them. That sort of stuff, starting with 3D printers and then, uh, onto labs, and then onto a number of other things and yeah, sure. We'll get into a bunch of that.
ALEX:
Oh yeah. So just to make sure everything's above board. Uh, I got to make a bit of a disclosure here. I've invested in Matt’s startup, GenFabCo (gebfab.substack.com). And if Dival would have let me, I would have invested in his too, but, you know (everyone laughs).
I don’t know, I don’t know. He's telling me he's not taking money yet. We'll see.
DIVAL:
We will, we will eventually. (laughter)
ALEX:
So, um, yeah, so again, uh, super excited to have you both here because I mean, I'm super into 3D printers, but as an amateur, I like to mess around with them. I, you know, I think through the, through the whole pandemic, they've been like a solace, like just kind of getting in touch with reality in a way that, you know, for, for me working on like business stuff, and when I do product stuff is mostly abstract model driven stuff.
Um, you know, having, uh, something I can touch is always like, I think I was missing it subconsciously, so I got myself into it. But you know, you guys are to the people who kind of know this stuff on a far, far deeper level. So I always look forward to when you guys are posting stuff on Twitter. And, uh, I thought, you know, if we're going to talk about what's coming down the pike, who else, who else can I possibly talk to?
MATT:
I'm actually coming at this from the perspective of a software engineer as well. And, and I can relate to the idea of like actually wanting to be able to touch it and hold the things that, you know, the things that I work on. There's been, uh, uh, you know, back, I mean, how all the way back in high school I did, I did a bunch of a bunch of work in industrial automation and it kind of feels like a kind of feels like a, you know, the wheel continues to turn here.
But yeah, I think there's a, uh, a lot of compelling things to be done in a lot of, uh, uh, interesting innovations to be captured, uh, at the intersection of abstract models and things that are constructed out of matter.
ALEX:
Yeah. I mean, in a very strange way that it doesn't feel like two distinct, like far away, uh, principles. It's more like, yeah, like the kind of weird horseshoe theory where the deeply physical and deeply abstract are needed in a bizarre way.
If I've learned anything from Dival’s Twitter feed, it’s that he's super passionate about, uh, feedback cycles as well. Uh, speeding that up.
DIVAL:
Yeah, yeah. Um, actually the thing is I've actually come from like a software background as well. Uh, I mean, I was doing computer engineering, but I was always obsessed with, uh, building hardware at home. And so normally I've been doing like firmware stuff, but then, uh, at home I've always been prototyping with a plastic 3D printer up here, a typical FDM printer. And that's how I got him to do all of it. It seems like we have a common theme.
ALEX:
Yeah, I know, right. This is a, I did not do this on purpose folks. This is just happening.
I mean, but, but there is, this is something that I've been noticing that, um, the hardware world, even from what I'm seeing in balena, um, is just very insular. It's like, it's, it's got its own language, its own way of thinking about things. And, hey, I mean, it's been working for a long time, and everything we've got around us from, you know, the technology we're using to even do this, do this conversation, um, has come about that way. But also, you know, when you were habituated to the, the beauty of what we have in the software world, um, the, the differences is kind of stark, I guess, when you got one foot in the one world and one to foot in the other.
So, uh, I mean, for us at balena that, that drove us to make, you know, a software delivery platform, because that was the thing that we saw was just, uh, excruciatingly painful when you, uh, sort of started messing with that world. But what we found out is that the hardware part, the hardware, hardware part, you know, like not the software on the hardware, the actual hardware itself is suffering from insane amounts of, uh, friction in terms of iteration. And you nevermind, you got to ship the stuff, like that's not going away, but between the two, uh, you know, how fast you actually get to the thing you want to ship? Uh, that that's just been a thing. One of the biggest surprises I've experienced in working at balena.
So where are you guys going to do about it? (laughter)
MATT:
Uh, build self replicating, uh, uh, factories and launch them into space, but that's like 20 years down the road. So I don't know. Do you want more immediate-term? (laughter)

How do we make hardware manufacturing as “push button” as today’s software processes?

ALEX:
So, I mean, what is the, what are like, if we put it out there more broadly, like, what are we, um, what do we have to hope to, like, if we're looking down like one year or three years or five years, like, what if somebody wants to make, you know, it used to be extremely difficult to do a software startup. Like, you know, if you look at like, um, stories about people who did software startups in like the early 2000s, or even in the 90s, like you had to go to Sun to buy the workstation, go to Oracle, buy a database license, put the thing in your office, buy like a backhaul cable, stick it on there. And, you know, like pray to the gods that your power didn't get cut, because that was your website. Um, and, you know, things got better gradually.
And then Amazon just showed up and it's like, press this button, you get a server, push your stuff can start from your basement. You don't need permission from anybody. Uh, and it feels like still, like hardware is, uh, at the place where, um, you gotta take permission because it's to build any, any sort of volume, I mean, Kickstarter, and that stuff helps. Uh, but still even the prototyping is quite expensive. So, you can't just do it like, you know, as a student with, you know, a great idea and no budget, like you're severely limited in what you can do.
So what are we, what are, what's the hope out of this mess?
MATT:
Well, I, I think that, I think that the, the, that is changing a little bit, you know, in, in various different areas. I think that, uh, you know, uh, the premise that we're operating under is that, uh, fabs are only as expensive as like the, the entry-level plastic part fab is only as expensive as it is currently, uh, for, for, you know, uh, just as a result of a lack of commercialization of, um, already mature, uh, FDM 3D printer technologies.
But I think more broadly than that, it's just a, it's the result of a, uh, a line of thinking in the, uh, in the, uh, physical engineering space that prioritizes, uh, capital density over modularity and, uh, other, other things like that. So, you know, people end up spending an enormous amount of time learning how to operate a machinery that can do things, um, you know, with, uh, with incredibly high levels of, uh, you know, of, of per operation efficiency, uh, but, uh, very low levels of, uh, uh, but very low levels of, uh, like reprogram, you know, reprogrammability like, you know, when you've spent $10,000, $20,000, $200,000 on machine tools, you tend to be, uh, sunk into a business model pipeline that's, that's pretty difficult to get out of. And, you know, that, that inherently limits a lot of how you do development, if all of your engineers are trained to develop things in that, uh, in that, in that, um, in that process and what that workflow.
Um, so yeah, I think a lot of it is, uh, ultimately cultural, you know, there's a saying that politics is downstream from culture. I think engineering is too, uh, and I, I think that ultimately, people need to actually be able to imagine a world in which they can, in which they can, uh, design things for a 10th of the cost or hundreds of the cost that they do currently, and, you know, enter serial production, uh, for a 10th of a cost or a hundredth of the cost, uh, before they actually go enter that world. Because, you know, you need to work your way and talk your way through the, you know, the, the second and third-order effects of that, so that you don't tank your business while you're trying to go do that. And, uh, that's, you know, that's, that's easier said than done, but a lot of that complexity can be eaten for people. And my hope is that some of the folks in the fleet of companies rolling out right now where we'll be able to do that for folks.
DIVAL:
Right. Um, actually going back to your question about, uh, if a student could like actually prototype, um, now-- I think it's starting to get easier, especially with like Arduino and seeing these low-cost microcontrollers, um, like the STM32 Nucleo , that's what I use. Um, and like with the 3D printer, like for me, I'm, I'm 21, I've worked with this 3D printer and some of the microcontrollers and other hardware, uh, and spent less than $2,000 and we're getting close to printing metal.
Um, so I think when, if an engineer thinks very strictly about, okay, how do I avoid machining? And, um, how do I use off-the-shelf parts? It's actually really, it's not too bad to get a prototype going. Um, not something that you can manufacture, but, um, certainly you can start to see if what you're working on will actually work. Um, and for me, I'm always obsessed about this iteration and just closing the loop as Alex pointed out earlier. Um, I think a large challenge with manufacturing, especially even with like, um, CNC, milling, or any kind of production like that, um, which is considered like, you can at least change the design quite a bit at the CNC machine. Um,y ou still have to think about fixturing and machining strategy significantly, especially if you're going into production. Um, and really what I hope is that in the next 20 years, with AWS, you don't really have to think about scale as much. Um, and hopefully we'll get there with manufacturing where you're able to just make a design, click printer, just click some kind of one click button, and then, uh, the part just comes out done. You don't have to think much about the strategy and you're still getting something that's economically feasible.
Um, and I think at some point when, you know, I think we're in hardware, we're still around maybe like “early nineties internet” in terms of how hard it is to do things. Um, and once we start to get something that's a little bit easier, then we'll start to see a lot more iteration, a lot more, uh, people trying risky, things that they wouldn't think would work. Um, at least for the smaller things. Like, I mean, of course, if you're doing like a supersonic or hypersonic jet, that's still going to be very capital-intensive. (laughter)
But, um, I do think we're starting to get to the point where smaller projects, um, I always go through like Hackaday and look at all the posts going on over there. And it's amazing to see what people are doing at home.
ALEX:
Yeah. It's, it's, I mean, definitely there's movement. Right. And, and, and a lot of the, I mean, the 3d printing revolution, I think over the last 10 years has done a lot, you know, it feels like we went from, you know, the, the MakerBots and like some sort of primitive, uh, riffraff stuff to, you know, the Prusas, and now, the next-generation stuff that's coming out there is, you know, I mean, resin 3D printing, it's like crazy, uh, SLA? Sorry-- is that the official term? I don't know what we, um, and, um, yeah, I mean, you can't say nothing has happened.

How can we make hardware design easier on people wanting to manufacture?

ALEX:
I guess, maybe another area. And I don't know what you guys think is around, um, sort of the design element of that, right. Because let's say I have my perfect, you know, omni-manufacturing, uh, you know, uh, machinery, uh, I still have to sort of put my idea down to, you know, in a form that it can actually handle. So, uh, is there, is there anything happening about that, that world?
MATT:
I mean, in terms of, uh, advancement in CAD?
ALEX:
I mean, if that's, if that's what we're looking at, that's what we're looking at.
MATT:
Yeah. I mean, you know, we've, we've well being a software engineer, uh, kind of, you know, pulling both pins and “Leroy Jenkins-ing” my way into the hardware world, you know, it has been a little bit of a, it has been a little bit of a, uh, an experience, uh, you know, diving into, you know, diving into all of this, uh, systems design that doesn't happen entirely in a text file, right?
Uh, you know, that sort of, that sort of stuff, uh, you know, and, and, and retaining like correctness and precision, um, you know, while you're operating in, in something that's representing, you know, a bunch of, uh, a bunch of things in three dimensions, um, and generally without like the sort of type system-level constraints that you can get in the graphics in the pure graphics world, um, you know, that's a big conceptual leap for me as a programmer. Um, and it, it seems like the tooling there is stuck in a lot of assumptions that were laid down in the, you know, in, in the, in the 90s.
I think that, uh, you know, Dival, as you said, like it does, it does seem like we're captured in the 90s in terms of, uh, uh, just hardware generally, but I, you know, th there are very specific things too, that that, that are literally done in precisely the same way that, you know, that they would have been back then.
Like it's still, uh, you know, we, we, I ended up finding out the hard way, why they, why they have those, like a “ball track” mouses. Uh, if you guys ever seen one of those, like a mouse with a big red ball that you just scroll by rolling the ball, that sort of thing. It turns out, I didn't realize this until I started spending hours a day in inside CAD, that, that's what that's for. You know, you've, you're, you're, you're stuck inside a set of things that, you know, have navigational primitives that are tied to, uh, you know, tied to a set of devices that, you know, might as well be a joystick. It’s so esoteric, uh, and so, you know, so unfamiliar and foreign to, you know, those of us who were habituated to track pads.
Yeah, it's, uh, if I had to say what the, you know, what appears to be the biggest gap and, you know, the biggest thing that the proliferation of new manufacturing technologies, it may end up the biggest block may end up actually just being people's ability to learn CAD, uh, you know, the developer tooling needs to advance along with, uh, you know, along with the form factor of the computer or the form factor of the machine tool.
DIVAL:
Right. I, the thing is for me, I, I guess, coming from software, like we're used to version control or used to all these things that just help us like collaborate with others.
MATT:
Type systems.
DIVAL:
Yeah. Yeah. Uh, I think the thing is, for me, I switched from... I've tried SOLIDWORKS, I've tried Autodesk, I've tried, um, tried to NX briefly, like I've tried some like Autodesk Fusion Inventor, like all these other things. Um, and I recently just migrated everything to Onshape. This is not an Onshape ad, but like…
ALEX:
I mean, if something's good, we're going to talk about it, right?

On hardware design, does the computer help you design, or does it just represent “what’s in your head”?

DIVAL:
Yeah, yeah. Right. Recently, like, it's nice because you can actually like branch off your design, you can create branches and, uh, like visualize the difference between two versions and compare things. And it's actually been surprisingly really good in that way.
Um, now I think one thing I always think about is like, how much design did you do with CAD is like in your head, and you're just trying to articulate it in the computer. And how much is it, how much of it is something that you're using the computer to help visualize maybe what you're thinking in your head or how much of it is just actually designed on the computer?
Um, for me, I always think that I do most of the design in my head, and then I'm just trying to like, translate it into the computer. So like a machine can make the part. Do you guys feel like design something you just do in your head and try to express on computer or it's something that you actually design on a computer?
MATT:
I mean, I, I still do a lot of like hand, like free-hand drafting, um, that pen and pen or pencil and paper, um, is still the, the tool that's the quickest for me. Um, but yeah, I mean, all ultimately like, uh, uh, drawing something, whether it's in CAD or on paper, is sorta like writing it down, you know, you have this abstract set of thoughts, right? And that needs to be refined into something concrete and, uh, uh, you know, the actual process of pushing it through that interface, whether it's language or, you know, or just, uh, you know, geometric representation of what you're doing. Uh that's, you know, that, that's a healthy constraint on the process that requires you to be, uh, very specific about where you're trying to do. Um, so I, I think, I think it's tough to separate one, one from the other.
Though, one of our engineers Curtis is, uh, you know, he, he, he does most of it. He, he, he has what he calls “Brain CAD,” and he does most of, most of his work in brain CAD, you know, doing laps of his block, uh, you know, listening to music, and then it comes down and just, you know, draws at once and it's done. Um, so some people are able to do that, but, uh, you know, I think for most it's, it, it really does, you know, the there's, the, there's a healthy constraint placed on the creative process by, by the, you know, the need to push it through an interface, um, that that might not cooperate all that well with you, because ultimately, like, you know, your manufacturing tools, aren't going to cooperate that well. Or, you know, if you're working on the software side, you know, your compiler might not cooperate all that well, if you just spit something out and then just say, figure it out, you know...
ALEX:
For me, like, I, you know… First of all, shameful admission here, uh, I was forced to buy an iPad, like the first time about an Apple product, after buying an iPod or my then-girlfriend, now-wife uh, like, I don't know, eight years ago is to buy an iPad to use Onshape. Uh, no, wait, what is it called? Uh, got it. Shapr3D, that's it. Um, and, um, you know, the way they've done the pen, the use of the pen and the fingers combo, is I thought it was blown away.
And still, it's, I find it, it depends on the, like, you don't only have to have the idea for what you're going to do, but also like design the, when you're in your head strategize about the sequence of steps you're going to take to actually make the thing happen. It doesn't feel like you're messing with a thing, you know, where you're like, ah, like it a bit like this, or like that, it's more like you're strategizing, like, okay, what sequence of steps can I, how can I fool it into, uh, you know, showing up the thing I want more like, than, than being...so it feels more imperative, still, like code than declarative as in, you know, I dunno for like the down the line, we're going to a world where we can speak to it and say like, “Hey, like this, but like, make it long. Can you, can you round out the edges a little bit? Or like, whatever.”
Um, but, um, I still feel, yeah, there is a, like the interface itself, maybe abstraction we've chosen. I don't know, maybe it's legacy or something, like it's coming from somewhere else that I don't know that somebody is going to like, clarify for me under the scales of both of my eyes, but like, it feels it's coming from a, yeah. I, even though, like, I spent like, as much as I, uh, as I, you know, as I could on making an interface, that would be natural for me, but my dream was to be able to do it without thinking, like maybe in the background of the code or something to be fooling around with something, it still felt like a very, very painful, uh, process.
I mean, maybe it's just, I'm not as good at it. Right. And if I get really comfortable with it and it'll make sense, but yeah, it still felt like quite arduous.
DIVAL:
Yeah. I think let's see. I think for me, like I kind of get the high level design for a part in my head. And then once I get in the computer, I can check, like, the thing is I won't be able to figure out, okay, how's this going to fit in the assembly perfectly in my head. Um, and that's what the CAD really helps me with is like, I'm just able to see how it fits into an assembly. I'm able to make some small tweaks, but then the high-level stuff, I kind of think of outside the computer.
Um, and I'm thinking like, maybe long-term in the future. Like, I don't know how many years, I'm not going to say how many years, but I think maybe we've seen GPT-3 with, uh, like someone's made like these things where you can just type a prompt and then it'll create a very basic React app. Um, I think maybe we could see something like that with CAD combined with parametric modeling or some kind of procedural modeling type of thing, where you're able to maybe pull from a part library, make some adjustments based on a prompt that someone gave and then design a part that way.
MATT:
Yeah. The constraint systems that they, that they're starting to introduce. And the interplay between those constraints systems in generative design technologies, like the stuff that Autodesk is starting to package, um, I think has, is that that space is underexplored as you say.
Um, I worry about the, you know, the, uh, pulling the human out of the, out of the loop when you have, you know, some, some sort of, uh, you know, fill-in-the-blank maximizer spitting out, uh, you know, spitting out various different, uh, you know, for example, spitting out React components or other, other things along those lines. Um, but yeah, I mean, it, it would definitely, it would definitely make, you know, being able to actually just go and, and, uh, uh, do, uh, design-by-example, the same way that you would do query-by-example with like graphql or something like that, that that would be enormously, enormously powerful.
And I think that might actually be the uh, the big get that, um, you know, the differentiates, like the next generation of CAD from, from existing, from existing, uh, uh, CAD systems that have the roots in 1970s, 1980s design technology. Um, there's a, and again, this just all comes down to habituation and retraining and that sort of thing. You know, we, if, if we open up a, uh, you know, if we open up, uh, uh, manufacturing technology to a much larger group of people, um, we increase the size of the potential size of the CAD market substantially, and therefore incentivize more people to dive in and, and, and start, you know, start, uh, doing development in that, uh, in that marketplace. And then you start to see improvements and, um, other things like that.
You know, my hope is that my hope is that we, we eventually ended up with, uh, some sort of like open source CAD.

Is hardware design thinking stuck in its 80s, 90s, 2000s priors?

ALEX:
One thing that I’ve been thinking about… Have you seen things like IceSL? There’s a couple of things that are basically like code, like IceSL... it's like Lua, I think, uh, to turn it into like designs. And the reason I'm thinking about that is because you said about React, and I'm thinking like, if we can get a code intermediary, then we can actually, it might be easier to go from there to, uh, to sort of some of some higher level interface than, uh, the, CAD itself.
I'm not sure. I'm just kind of thinking out loud really, but yeah, it's, it's, um, it's really fascinating because we keep coming back to this theme, right. That technology sort of is stuck in whatever first worked. Um, and then we're just continuing to develop that into, it's always easy, I guess, to add a step. Right. Okay. We're going to take that thing that everybody knows, and I'm going to make it a little bit easier, and that's always like such a compelling proposition, but it's, uh, it's hard to say like, “Hey, I got this new paradigm.”
It's like, well, you know, join the queue with all the other, uh, cranks that have “the new paradigm.”
MATT:
Right. Well, it's, it's like, it's, I mean, we've, we've known that Unix was suboptimal for doing what Unix does, uh, ever since, you know, five years after Unix came out, when all of the people who wrote Unix came back and said, “Hey, there's this thing called Plan 9. It's, it's categorically better at all this, you know, network computing thing, stuff.” And then, you know, and then everybody was like, “Oh, well, Unix is already very, very good. Um, we don't require additional improvements on this.” We're just going to run this, you know, we're just going to continue beating this donkey down the road for, you know, for, for another 30, 40 years before everyone realizes that they've essentially gone back and reinvented Plan 9, except in the worst possible way by grafting it onto the kernel.
(laughter)
I think there's a lack of awareness about the need for in like very serious, very targeted investment in completely new paradigmatic approaches to, uh, technologies, whatever they may be. You know, we we're, we're on, we're on 60 years of wide body of the same wide body of airline technology or airline or technologies, you know, we're on, uh, you know, we're on, on, on 70 years, uh, uh, deep into the same, uh, suboptimal light water reactors for, for, for, uh, warships that have never been pushed out to the commercial market. Um, you know, uh, the situation with reactors on land is even worse, you know, and, and, and, uh, I think folks just kind of sci-op themselves into this idea of path. Uh, you know, that the path dependence is a far greater thing than it actually is. And, uh, you just leave a lot of low hanging fruit when you do that for, uh, for a generation or two.
DIVAL:
I'm going to jump back a little bit to where you're talking about general design, and then I'll try to tie it back into what Matt said. Um, actually when I was doing the other day with Onshape was like, we're designing nozzles for the metal printer, and they're going to be like a cast out of like, it's going to be a ceramic material for the nozzle. So we have to cast it. And what's nice is you can set up things like variable tables, functions, and basically everything. Um, so that I can just type in, okay, change this dimension in the table, the model adjusts. And then I'm designed to mold to caps that nozzle. And that'll also adjust all in real time based on just like one variable change to like change, maybe so with, um, a flange or something like that. I thought that was pretty incredible. Um, it saved me at least like an hour or two of just spending time in the CAD program.
But, uh, I think, uh, it needs to be really easy to change things like that. And like, the thing is with what Matt was saying, like, I think part of the problem with like, getting people to switch to like a new CAD system, or even, you know, some of the other like actual like nuclear reactors or something, like, but especially with CAD, like, um, it's a lot of it is like we have these like really, uh, I guess like senior mechanical engineers that are very they're good, but they're very rooted in their way of doing things.
Um, I remember a couple of years ago, this was when I was in high school and I was, I was trying to get an internship at GE, but I didn't get, but, uh, they're talking about like, the problem we have is like, we have all these engineers that just do everything on clipboards when they're doing their inspections and we'd like them to use like iPads or something, but then I'm sure that like, whatever app they wrote was probably really s***, but, um, I probably don't blame the engineers too much, but also it's like, um, it's also hard to get people to switch, uh, something like a CAD tool because you get so accustomed to it and you just don't want to switch out of something like that because it takes a long time to adjust to a new environment.
MATT:
Yeah, for as much s*** as I talk about people not adapting to new technologies, like I'm, I'm on Emacs all day when I, well, when I, when I'm writing code like that, and I see absolutely no reason to move to this whole modern IDE thing that people that people, uh, have, have pushed for us. In my defense, it, you know, Emacs is a highly modular system, but, uh, you know, I guess, I guess, uh, we are all guilty in some measure of that sort of, that sort of a commitment to, you know, to certain types of tools.
ALEX:
Think about it now that you have an AI, right. And go for you now.
MATT:
Well, I mean, you can, you can plug, uh, you can plug the, um, uh...
ALEX:
The GPT-3 thing? Have you seen the whole, uh, open AI, uh, GitHub collaboration or...
MATT:
No, I, I heard GitHub had pulled the trigger on that. Uh, that one of my greatest fears is, um, is not so a lot, a lot of folks are afraid of like runaway, um, you know, like paperclip maximizer or of runaway grey goo. My personal AI enabled hell is, uh, is, is, uh, is an agent generating, um, components off of, uh, off of a training set of all of the like published open source trash JavaScript out there, uh, creating an eternal, like an eternal bottomless pit of JIRA tickets until the heat death of the universe.
That's that is my that's, that's my personal hell. That is what I pray to the God, I don't believe in, uh, that we avoid, um, someone should tell Roco about that one. Maybe that'll keep them up at night, but, uh, yeah, it's, it's, uh, that, that sort of stuff. I, yeah, I, I don't know. I'm sure I'm sure GitHub is, uh, uh, full of people who are cognizant of that risk, but, uh, I don't know. We'll, we'll, we'll see what ends up happening.

What will the first AI-created programming language look like / on solving problems with a language-creation mindset

ALEX:
I mean, this is a fascinating question because if you think about it, like, um, you know, the, um, the AlphaGo Zero, uh, type thing, you know, it trained on sets played by the, you know, grand masters get invented a style of play that was alien to them.
So I don't know, maybe, maybe it can get, you can break us out of the master, and invent something that makes sense. You know, actually this is a decidedly out of the, uh, the harderware space, but I do wonder when the first AI designed language will come--programming language.
Like if you think about it, like you got people, uh, designing programming languages based on assertions of like, whatever, you know, like just, oh, you know, I think this, or I think that, or whatever, it's like, we don't, if I've been, I've been toying with this idea, I don't know if it's going to be easy to describe, but like, if you take all of the code that runs on these easy to, uh, any easy to instance, right?
Like Amazon takes all the, uh, the assembly language that is sent to its CPU's and it just runs a massive compressor, right. That creates a dictionary, right. That dictionary is a set of common constructs its human sends to use in assembly. Right. So it could be a hint at a language that we could make there'll be efficient, but I don't, I've seen some people experiment with the compressibility of programming languages to figure out like, you know, what to, um, you know, relative expressiveness and whatnot, but like, I, yeah, I dunno, it'll be, I think it'll it'll happen. And it will be fascinating when it does, because it's, you know, programming languages are stuck in their own ways. And they're in a very similar sense as in everything else, you know, we're still not sure if we should do the call on equal or equals, like, you know, we're fighting about tabs versus spaces years, 2021, apparently that is a relevant topic.
MATT:
PL design-- So I, I spent, uh, a good chunk of the, you know, the, the last few years doing, um, you know, doing non-academic institution affiliated, uh, but still non-trivial, uh, PL design research and things, you know, things around that sort of stuff. And it is striking to me how much, uh, how much further you can get studying a problem, uh, or, or how much further you can get developing novel solutions to a problem when you can, when, when you just treat, uh, your development paradigm as, as language creation, um, you know, or as defining a set of, you know, uh, defining, uh, uh, functional space, um, uh, uh, and a toolbox around that to, you know, to, to manipulate something with, um, you know, with the sorts of precision that we can get from modern, uh, you know, from modern type checked, uh, uh, type check programming languages.
Earlier today I was messing around with, uh,you know, like basically like whenever anyone's trying to do something like that, um, they're probably using, uh, like an ML. Um, you know, I've seen people, uh, trying to essentially plug, uh, game-theoretic decision agents, uh, into Haskell.
Jules Hedges a researcher out at Oxford has, but done a lot of really interesting work on that sort of thing. Using the type construct, uh, uh, using the, the type systems construct, uh, uh, or using a type of construct called open games. Um, there's some interesting Cornell research on, uh, applying type systems to, uh, you know, to geometries and graphics, uh, that, that I was actually reading and tweeting about earlier today. Um, if anyone who's listening to this knows anything about that line of research, do please reach out. It's something that I've been doing a lot of intensive reading on lately.
ALEX:
Is that a path between sort of that world and sort of physical objects? Or am I, am I hearing things?
MATT:
Yeah, yeah. A hundred percent. So, uh, I mean, we, we, you know, there is, there's this fast field of like, um, uh, uh, uh, type systems, um, uh, as applied to geometry and topology and a bunch of other things like that. Um, this is waaaaay further in the mathematical deep end than I, than I can swim. Uh, but there's a lot of interesting thing going on with like homatopy type theory and drawing, uh, you know, drawing connections between, uh, different mathematical domains, uh, such that you can apply, um, the sorts of type system methodologies that we use to make very powerful programming languages, to, um, you know, all sorts of different, all sorts of different areas, whether it's a game theoretic decision agents or whether it's, uh, you know, whether it's, uh, uh, alligators, rendered alligators. Let me pull up the links so that, uh, uh, you guys can post in the show notes, but, uh, yeah, a lot of interesting stuff happening in that sort of thing.
You know, it all does seem to come right back around to, okay, what language do we want to put around this? And then how do we want to define our terms? And then what, uh, you know, what, what can we, you know, what, what can we do having, having to find all of this such that we can, can compose a bunch of new components together and, uh, uh, you know, do do interesting things with it.

Is additive manufacturing the new incubation space to build the next, more intuitive CAD system?

ALEX:
Honestly, it sounds like a race to get around the, um, you know, old school designers and to the masses, like whoever managed to make an abstraction that can bridge all the way to like regular people is the one. Like, I don't know if it's like scanning stuff with lidars and like, uh, that's on every iPhone now, you know, or, or is something that gets, so you got a 3D object and you're like to take this and like, make it longer or smaller, whatever.
MATT:
If I had to, if I had to like, guess what the path is towards that sort of thing for CAD, I think it would be one of these, one of these like programming shapes for nine year olds, languages getting really mature really quickly. Cause you know, some, some extremely talented grad student gets inspired and decides to, you know, f*** off on his project for a summer and just do that.
ALEX:
You know, that, that feels, you know, like I don't know about that and was just going to work, but it feels like the sort of thing that might, because what I've found new technology doing well, it's found in the incubation space somewhere where it's not people aren't worried about it. They're not sort of putting in an undue amount of pressure. It's not coming out like claiming, oh, we're going to change CAD forever. It's like, no, this is for nine year olds.
So, you know, like it's that kind of the, the raspberry PI or the, even the, the original PC or whatever, right. It didn't like, say like take this PC and take a lot of them and put it in your, you know, replace the mainframe. They didn't say that like, they, they just went for another market. And then from there, it actually grew back into the original, but you need to find an incubation space is what I find always.
I'd never thought about children as that place, but it does. It makes kind of sense for a lot of technology. I'm actually looking at my son, there's like two, right. Um, you know, he'll go to screens and start like swiping, uh, as if that will do anything. Um, you know, he's, you know, he's, he's absorbing you know, UX paradigms. Uh, so I don't know.
DIVAL:
Kind of do think that additive might be incubation space for at least new CAD systems. Um, just cause like a lot of what you're making with an additive manufacturing thing is going to be like something that's generally designed or it's gonna be like this crazy, like gyroid or like heat exchanger or something that you just can't do in normal CAD.
Um, and so like recently I've seen like in topology is doing a lot of interesting things and like, that might be the wedge you need in order. Cause like you can force them to say, “Hey, this is a slicer, you know, it's just, it's just helps. It's a software that's helping us prepare the file for printing.” And then you can slowly just add new design features on top of that. Like, oh, you can do generative design. Oh, you can do some kind of topology optimization for your parts.
And then slowly you can just keep adding features that like kind of replace CAD and a lot of ways that might be, that's something I've been thinking about quite a bit. Um, but also it's like, I feel like there's a lot of low-hanging fruit within, within just, um, taking things that worked in software for CAD.
Um, like there's no linter equivalent for manufacturing and design, like manufacturer's design. Like if you could make something that acts as a linter that could go and check like, okay. I mean, we kind of have this with 3D printers sites. It's like the overhang, like it'll highlight overhangs and that kind of thing. Um, but I think that having something like that built in directly to the CAD system would be very helpful.
Um, and eventually I think it'll kind of be similar to maybe like a hardware loop type of simulation thing. Uh, like the way we have for embedded software. Maybe you have like a machine tool that like has, uh, some kind of, uh, structured light scanner or something that is able to check, “Okay, your tolerance was off here. We're going to make some adjustments to the camera or the mold,” and then it'll just automatically do this in a loop. Um, I think that's what would be a really exciting future to have.
I don't know if that's the one we'll have, but, um, I do think that linters are very, that's a very low hanging area to be just for simple things. Like maybe your tolerance is off. Um, and like the way it works now is like you design a part, then you you'll do your FEA analysis, like fun element. And then this tech like, “okay, it can't handle this level of stress. We need to make the bracket thicker.”
I think those automating that pipeline would be a good place to start where you're able to set constraints and then, uh, like automatically import the CAD model or something like that. Or it's built in directly into whatever CAD software you're working with.

What are the differences in how academia, the startup, and industrial worlds tackle developing new technologies?

ALEX:
Actually Matt testing on a topic before I wanted to ask you about, Dival, which is, um, the difference between sort of the academic and the industrial and the startup world in terms of developing new tech. Uh what's uh, do you have anything, like, I know you've, you've mentioned a few things again on the social media about that. Like I'm sure you have thoughts, but I'd love to hear.
DIVAL:
Yeah. Yeah. I think with academia, they tend to do a lot of like intellectually industry, interesting things. And then just not thinking about commercialization and then industry tends to do the exact opposite where they're only focused on commercialization. Oh my audio pop back up. Um, but like, yeah, they only focus on commercialization or, and not really think of any kind of novel tech.
Like I think what industry should do more is try to take some novel technology and find a way to force the market to adopt it. I don't know how that works exactly. But, um, I think that's kind of what's happening in metal printing is like a lot of these capabilities are there that, you know, normally people don't look at as like something they need because they're not thinking, oh, I can make this heat exchanger or, oh, I can design this part with conformal cooling channels. Like, because that's like some crazy sci-fi pipe dream idea.
Um, and then like once you show that it is possible, then people tend to kind of adjust their priors and then rethink like, okay, maybe I can do this and use it. And maybe we can see a little bit more of that. And it's like companies that just forced their customers to adopt a better technology, especially within hardware.

Is talent drain a problem? Does all the best talent get sucked into government work?

MATT:
I think part of the problem with that is that, is that, you know, as, as you said, insufficient pressure to do new things for people who already have working commercial channels and insufficient pressure to do, um, to do workable things, you know, from academic folks, because it has no bearing whatsoever on whether you can fund it or not next year. Um, I think both of those are enormous issues that are exacerbated by the fact that we just have talent drain, uh, away from, uh, organizations that would be able to work in the, you know, sort of golden mean between those two failure modes.
Um, we have talent drain, uh, of those people away from a regular industry and a way from, uh, in a way from academia and into the aerospace and defense world, at least in the west. Um, the, you know, there are so, so, so many brilliant people, um, electrical engineers, uh, and mechanical engineers primarily, um, working in, uh, the aerospace and defense world who are doing, uh, duplicative projects essentially as serfs for the NASA, for NASA or the DOD or Boeing, or all three in some combination. And that is a, that is a, it's a really bad state of affairs to be in because, you know, we've created a, we've created a situation where absent, um, absent some way to pull a lot of those people out, you know, you just have it, you know, you have a big chunk of your brain power captured in highly unproductive, uh, you know, in highly unproductive ventures, you know, redesigning sub components of the F-35 for the umpteenth time.
Those folks would be able to do a lot more, um, you know, if they weren't, uh, tied into doing that, you know, and we see, uh, you know, we, we see a lot of brain drain in the software world, um, into big tech companies. But, I would argue that that phenomenon is far, far worse in the world of atoms than it is in the world of bits.
ALEX:
That's fascinating. I've never heard of that phenomenon. I'm in fact, I'm shocked that any part of, uh, government related technology is ahead. Uh, this is not how I understood the world-- do tell!
(laughter)
MATT:
ITAR! International traffic in arms regulation, AKA the way that the, the way that the executive branch ensures that a technology does not proliferate out of the, uh, out of the defense establishment and remains something that, uh, something that is, uh, available to only a very small segment of people. I mean, they, they arbitrarily go around and classify or, or restrict technologies all the time. Um, on behalf of, uh, on behalf of large industry players, um, it's not as bad as the impact that the patent system has on, on development, but it is pretty damn bad.
ALEX:
So, could it be that like a technology that is invented for non-defense purposes can be classified as like, just come into your startup. And they're like, “Dival, you know, a metal 3D printer, like, that's sorry, too good to be allowed to show to the commercial market.”
MATT:
The chainable Raspberry PI of GPU's was pulled off the market a few years ago. Um, and nobody knew why until the guy who, uh, was leading the open-source project turned up, uh, living in a completely different city, uh, working for DARPA.
Um, and apparently not by his own choice. So… (laughter)
Yeah. It was back during the early Obama administration when, when a bunch of folks, uh, realized that, uh, microchips were a national defense priority. So instead of doing the thing that would've made them, uh, that would've made microchips plentiful enough for us not to be worried about that, they just classified the technology and made sure that no one could commercially develop it. So now NSA has it and no one else does.
ALEX:
Given that “we can't have it, nobody else can,” rather than saying, “we're going to make it so plentiful and it doesn't matter anymore”

What causes “stasis bias” in technological advancement? Does it require war or conflict to stop it? Do we really want to live in a world with no driver or perceived need to innovate and make new stuff? Do we need existential risk to drive innovation?

MATT:
Uh, the, the, there's this, uh, there's this old, like kind of whiny left-wing essay about Silicon valley. Uh, God, I think it's called the Californian Ideology. And most of it is just kind of whiny, Marxist talking points, but they do make a really good, uh, statement about, um, how, you know, uh, the United States and more broadly the west, uh, does have industrial policy and in a very similar way to, uh, you know, to what the Soviet union did, um, or what the, what the PRC does. We just call it the Pentagon. And, uh, I think that, uh, that is, uh, for better and, or for worse, pretty accurate.
ALEX:
Um, I mean, people in Silicon valley, don't, don't like to admit that, um, Silicon Valley is a U.S. D.O.D. project, uh, pretty much.
MATT:
Essentially yes. I mean, Oracle's, Oracle's first customer was the Pentagon, microchips are… the whole raison d'etre of microchips are missile guidance systems...
ALEX:
After World War II, they basically just poured an, ungodly amount of money, uh, into that space. Yeah. After sort of Intel and after HP. Yeah. It broadened out and it was like all sorts of things, but like the first two or three generations, it was all Cold War tech.
MATT:
Well, and that, that coincide, that, that was like the move from, from of like big science had occurred at large-scale before. Right. Like you, you had, uh, you know, you had the initial, the initial period where it was sort of like, you know, the industrial revolution was seated by the Royal Navy in effect, so yeah.
ALEX:
Do tell! I’m interested now…
MATT:
Oh yeah, yeah, no, the Royal, the Royal Navy dock yards, the Royal Navy dock yards were, you know, that, that was where like James Watt was lived right down the road and went down there and said, “Hey, we've been using this little steam engine thing in a toy factory. Uh, would this be useful for any of the stuff you guys are doing building ships?” ...and, you know, the rest is history.
Uh, you know, that, that sort of stuff was, uh, you know, was, uh, uh, a thing of enormous consequences, like the guild system that rose up in the Mediterranean as a, uh, you know, a, an outgrowth of how the Roman army supplied its troops and then how the Byzantines did, um, you know, there's all these, uh, you know, there are all these other, all these other, uh, things that are driven by a defense investment, um, and not necessarily at like the state level either.
Um, and I think that that's, uh, uh, one of the primary drivers of technology. And I think a lot of people, a lot of people ignore that when they look at the history of technological progress, um, or at the history of, uh, technological stagnation. you know, um, the, the, uh, uh, you know, Japan stagnated because the samurai had all the swords and use those swords to take all the guns when the peasants couldn't find gunpowder.
You know, it was, uh, uh, that, that, you know, it, technological advancement didn't occur because somebody in the, in the Japanese state said that they were going to commit themselves to stasis for a few hundred years. You got what you got.
ALEX:
Look, so it may go that well, you will, but at least the, it means there's hope for, uh, you know, containing runaway AI. We just need the Samurai back! (laughter) I really mean that I don't really mean that. I mean, I've got a thread on Twitter now tracking all their, all this scary shit that's happening with, um, uh, you know, robotics/AI crossover.
Um, I don't know what to make of it, but I'm, I'm just tracking it to see what, like, you know, sometimes I just make a list of things. And then when you have the list, the conclusion emerges without you having to do much to force it. You're just like, oh, you know, because you have all these like, isolated examples in your head, and then you put them together and the picture sort of paints itself rather than you having to like, torture your mind.
I don't know… apparently it’s a very male, typical, uh, behavior to make lists of things. And I just embrace that.
So, so, yeah, the, so what's the, what's the, the takeaway there though? Is it that, um, government sort of, or centralized or whatever you want to call it, big money, uh, can drive a technology far more than is believed? Or, is it that it can’t happen any other way?
DIVAL:
Oh, Matt’s gone. (due to internet issues)
(laughter)
ALEX:
Or, oh, government or big tech just took away Matt... (laughter)
DIVAL:
Yeah! That’s exactly what I was thinking. The FBI showed up and he’s gone! He’s gone! (jokingly, laughter)
(laughter)
ALEX:
He said too much!
DIVAL:
But I do think like, uh, back in the sixties, like the DOD and DARPA was actually like pretty effective. Now, maybe I don't know what the thing is. Like, you know, I've never experienced working in the defense industry or anything like that, but I think that, um, they just had a lot more autonomy, right?
(Matt reconnects…)
Like where each person can have individual agency, whereas like today, like, you know, you could be an engineer at a large aerospace company and spend your entire job, uh, optimizing a single bracket for like a year. That could be your thing. Um, and so you just never get to see the rest of the airplane or whatever you're building.
And like, as an engineer, there's just no, like sense of like, okay, I can do these things. Like, you just don't think like a systems thinker at all. You're just focused on your one little area. And I think that, uh, you know, having the, to see the whole thing and see the whole picture is pretty helpful.
And I think that partially, yeah, partially this is due to like, just having really complex supply chains that are just like, overseas that you just never get to see the factory or anything, like as an engineer, I think that's really important to be able to see like, oh, this is where things are made is how it's made. And then you start to realize, okay, we can be so much better. And like, you know, this many, like, different ways.
(to Matt) Um, yeah, they're cutting your internet Matt because, because of what you said earlier, um, but, uh, I can hear you.
[group troubleshoots internet woes...]
MATT:
Sorry about that. I'm having my classic conductivity issues. Um, I mean, with respect to that, uh, you know, the diff the difference, or with respect to sort of the siloing of, of things in, in the defense industry, and then that siloing kind of repeating itself out in major corporate entities that copied the defense industry, organizationally, which is the vast majority of them in the west, um, that that's not necessarily a hard and fast organizational rule.
Like there's this whole, um, body of literature about, uh, Lockheed Martin’s skunkworks, for example, um, that goes into, in fact, I believe there's a book called Skunkworks right over here [looks at bookshelf]. I’ll grab it later. Um, uh, that goes into, you know, that goes into like how, uh, the, the most high performing labs and development operations at Lockheed worked. And the crazy thing is that if you compare the organizational model, which is a flat, um, uh, very leadership driven, uh, and, and highly integrated, you know, both design, like design, production engineering, everything happens within one complex.
Um, uh, at least everything did happen, uh, in one complex in Lockheed’s development arm, uh, you know, from the, from the Starfighter all the way out to the F1-17 Nighthawk. Um, these extraordinarily complex systems were, you know, were developed by a pretty tight-knit group of people with very short supply chains. Uh, and if you look at the other people in the world who do that, um, you know, it's, it's the Swedes who can produce, uh, you know, the could produce top tier weapons systems and the Finns as well, who can produce top tier weapons systems and top of your space systems, um, you know, while with, with a fraction of the budget, uh, that, uh, you know, that, that, that the United States or the UK has with its distributed contractor system.
Uh, and, and then it's also the Russian design bureau system, which is the, the, the world's most impressive aerospace program. And no one else is even close, just because they're able to do, uh, they're able to do, uh, things that challenge the United States directly with the economic resources of Italy [comparatively, not literally].
ALEX:
And the press free freedom of, well, Russia. No metaphor needed there!
MATT:
I guess it goes, I guess it does. It does, uh, it, it is helpful when you can essentially enslave everyone in the top 5% of your, uh, IQ distribution, uh, and make them work in secret cities, uh, you know, posh, posh prisons near Krasnoyarsk.
ALEX:
I bet I bet you, the top 5% of our, uh, distribution would not mind sometimes, but present them with this deal, right. I will take you to a village and there is no Twitter there. The internet is fine, there’s just Twitter. It doesn't work. And you can just focus on your stuff. There's a factory here. Let's do design here as manufacturing over there, uh, for, for prototypes, whatever do you want basically. Or you can stay near Moscow and, you know, be judged everyday on the Metro for not wearing the right clothes. Up to you!
MATT:
One of our co-founders, uh, Nick, uh, COO, um, he, uh, he spent a bunch of time at Oak Ridge, uh, which is the American version of, uh, the American version of the, of the secret hidden city near Krasnoyarsk. And, uh, I definitely think he would, he would relate to that idea of, you know, being able to retreat monastically to some, you know, to some like places like a place of, uh, uh, great scientific, uh, you know, learning and detachment.
Um, I think, I don't know, that's kind of the pitch for colonizing Europa. Isn't it like, you can go get, you can go here five or six hours out of the light of the, you know, the Terran light envelope and, uh, you know, and, uh, uh, avoid all of the, all of the, the, the nonsense while we, you know, while you go build a big fucking Halo Ring and a Matroishka brain.
ALEX:
Your mom's basement, but planet-sized.
(laughter)
MATT:
Yes, exactly. Exactly. It's going to be, it's going to be me [inaudible] and 30 other people on a Stanford Taurus, all going insane.
ALEX:
All right. We've totally done this into the, into the rocks
(laughter)
DIVAL:
Yeah, actually one thing I don't get is if Lockheed skunkworks, why don't they Institute it? Maybe I'm just really naive.
Like, why don't they apply that to every part of their company?
MATT:
Congress.
(laughter)

What’s at cost if we completely inhibit new ideas and unnecessarily limit technological advancement?

ALEX:
Wait, but I mean, my understanding, and I don't know whether you guys know about it, but my understanding is that even those systems have been sort of skunkworks today is not skunkworks back then. You know, it's a department, got it, blah, blah, blah, the career ladders, you know, all that stuff.
So, and, and honestly, like I have a personal principle against wishing for war, but I keep [inaudible] talking. I just keep hearing, you know, there's a, the paracletus is, uh, sort of adage, which is like, war is the father of everything. And it just, it just keeps playing. I can't stop it. Like it's like, war is the father of everything?
Like, there's no talent, there's no war, everything starts becoming proper and prophesized and, well, we can't do it like this, you know, we can't just let any engineer do whatever it has to be done properly, the proper child-- whatever.
Like, and then before, you know, it it's like, “Well, when was the last time we invented a new kind of thing?”
“Oh, I don't know.”
“And who does it?”
It's like, “Oh, there's this weird guy, you know, Elon Musk that we send the SEC after.”
Like, you know, like, it's like, yeah, it's, it's like, we don't want new ways. It bothers us. There's no driver and there's no kind of perceived need in a way COVID maybe was that, um, if we would let it, um, but there's no like, sort of, yeah, no, no perceived need, it feels like, uh, for innovation.
So people are, would, would take comfort over over that anytime.
MATT:
Yeah. I would like to think that conflict. Uh, I, yeah, I don't know. I think, I think the, the optimistic version of that would be that that existential risk is a driver of human advancement rather than just purely conflict, and that conflict as a subset of the existential risks that drive us to do things like, you know, migrate over continents or, you know, or, or, uh, do other things along those lines.
The profit motive is also obviously rather strong and, and we would be lying to ourselves and perhaps, uh, perhaps engaging in too much economism if we didn't admit that there's a real serious aesthetic and like moral component to, to, um, the model of, you know, uh, technological advancement that some people adopt. I mean, you know, I, I think that, I think that, uh, um, uh, you know, there, there, there is, there is some sort of like moral glory in being able to do new and interesting things and building stuff.
Um, but yeah, that, that's not a population-level motivator. That motivates people with a “hint of the ‘tism,”-- that does not motivate that does not motivate politicians, um, in the same way. Uh, it definitely doesn't, uh, motivate people with, uh, you know, an even more profound stasis bias.
ALEX:
Yeah, stasis bias is a great word. One day I’m going to make happen on Twitter, like, uh, one day I'm going to write up a little bit more about it, but this concept of like, you know, 1 trillion humans for a start. I feel like we need to restore, um, the, some objective for us, right?
Like right now there's like, what's the objective? What are we going towards? Like maybe pre COVID times? Like, that's like, we were hoping you'd go back? Like, who cares? Like what what's, what does that excite?
Um, you know, we, we, we, we need some things to be aiming at, and maybe that's the, you know, maybe war is a proxy for that, right? Like with war you, you know, you, you know, you need to beat that guy, and that guy is motivated to beat you. And that is a very clear and understandable motive, I think for humans and the rest of it is like, you know…
Now we're slipping into antinatalism and whatever. It's like, “oh, we need to solve climate work. Bring how do we go to the warming, whatever climate change, uh, how do we do that?”
“Well, let's just make less of us.” It's like, I mean, sure. Maybe, but like, is that like, every time you have a problem, you just going to retreat? Like, is that, you know, if you're not growing, you're dying...
MATT:
Uh, there's a, there's a Twitter account. Um, there's a Twitter account, uh, uh, run by a guy named roon. Um, he's, he's, he's kind of on the he's on the, the, the, the, uh, yeah, it's, it's short for, it's short for his fault full name, which I don't want to disclose cause he's pseudo anonymous kinda. Um, he, he, he, he says that all, all questions of human politics ultimately will, uh, collapse into, uh, uh, uh, “homesteading the local group versus ‘Ted Kaczynski.’”
And I think that...he’s not wrong?
(laughter)
ALEX:
Right. Antinatalist-- so basically “Kazinski light”-- got it.
MATT:
Yeah. I don't even know if it's Kosinski light anymore. I mean, look at some of the, you know, look at some of the, the “live within our means”-type rhetoric that you see from, uh, you know, from major political figures. Or, or, you know, a lot of the, um, you know, the orientation of, uh, of some of the more, um, delusional environmental groups, uh, uh, you know, folks, folks who, folks who think that, uh, we think that like de-growth is, is a plausible thing to sell to people.
I think that there is like a substantial chunk of the, of the, uh, you know, high status population or high status, uh, global high status population of, of, uh, you know, sort of English speaking university, adjacent elites that does actually think that, uh, constraining and managing, um,
uh, humanity is the right way to go.
And I don't think long-term, that's viable, um, you know, really, really difficult to push that hard against entropy.
ALEX:
It's so novel to our psychology. We're not, you know, we haven't done that as a, as a species to, uh, uh, to, to sort of say like, oh yeah, let's just live within our means or whatever it is, like, you know, “one child policy” or like, I don't know what this boils down to.
But it's like, I can't see, like, it's kinda like, you know, right idea, wrong species maybe, but even, so I just don't, I have this permanent, um, uh, permanent sort of, uh, you know, how to, I don't know where to, I don't know how to phrase it. Um, yeah, it's, it's sort of, uh, yeah, so, so, so it's just, doesn't it doesn't, it doesn't can't, I can't make it up to sort of go backwards.
Like we keep underestimating what we don't know. Right. It's like, we, we, we pretend like we have a full accounting of, of solution space and problem space. And we're like, okay, well, we've got to apply the solutions that we know to the problems that we know. And we have not full accounting either, right?
Like, you know, this, like nobody, very few people will come out and say like, if I don't know, it, it doesn't exist, but many, many, many people will behave as is. Um, and that's what really worries me.
Like, um, let's say you apply this, you know, really concerning solution to this problem that you, uh, perceive and I'm with you, like, you know, mathematically, it makes sense, right. Use less, uh, produce less waste, uh, you know, pollute less. I, I follow, but you know, what happens when an asteroid comes? If we don't have the technological capability to fight it off?
What happens when an epidemic comes and we don't have like something else, uh, the global cooperation that we need? You know, the dinosaurs did not require, you know, overpopulation and high tax to get extinct, so...
MATT:
Right, right. And ultimately, ultimately like the, you know, I think it's, it's very difficult to make the case, um, just from a purely like Darwinian motivation perspective that a species should do anything other than move towards the largest watering hole that it can find and fill them. Like, we, we are doing something that is-- if we are finding ways to shrink our population, finding ways to do, uh, uh, less with less, uh, and feeling good about it, I can't think of a more fundamentally a-human, um, uh, uh, you know, just philosophical position in that.
I guess the good news is that anybody who's committed to the, uh, you know, recursive improvement, uh, path, uh, very quickly obtains the resources to make sure that the people who are committed to the “stasis path” are no longer a problem.
So, or, or they're a problem that's a, you know, restricted to, uh, you know, restricted to a, uh, uh, you know, restricted to a planet and its near orbital environment. Um, but, uh, uh, yeah, I mean, it's, it's, uh, I think that sort of “Kazinski versus Iain Banks” is going to be the dimension is going to be the, that's going to be the fundamental question of the 21st century, which, which path do we want to stake most of our resources on.
Um, and the, the, uh, the, the next level of question that that's going to answer is how quickly do we move towards the growth path, because we're going that way or we're disappearing. And it's just a question of how quickly do we want to get on that train?

Why is there vilification of those with the resources to innovate and explore new technologies? Is this actually showing our single points of failure?

ALEX:
Yeah. I mean, one of the, you know, people like to hate on Elon and, you know, I'm well known on the reverse for being, uh, I mean, to the people that know me anyway for being pretty fanatical and support him. But it's, I do see it as embarrassing that we have, let one, dude, basically she'll show up all of, like, I dunno, several industries worth of technologists.
Like it is, I share their sense of humiliation. I don't react the same way to it, but like, it is absolutely just shameful. How, how much, like, I can't bear the single point of failure situation we're in right now.
MATT:
One of the fact that he had to start in FinTech to even aggregate the resources required to take a swing at the substantial more consequential problems of batteries and orbital access. I mean, that's, that's, I think that's quite the commentary on, on, on our capital allocation infrastructure.
ALEX:
Right? Exactly. I mean, it is, it is fascinating that, uh, this is one of the most hopeful things that as soon as he did prove with basics, that it can be done. Several space companies have now started, which is a relief. Like if he, if something were to happen to him-- like he's 50 years old, right? Like, I'm not saying like he's taken out by the Russians, which is also definitely this distinct possibility we have to model. Um, you know, it's just like whatever, like 50, 50 years old there is like a, um, there is a risk about that I consider unacceptable.
At least space industry. I think, you know, you got Rocket Lab, you got, uh, uh, the guys that are 3D printing metal. Come on, Dival, you got to know those guys, what's their name?
DIVAL:
Relativity Space. Relativity Space.
[inaudible]
MATT:
Are those the guys with the big macro scale, uh, with the big macro scale, uh, airframe printer?
DIVAL:
Yeah. They got to like the robot arm and they use welding wire.
Speaker 4: (01:14:35)
Yeah. That thing is sick. Is, is that a, is that a, is that a, uh, DOD tech or DOE tech transfer thing? Or did they in-house that?
DIVAL:
I don’t know actually. I don’t know. Um, they're in Los Angeles. I should find a way to visit.
ALEX:
It being space ttech, maybe they are in the ITAR umbrella and can do whatever they want, I don’t know.
MATT:
Nice thing about it is that all of space is America. So once you're up there, put an American flag on the side of your hat and you're like, “I'm not exporting anything! What are you talking about!”
And specifically, asteroids are Texas.
DIVAL:
Yeah. One thing I do want to do is actually want to sit down and like write a rebuttal to Ted Kaczynski, his manifesto. That's something I've been wanting to do for a while. That'll take forever though. So I haven't gotten around to it.
ALEX:
Um, that'll be amazing.
MATT:
And you read, have you read Ted Nordhaus and, uh, uh, the Ecomodernist the, their, their document, the Ecomodernist Manifesto?
DIVAL:
I haven't.
MATT:
It is essentially an argument for the disconnection of, of, of human beings from Earth's, uh, uh, ecology so that we can maintain a very, uh, high and increasing level of biodiversity, um, uh, and also simultaneously extract more and more and more energy without, um, you know, and, and deploy more and more energy and just, you know, throw Joules at problems without, uh, you know, without, without causing a whole lot of damage.
Um, and, uh, I think that, uh, I think that, that, that view of things is, um, you know, definitely more in the Iain Banks, rather than Ted Kaczynski, Ted Kaczynski angle. I don't think it's a direct reputation of industrial society and its, uh, and its future. But, um, uh, it's, it's, it's skating in that direction. I definitely encourage anyone listening to this finds finds this sort of, uh, bloviating about space stuff, to be interesting to go read that. Cause it is, it is exceptionally high rate content in that, in that, in that particular area.

Are we heading toward a VR or space-absolutist future? Would we rather amass virtual wealth and be comfortable or explore space and advance humanity and technology?

DIVAL:
Yeah. Yeah. Actually one thing I think is really dangerous is basically we just vilify material wealth. And so I think what’s happening is like, uh, now, I mean, this can all be done virtually, like you can acquire as many resources virtually when you're playing video games and that's not frowned upon at all. It's encouraged, right? It's like people are living out there. Like, uh, I don't know, I guess like capitalist dreams in like virtual space and one thing I've been thinking about a lot and I've been meaning to thread on this... This is like, uh, if we have like this VR-absolutist future versus like the space-absolutist future where like, you know, both of them kind of would fulfill our desire to like acquire resources or experience things.
But I think it's, uh, very important for us to actually go out in space and colonize space and go that way. And then VR can be like a small thing rather than like the main, uh, area where we experienced reality. Like, it seems like a lot of things are trending virtual right now and then like Elon’s the exception to that.
ALEX:
You know, I I've been tracking Elon's technology for a very long time. Actually. I have old blog posts that I actually like created like a tech tree from his work and how, you know, like, uh, this SpaceX is we, you know, creates, um, Starlink, and Tesla creates, you know, the Battery Ventures and, you know, like all of the houses all expands and the one I can, you know, Boring Company with, with Tesla have a clear sort of cross cut.
The one thing that never fits for me, Neuralink. Like it doesn't, you know, um, it's it doesn't, you know, like you can be like, Hmm. You know, all of the other pieces are basically the building blocks for a new civilization on Mars. Like you can see we've got a new space internet, we're going to need tunnels, we’re gonna need, you know, uh, vehicles to go around solar panels.
Like you can see it all coming together. Even his brother has invested in, uh, uh, in one company doing, uh, food and containers, and the other company doing artificial meat. It's like, it's almost like it all clicks together. Um, but, Neuralink is like… go ahead, say it...
MATT:
And have, have, have you read, have you read like any of the Culture novels by, by Iain Banks?
ALEX:
I keep meaning to. I've read the first one, which was amazing. And I started the second one The Player of Games and...
MATT:
So, so Neuralink is, is the neural lace from those novels? Quite it is literally just that, um, there is some sense in which, in, uh, in which, uh, in which Musk took these like utopian sci-fi books from the nineties and said, “This subset of technonology, like we're not going to be doing hyperspace anytime soon and we're not going to be doing. And you know, we're not going to be doing hyperspace anytime soon. We're not going to be doing planet-scale sort of like a mass mass mass stream, not going to be doing Dyson's fears. We're not going to be doing a FTL. We're not going to be doing like, you know, big Halo Rings. Uh, but brain-computer interfaces well within reach.”
You know, significantly more sophisticated AI than we have right now, also all within reach. Um, and I, I think he's honestly, I think he read a lot of those books and said, Hey, that's, this is, this is the direction we're skating. This is what I want to go do. Um, there was a whole group of people who kind of hopped on that hopped on that train. I think, um, you know, uh, uh, the Orion's Arm Universe Project was this big, uh, essentially like sci-fi story-as-a-Wiki, uh, that was started back in the back in the early back in the early aughts. And it's, you know, it's, it's, uh, comes from sort of the atomic rockets lineage of like very hard sci-fi that, uh, you know, that does its absolute best to stay within scientific constraints, um, while still sketching out, you know, glorious space opera.
And, uh, I think that, um, I think that, uh, uh, many people looked at that sort of stuff, myself included. Like, I, I, you know, I grew up on, on, on the stuff that Andrew Sandberg and, and, and, uh, some of the other people were producing, um, you know, on that project, it was one of the, one of the first, like, you know, truly beautiful internet, rabbit holes I ever found when I was a kid.
Um, I think that a lot of people grew up with that sort of stuff, whether it was, you know, pulp paperbacks or Wiki, as you know, or Wiki as, as a, uh, as art form, um, looked at, look to that as their refutation of industrial society and its future. Um, and, and of Kaczynski-ism.
Um, and they, they, they, you know, so may maybe, maybe in some sense, like the, the correct refutation to, um, you know, to, to Kaczynski, isn't like a concrete examination of the, of the, like the energy constraints that we're subject to on earth, but rather a more spiritual, uh, maybe, maybe it's a more spiritual, uh, sketch of, of, uh, you know, what humanity might do freed from those constraints.

Theory time: Does growing up with sci-fi change the scope with which we problem solve?

ALEX:
You know, I, um, I want to ask actually Dival as well if he's, uh, sort of grown-ups in this fiction. Cause I have this, um, have this pet theory and I don't, I haven't been able to prove it, but I know like for instance, Elon like absolutely grew up with that stuff. Yudkowsky, who I also read a lot of, um, who keeps saying how he grew up with sci-fi. And I did not, by the way, right? I was too busy imbibing fundamentalist Christianity, but you know, that is a story for another podcast.
Um, and if you want anything to know about theology, let me know. I've, uh, probably enough to qualify for a minor.
Um, but, um, the, uh, you know, I wonder if having to that stuff as a, as a, as a child sort of locks your “zoom factor” out. What I've seen, both Yudkowsky and Musk be able to do is you ask him a question and from your scope, it's like narrow, like, and they're able to just zoom out so effortlessly. They're not trying to do that. They're like, this is just natural to them, but because I do it so quickly. Make an answer to a very long-term kind of way that just absolutely dissolves whatever, you know, you give them a problem and they dissolve it so elegantly that it's like, it's not a problem. Like I remember one of the earliest things I heard from Musk and I was absolutely impressed. One journalist is like, “Oh, you know, are you going to put Android on your, on your cars? Or are you going to go with Apple?” Right? Which is like total, like, you know, tech press “gotcha” kind of questions.
Um, and he's like, “Our mission is to, uh, you know, accelerate a sustainable task here on earth. So I guess we're going to do whatever helps that.”
Right. And it’s so like WHOOSH!
You know, like, but you have to be able to think on that, at least on that zoom level to say, like, that doesn't matter, what is it, what difference does it make if I use Android or like, who cares? Honestly, it's pointless. But, if you're just locked into the drama, day-to-day, it's like, Hey, this is, this is the gossip thing. Right. You can call the same or you can go with them, you know, who gets the win? Um, and I see him doing that so effortlessly, I wonder if it had something to do with it. I don't know, develop as you, did you consume like an unhealthy amount of science fiction as a kid or what?
DIVAL:
Yeah. Um, I did a bit, for me, it was always comic books. I'd read a lot of Spider-Man comic books. Uh, actually the thing is like for me, Ironman, I was eight years old when the first Ironman movie came out. So, I watched that like probably 40 times probably?
I'm obsessed with Iron Man. Like I know this whole workshop is like a dream come true where I like just work down here.
ALEX:
Metal printing! Now I know why they all intersect!
DIVAL:
Yeah. Like a metal printing is like, you know, I would just, I just want to like dream life is just like being able to build whatever I want whenever I want to, and not have to think about how to make it. And I guess, metal printing, like kind of as an extension of that, um, like just giving me the ability to do that.
Um, and I guess that's why I'm obsessed with CAD and all these things, because I watched Iron Man a lot. Um, and I was, as a kid, I was always obsessed with like making jet engines out of Coke cans. For me, like it wasn't much sci-fi. I would watch a lot of YouTube videos about people building things. Like, I didn't know Hackaday was a thing. So I would always scroll through like Instructables where like people would be building things and it was really, you know, it was really hobbyist and it wasn't really properly done as engineering, but like, I think that's where I started thinking of things a lot.
Um, one time I tried to make like this sheet metal, like a robot arm that you can wear around your own arm, like an exoskeleton alarm because of the Ironman movie. Um, and that's got me into like using Arduino and microcontrollers in middle school.
So like that's what drove me. Um, yeah. So that's, that's how I think things actually, one thing is like, I did meet Musk once at the Hyperloop Competition. I got to ask him a question. I was trying to find I was fishing for startup ideas. That's what it was for me. So I asked them like, “Hey, what's an area of technology that you're excited about, but you just don't have time to do?”
And he’s like, “Um, supersonic, electric jet.”
Um, uh, what I noticed is that he tends, yes, he jumps to that question of like, does this advance the goal on pursuing your, not like it's either spacefaring civilization or something related to AI or, you know, your sustainability, uh, like mission that he has with Tesla.
But, um, what I noticed is he tends to make decisions really quickly based on his priors and then just act on it, just like that. And he doesn't hesitate at all. Once he's decided, then it's done.
Like someone asked them a question about, uh, “Hey, would you be able to extend the Hyperloop track?” Cause like right now we're the Hyperloop track outside spaces is like a mile long. But yeah, the top speed, the top speed is limited basically just in the distance we have, you can't really go above like two or three GS of acceleration at that point when you're using like something on a track. Um, and you wouldn't just, you just would run out of braking distance to, uh, so someone asked, “Hey, right now we're struggling to reach top speeds because the track would it be possible to extend the track?”
And he was like, “Yeah, maybe he looks over at Steve Davis and he's like the CEO of the boring company. And he's like, yeah, we could extend it. Maybe like 200 meters. So we'd have to buy the parking lot over there and extending them there, but maybe it's possible.” And then later that night he tweeted the next Hyperloop competition will be in this 10 kilometer underground vacuum tunnel, which will take years. So they're actually, that was the very last Hyperloop Competition, and will be the last for a while. Um, but you know, it was like almost like this impulsive decision, like, yes, we're going to do this at some point.
And that's how a lot of decisions and statements he makes on Twitter are done. Um, there's a huge action bias.
MATT:
Right, right. I mean, I think, I think what's interesting about how Musk makes decisions and how, uh, like other large centrally planned organizations make the, that, that do things well, um, make decisions is that there is a willingness to just execute on something once you've done some pretty simple physics-based math, right. And that's, that is, um, you know, physics or materials based if you do it like a surface level engineering study is oftentimes all, you really need to understand that something is absolutely viable. And, uh, past that point, it really just is a question of execution. And you, you, you know, you, you need to have a strategic, uh, strategic, uh, uh, process that can do good surface-level engineering studies quickly.
And then you need an organization that can execute on that. And I, I find, uh, you know, to key back again to large centrally-planned organizations, uh, two of the very big ones during the 20th century where, uh, you know, where the, um, uh, well, I mean the engineering corps of the Soviet Union and Maoist China. And, uh, uh, they, uh, you know, they, they, they, uh, are very famous for taking surface-level engineering studies and then running with it.
You know, that's what the, that's what the Three Gorges Dam is. And you, you can get yourself down a bad path with that sort of stuff, and you need to be able to, like, you need to be willing to drop things. If it's clear that it's not going to work, or if there are second or third order effects that are bad.
But I think that that's one of those things that, uh, you know, where, where it starts to become useful to actually introduce, uh, you know, to have your centrally-planned organization, be some sort of, uh, you know, be some sort of firm that is subject to, uh, you know, that subject to, uh, uh, requirements of, you know, return to investors because it's, uh, you know, it's a, uh, it's, it's a virtuous constraint that, uh, requires you to be pretty good at execution.
Um, all things considered as opposed to the sort of constraint that like Soviet engineers would live under, which is...
ALEX:
Don't get shot?
MATT:
Yes. Yeah. Don't go get shot. Don't lie too much about what you're doing, but don't not lie. Cause, cause that would be a problem too. Uh, you know, and, and, uh, uh, uh, and definitely don't allow your project to be finished ever because you don't, you don't, you don't want to see what your next boss might look like or where he might be.
Because they always need people for that. Uh, for, for that, uh, uh, nuclear aircraft, propulsion research lab that kills people all up in, uh, you know, up above the Arctic circles.
ALEX:
But, um, so the one thing I know about Elon that I, that I understand he, so my number one thing for startups is get good at failure recovery, right? Like, if you don't have that, you're gone and, and not because like, um, you suck if you fail, but at least you recover.
There is no way not to. And that is just a given and even trying to not fail is folly. And it, you you're, you, you failed already when you're trying not to fail anyway. So, accept that, and get good at coming out of it.
And one thing I'm noticing about Elon is he's got that to the absolute unimaginable max, I did not think he could do it that much, but he basically, you know, how they say the best way to like, find the answer to your question is to give somebody on the internet a bad answer and wait for the hoards to like attack you? I feel like he's taking that and weaponizing it because he doesn't, he doesn't care. His stress tolerance is probably his only superpower.
Um, and his biggest superpowers is being able to just take unbelievable amounts of flak. So he's like, “Well, if I can do that, I can just basically go to Bitcoin and tell them that their stuff sucks. And I'm going to do Doge, and just wait around, let's see what they do.”
And, you know, he doesn't care if the press like destroys him over that, or people start complaining about how he shouldn't be investing in Bitcoin or whatever. It's like, “Whatever, you know, I've been through worse, you know, I've had to put my last money into splitting it between my two companies. What's the worst that can happen?”
Whatever is broken in that man and something is definitely broken. It is making that sort of trade off works so well for him that, yeah, I think that, honestly, that's my operating theory for him. It's not that he is like, I mean, he's got a lot of, you know, uh, factors on the high level, but a lot of people do, he's not unique in that. It's just that he somehow is able to weaponize the internet into his advantage.

How does the 2021 state of the supply chain influence advancement and accessibility of hardware for the masses?

Um, but I'm coming back to the, the-- there's a topic we touched on a couple of times so far and I want to get your guys' takes, which is supply chain, right. I'm sure you're both suffering in your own ways from that. And I heard, uh, I don't remember who I mentioned the fact that, you know, short supply chains and innovation are like inextricably linked.
I think you both actually mentioned something around that. So do you think we're looking at a shortening of the supply chain, uh, coming soon? I'm now going to be the guy who says the wrong thing, the internet, wait for you guys.
MATT:
I think we're looking at a shortening of the supply chain by necessity for anyone who wants to do anything interesting, because quite frankly, it is a s show out there. The s show is compounding and not slowing down and it's hitting every category we can see.
So just last week we found out the hard way, trying to order a big batch of aluminum, extrusions, that the primary supplier of that, uh, you know, of high quality aluminum, extrusions, at least, you know, uh, in, uh, the primary one that you could kind of turn on and turn off, um, as a, you know, as a demand, a demand actually had a tap for that sort of stuff, uh, is now 60 days delayed and it's getting worse by the day, right?
So out of nowhere, we're like, oh s***, well, we need to, you know, we need frames for four, literally hundreds of 3D printers for the next couple months and for, you know, and for another, and then for another like five dozen test rigs for other other stuff, you know, other tools we want to be running on this on the same robotics.
All of a sudden we're, we're having to look at bent sheet metal for, for a bunch of things that, that, you know, we didn't have to before, because if, you know, because the factory that makes the sheet metal is in, uh, is in Indiana and there are shops that can laser cut it and bend it down the road from us here in Michigan, and it's suboptimal relative to, you know, relative to erector, you know, industrial erector sets, uh, in the form of like T-slot extrusions, but it very easily, could be the only option we have to make the machines to build other machines.
And that's, that's, uh, I dunno, I get the feeling that, um, that sort of problem is going to be exactly the kind of thing that, uh, uh, people across, um, the entire developed world are going to be experiencing, uh, trying to do anything industrially complex over the next couple of years.
I mean, the geopolitical uh, direction of things is very, very bad, far worse than most people think. Um, the, uh, just the, the base-level, like supply chain coordination issues, particularly with the concentration of, um, logistical systems, uh, and of, uh, just like sale point-of-sale, uh, into, you know, into, um, Amazon and Ali[Express] and a few others that's been extraordinarily pernicious, um, uh, particularly because when it does work, it works quite well. And when it doesn’t, it sure as s*** does not, um, you know, that sort of stuff is that sort of stuff is, is becoming a huge issue.
Um, I, I basically think that the ability to create short supply chains is going to be a distinct competitive advantage for anyone trying to do anything in the industrial space, um, going forward. And it'll be a big advantage for their customers too, or for the customers of whatever firms are doing that because, um, you know, if you can provide, uh, uh, predictable lead times that are short, that are high-volume, uh, you know, in that order, uh, then yeah, then you can actually keep the industrial engine going right now.
And, and as we're seeing “just-in-time” logistics now means “not-at-all logistics” for uh, increasing the volume of, of, of the industrial inputs. I mean, we lucked out, um, with boards, uh, driver boards for, um, well, for a number of different things that were plugging in. You know, um, we, uh, yeah, you know, uh, FYSETC, uh, big producer, um, did a run of exactly what we need, and that is hopefully gonna save us like a month and a half of BS trying to chain together Creality's Ender 3 boards so that they can run eight motor, you know, so that several of them can run eight motors while slaved to a Raspberry Pi. Uh, and that, uh, you know, that that's, uh, we got lucky. You know, we're not going to get way in over the, over the, oh yeah.
We don't want to get lucky. I don't like being lucky. I would very much like for, I would very much like for our, our, uh, our, our microcontroller problem to be, uh, one of writing Verilog for, for FPGA, for, you know, for dirt-cheap FPGAs that are, you know, that are turned out in some state-sponsored factory in Texas, where frankly, you know, I, I'm not, uh, I'm not...
ALEX:
You're not specific at all. (laughter)
MATT:
Next time, next time I'm in D.C., I'm going to be knocking down every, every senator's door that I can find. Uh, and every policy shop store that I can find, uh, about that specific thing. I mean, we, so, so, so badly need to take chip design and make it more general.
Like a, it just is, I don't care if we're using a 20-years-out-of-date process for that, like it's gotta happen and it's gotta happen like very, very soon. And, uh, yeah, I mean, it's, uh, yeah, I dunno if we don't show them that supply chain, the entire f** economy is going to grind to a halt in like a year and a half.
ALEX:
I was, I've been, uh, I've been thinking a lot about sort of new hardware and, and, and, you know, I think about things from the level that I am sitting at, right. Because I don't have a lot of access to that technology, something about it on the PCB level or whatever, where I can actually visualize it and think about it. Um, but I guess if it was easy to design chips at-will, then you would not have as much problem with all the little components that are the extras that are needed around it because those are needed to retrofit, essentially something that was made for one situation to another. Right. And you need all these little adapters effectively, like to turn your voltage into what you needed, or to turn it into some signals what you needed or whatever.
If you're just making it for what you needed, then you wouldn't need that. I mean, if you look at Apple's, uh, phones, they have less and less components in them. They just make the iPhone chip. They don't need anything else.
But I'm wondering if Dival has any problems with supply chain or, I mean, when, when your startup works, you know, you can, you can sell us one because you know, there's no problem with extrusions, but, um, how are you, how are you doing right now?
DIVAL:
So right now, I’m very early in this prototyping and developing the printer. So, I haven't really been affected by it. It's more like I just have to look around and see like, “oh, I need this. My grid troller is what's the lead time on this?”
And I did order a bunch of like STM 32 and development boards just in case something happens, but it hasn't really affected me yet because I'm just working on one or two printers. Um, and I guess the way we're working on these metal printers is they'd need this like huge, like almost like CNC mill-type of machines. So even at scale, like we'd be making like a couple of years down the line, maybe 30 a year, or 40 a year. Um, so we don't have to deal with a lot of the supply chain issues that like a lot of other companies are facing, um, which gives us an opportunity to kind of like vertically integrate into the supply chain as much as we can and do as much as we can in-house. Um, and so that's how I've been thinking of things.
Um, actually just for prototyping, like the thing is like, I can't afford to go make, go to a supplier and get my nozzles made. So I just like figured out how to slip, cast them on my own and, uh, work on them internally. And that, I mean, not only is it cheaper, but it was good for me to learn.
I actually do think, like it's very dangerous for companies to rely on external suppliers as much as they can and off the shelf components. Cause like, uh, you're losing a lot of like, I guess lessons and just like learning how to design these things. Like, so you're actually weakening like your internal like company and in their engineering prowess, because you're outsourcing as much work as you can.
So I think like, uh, not only is it good for like, you know, margins and making your product like, uh, but it's also really important for your engineers to be able to learn how to do these things, then you'll have much better engineers than your competitors do.
So you're competing on multiple levels basically. Um, and so like you, you avoid all the logistics and shoes you avoid, uh, like a lot of the existential risk of like, “oh, I can't find this part because my supplier is having a 90 day lead time” or something. And then part of what we want to do with the metal printing stuff is like, we want to prevent these types of issues for like people who are making molds or aerospace or something like, uh, where the unit economics works for using a metal printer for your end-use part.
Um, like one thing I thought was really interesting is like the way apple manufacturers are Macbooks is that they just mill everything, which just blows my mind. Like it's not a cast. Like they, they take metal and just mill it out and it's like, that's what people do not appreciate how insane it is to make this unibody design.
It's absolutely absurd. Um, but it'd be cool if more companies could do that. Right. And that's kind of what I hope that we can help other people achieve is that they can use you as this one machine to manufacture a lot of things for low volume use and then not have to worry about retooling, like even with CNC, with CNC milling, like you have to think of how the thing's fixtured. So, you're kind of locked into, at least for production. You're kind of locked into a specific design that you iterate on.
And I think like a lot of innovation originates from like being able to switch design completely or start from scratch, or like start a whole new architecture rather than just iterating and refining one little thing. Um, so then the retooling, fixturing, and thinking of all that with milling becomes a limiting factor on what you're doing with building, if not the design, like it's one of those constraints on your rate limit for building something entirely new.
Um, so what we're kind of thinking is like with metal printing, you just weld to a plate, you don't have to think of fixed string. You can. And when you do think of, I guess, fixturing your positioning or like your machining strategy, that's all done on the computer rather than having to go out and retool the actual machine.

How do we make manufacturing more configurable and accessible to the masses? Get the unit cost down, bring it to more places, give the advantages to everyone, not just rich shops

DIVAL:
Um, so I think like the more we can make manufacturing, like this FPGA type of thing, where you can just, uh, re I think traditional manufacturing will kind of be like these Asics, and then you'll think of like a use case for something like this similar FPGA-type of thing where you're like maybe less than 10,000 parts a year, and you're, uh, changing the design constantly and iterating on it. And you'll be able to do that in production.
Like, imagine if you could just like design something as a prototype and then, uh, like what software you used to like write an app deployed on AWS Lambda and scale that way. And that's amazing that you can just do that. And in hardware way, we don't have that. And it would be nice to have something similar to that. And that's what we're trying to do.
Um, or was that I was getting lost on something. Oh. And then one thing is like the one thing I am annoyed with 3D printing on is like, most people just use for prototyping and not for manufacturing. Um, and you know, there's like, I guess, unit economics problems to deal with there, but, uh, like it kind of defeats the purpose of the prototype if you have to switch to a new manufacturing method to manufacture it, um, in some ways.
So like, what would be nice is if we can make these machines that are capable of using the same process for your production and prototyping, cause then you're like basically like testing, like your production method on like a desktop machine, and then you can just push it to like, whatever, I guess, print class or a factory that you have, and then just start rolling out things in production, and then you can iterate much quicker on your design and not have to worry about like, you're... like the whole idea of like just-in-time manufacturing is to like, um, compress the iteration.
Like you don't have inventory cause like make a mistake in the design. It's like stuck for the next like thousand parts of something that you already manufactured. And you just have a lot of sunk capital that way. Whereas if you have this kind of like, I mean, they call it digital manufacturing, but like, you know, what does that mean? But the idea is like, you know, if you can do this where you're able to change the design and not screw over your inventory, like that's going to, I think, make it a lot easier for engineers to try new things. And that's what I'm really obsessed about.
MATT:
Yeah. Well, I would add to that also that, um, you're totally right there. I would add though that, uh, I think people radically, mis-priced the unit economics of, of even FDM right now, um, especially FDM like the quality of the, of the machines that you can get for right now around $2,500, probably next year, $2000 the year after that about a grand, um, is far far above anything you could buy for $50,000 and up, 10 years ago.
They are faster. They are better, they have higher yield rates and they have yield rates that are, that, that when properly tuned and when you're not dicking around with, with materials that that are, uh, badly specced are dealing, you know, ha have, you know, a bunch of other issues, um, you know, you can, you can actually start to achieve yield rates up in the nineties and, you know, and then, and then higher and then, you know, and then, and then you start going to repeating digits and, uh, that, you know, at least for certain, certain sorts of things.
And, um, I just, I think a lot of people aren't cognizant about that. I don't think they, they certainly don't design around the assumption that you can do the first 10,000 or 20,000 of whatever part you're doing. Um, you know, three days after you finalize the design or gotten close to finalizing the design, uh, on, on a side of print farms and that you can do that cheaper and with probably less lead time than it would take to, you know, to, to throw the football over the Pacific and have somebody in, uh, you know, in PRC or Indonesia, um, you know, do all your tooling and injection molding and all this other stuff.
So, for an increasing volume of plastics applications, the unit economics have changed already.
I would also say too, um, part of the issue that, you know, like what you were talking about, uh, with, uh, you know, the scale imposed by, um, or sorry, the iteration uh, timelines that are imposed by the need to do mold designer workholding or all this other stuff. I think that that is going to change a lot as we move towards lower force processes, like the induction, um, like the induction-based, uh, uh, filament deposition that you guys are doing. Or, uh, you know, or, or subtractive machining with electrical discharge. That sort of stuff, you know, your workholding becomes, uh, you know, becomes a, becomes just one more, one more chincy robot arm, uh, once you start doing stuff like that. And, um, that is wildly, wildly different set of constraints for, engineers trying to, uh, do, you know, do very traditional things with a block of aluminum, um, that, uh, you know, with a block of aluminum or with a bunch of, or with filament.
Um, so yeah, I think, I think that, uh, in this brings me back to, you know, uh, engineering is downstream of culture. Like if you don't, you know, if, if you're acculturated to this idea that you're reliant upon extreme capital density to do anything, um, you are reliant on extreme capital density to do anything. So you can't imagine a world in which that's not the case. So, yeah, I think there needs to be a lot of like demonstration and proving out of, uh, of, uh, uh, capabilities that, uh, right now are considered to be the, you know, in the province of big aerospace companies or of, uh, of, uh, the gentlemen engineer working, uh, working in, in, in a garage somewhere or the, you know, or the college student in her dorm.
DIVAL:
Yeah. That's exactly how I feel. And yeah, I think it's getting the point with embedded stuff. Right. But then, like, you know, still, like, I mean, I guess, like, we're starting to see more, like, I guess, makerspaces with CNC machines on college campuses, but those are still kind of like pretty difficult to use and like, you know, you, it does cost money.
But I think like for us, like, uh, when I was at school, it was like, um, so we're between 20 and $60 an hour? Which isn't actually that bad. But like, as a college student, it's still like a lot, but like for a company, like for us, that's like amazing to be able to do something at 20 or $60 an hour, um, something in that range.
But yeah, I'm hoping that it's going to change over time. Like for us, what we're thinking is like, I mean, it's gonna take a long time for us to make these machines. But, uh, we'd like to do is like offer a service where you can basically just, and then we'll like, have these centers where you can just like, uh, in, in like major cities or like high-yield cities, so like the lead times are low kind of like, it would be a dream if we could be like the additive McMaster-Carr, that's what we're hoping for. And then like people who use it a ton who think like, okay, maybe it's cheaper for them to just actually like lease a machine and have one onsite. That's what we're thinking might make sense. But, um, yeah, there's definitely a small group of people that need like metal printed parts, um, onsite and need a machine for that. But as things become more common, maybe obviously that leasing becomes more common than actually the service itself.
MATT:
I think it's, I think it's a question of the unit cost of the actual machines. Like a, you know, if we're talking about, um, you know, like, uh, big, uh, big SLS printers, uh, selective laser centering printers that have, you know, that have a bed of, uh, you know, that have, uh, just doing this for the context for solicitors, they have a bed of a bed of powder that, that, uh, laser goes over and, you know, uh, built, uh, you know, builds a, uh, uh, arbitrary, arbitrarily, complex geometry part out of, you know, out of the nylon powder that way. Um, you know, that sort of stuff currently costs, you know, HP will, will ding you for a quarter of a mil to buy one of those machines. And once you have, you know, you, you know, once, once you have it, you have that capability.
ALEX:
It’s probably as good as an HP printer.
MATT:
I s*** you not, they actually make you use printer cartridges.
(laughter)
Yup. Somebody in HP sales took a look at that and said, “Hmm, what do we, what do we make most of our money on? Yeah. Cartridge sales!”
ALEX:
Are they chipped as well? So you can't refill it?
MATT:
I'm certain that we will probably end up having to use one of those things down the road for some sort of process, probably just renting time on one. Um, I'm not looking forward to the moment that we do, because I want nothing to do with that technology stack. But, um, when the, you know, when the, when the entry level, uh, machine for that sort of process is a quarter of a mill that advantages job shops and, and, and, uh, Protolabs, right?
Um, when the entry-level cost for a machine like that is $25,000, and it can do the same if not greater output with, with similar, uh, part performance characteristics, that's very different world. And when it's $2,500, uh, the whole built environment changes. Right? So, you know, we, we think that, we think that, that, um you know, walking that improvement curve, uh, is the most consequential thing that, you know, that, that, that we can do.
If you can get the cost, if you can get the unit costs down, you can get it in more places. And once it's in more places, it improves faster because every, you know, you have more sets of eyes on it that are contributing fixes back, especially if you're not being an idiot and you're letting people do your research for you because you're keeping it open source.
I think that it's, uh, uh, I think that, that, uh, that improvement cycle could happen a lot faster than people think, but, you know, it does still need to be shown out like that, you know, that that needs to be, that needs to be demonstrated for folks. And it needs to be, um, actually made real, uh, we've seen that with printers, uh, you know, with, with FDM, uh, uh, FDM printers, um, and what I'm wondering when we're hoping to find out, uh, is, you know, what, what is the next, what is the next technology that expands the capability envelope of somebody who has two grand to spend on a machine tool, right?
Like how do you democratize and proliferate that?

If you had $2000 to mod a 3D printer, what would you do?

ALEX:
So, if I have $2000 today, what is, and I have a FDM 3D printer, what is my next best use for expanding my manufacturing capability envelope?
MATT:
Capability envelope? Um, if you have two grand and an Ender 3 right now?
ALEX:
I already, yeah, I have a Prusa. I don't know if that makes a difference.
MATT:
Expand the, the, the volume of capability that you have in terms of materials. So, uh, enclose your printer, get better software so that you can run it a bit quicker and with higher accuracy or, you know, with higher, like dimensional precision, um, uh, on your finished parts, uh.
ALEX:
Klipper basically?
MATT:
Those, for the uninitiated Klipper is, uh, essentially, uh, that the kernel, but for robotic motion systems, or it will be that, uh, you know, given another three or four years of development, um, yeah, I mean, you know, I, I would, I would, uh, I, I would do those things. I mean, really the, the correct answer here is, uh, wait a few months and then, uh, uh, uh, by, uh, by a Genfab rig, uh, and then, uh, help us, help us, help you figure out how to turn that into a business so you can buy five or 10 more Genfab rigs!
(laughter)
ALEX:
Dival, What would you have me buy a couple of grand to improve what I can make in my house?
DIVAL:
Sure. Um, I think what Matt said was pretty accurate. I mean, I'd pretty much agree, uh…
ALEX:
Basically double down on making my printer better, but not getting another one. Fascinating...
DIVAL:
Right? Yeah. Um, let's see if you, if you made that like $5,000-$7,000 I would say, maybe do a DIY, like, I think, like having an EDM machine where you can cut away metal, that might be pretty useful. And I saw this, like, I was looking around about a month back for... if there's any kits to do this. And I found like this one that's got where it is... I’ll have to look around after the show, but, uh, uh, basically like if you build your own motion system, you can build your own like EDW machine at home and start cutting away at metal, which, which would be pretty interesting. I think Matt’s giving me the eyebrows, maybe he was working on something. Um, but if you are, I'd like to buy one.
MATT:
I think the open-source community is about 18 months. Uh, I think the open-source community is about 18 months away from blowing everybody's socks off in terms of the capability of these things, uh, electrical, discharge, uh, machining, um, such that I think we're probably gonna have to go rebranded as something a little more charismatic.
I personally like electro milling, uh, but that's just me. Um, yeah, the capability is there are nuts. I mean, the, the ability to essentially take like a CoreXY Voron, or equivalent, which is one of the projects that we're drawing a lot of inspiration from for, for, uh, you know, our, our initial, uh, you know, for our initial chassis, um, yeah, the ability to take that swap, uh, swap it tool head on there, swap the build surface for a, for a build, for build tank or for, uh, you know, uh, for electro milling tank, and then just, you know, the only real unsolved problem is, uh, right now is the workholding aspect of that.
Um, you know, the rest of it needs to kind of just be a, uh, you know, there's, there's a lot of, there's a lot of programming that needs to happen for that, uh, you know, for that to integrate with these various different systems. But, um, there are enough people working on it and working quickly on it that, uh, you're, you're gonna see, um, I think within 18 months, uh, systems that can do commercial stuff at scale. And then, and that means you're taking a significant fraction of the capabilities of a, of a, of a $50,000 to $150,000 six-axis rig and
Knocking that down to $5,000 and then to $2,500 and then to $1,500 and then the whole built environment changes.

We need to think we can do the impossible, to, well, do the impossible and let technological breakthroughs happen-- THings like Can electro discharge machining bring us closer to the dream of self-replicating industrial machines? Go out there and do it.

So, um, yeah, that's I think the, honestly, yeah, electrical discharge machining is the single most under-appreciated manufacturing technology out there right now. Um, it has the most out of whack ratio of potential to interest of anything I've seen. I mean, it's, uh, the, the, the amount of, um, things you can do on the other end of maturing that technology are like at least if two orders of magnitude more impressive than anything that's ever come out of FDM.
That's what's so cool about all these low force processes. Um, as soon as you can automate this stuff into many thousands of little cells with swappable tool heads and swappable build, build volumes, um, running on largely the same robotics with, uh, you know, with a ton of shared software, um, we, you know, we end up in a, uh, you know, we ended up in a world that is wildly different from the one we have right now.
I mean, um, cause, cause that's, you know, that, that, that set of capabilities is going to make it that much easier to make more machines that can do that. You know, and it's gonna, it's gonna make 'em, uh, fan out fan out and iteration on, uh, you know, on those, on those machines, uh, that much quicker.
ALEX:
I think the, you know, I've done a lot of thinking around self replicators, uh, just for fun. Just because it's like the funniest thing I could think about to think about. Um, and, um, it's shocking how much you can, how far you can go with it. And a lot of it is about giving up the kinds of things that don't like, that are the way they are because of things are the way they are.
For instance, like, you know, we care about like the weight of things. Why? Because we need to ship it very far. Why? Because we make it all in one place, you know?
Uh, but if we didn't make it all in one place, but it was within the continent, so it could be railed into wherever you wanted it, then you wouldn't care about the weight. And then all of a sudden, a lot of things, you know, things go away or we care about reliability to a crazy degree, but if you have a machine that can make its successor, right, maybe you can meet a lot fewer repetitions in it because “Hey, when it reaches 10% of its, of its useful life, it'll just make it successor itself to whatever, right?”
Like the one master capability of replication, um, can pay for a lot of flaws. I'm talking like lethal flaws today. Right? So, so, but, but so a lot of the technologies there, um, but it's not being unified. It's not being thought of that way because it's not, it doesn't make any sense unless you have the full cycle of self-replication.
Otherwise, if you need one part shipped, you're screwed. Right? So, or something, even though, I mean, I'm sure that you can do some things with one part shipped, but there's, you know, once you reach zero, it's a whole other ball game.
MATT:
Or, if the parts that you need, you know, the, the, uh, “Adrian Boyer's vitamins,” uh, know the RepRap, vitamins, you know, the technical nutrients that you can't, uh, you know, that you can't re, uh, generate yourself. Um, if those are really low mass...
If you have a general purpose, if you have a general purpose computer, or a shitload of general purpose computers, and you can just send rockets full of general purpose computers, that's the one thing that you can't make in, you know, in, in the Jovian system, um, you can, and you can make pretty much everything else in the Jovian system. Um, yeah, you're, you're, you know, you're reliant on a lot fewer, uh, uh, on a lot fewer things.
And, uh, as far as the, the, you know, like weight constraints and everything else like that, uh, is concerned, um, you know, all this just gets easier when we're not, when we're, when we're talking about out of the gravity well, like, you know, not even a decade or two from now, right? You know, um, once we're out of the gravity well, and you know, it is in is, and it is relatively inexpensive and also relatively simple to just plot a trajectory, send a bot to go capture an asteroid, and then, and then chew it up as it's bringing it back, then all of the, all of the, uh, yeah, all of your issues just end up getting way simpler.
ALEX:
I think that that will require a culture that I'm coming around to your idea, um, culture, uh, that will require a culture that, um, is acclimated to the development of new technology because honestly, my number one problem, like literally everywhere is convincing people that crazy stuff can be done. Like, literally, I feel like I'm, uh, most of my work at this point is to keep repeating the same point, which is like, yes, crazy things are possible. No, just because it hasn't been done doesn't mean it can't be done. It just means it hasn't been done. Um, because everything that was done was done well at one point for the first time, like, you know, this is not difficult to, to reason through, right. Like things that happen have happened and therefore things can happen.
You know, it's like, we're not talking about [inaudible] here, but somehow we, I think there's like this overt, like, I don't know, maybe it has to do again with like the distancing from the manufacturing, uh, spirit, but like, there's this weird sort of deification of the, the makers of things that sort of, they know, and nobody else does, uh, because they don't have the right DACs or whatever, and therefore it's not even worth trying to do stuff.
Um, I always have my stupid, uh, you know, like made by, by my sort of copper plated PCB right next to me to remember the, you know, this was thought to be impossible. And I was like, what I have to lose? Like, can I copper plate uh, conductive filament. And it was like, holy shit it works.
Like I was like, I, you know, I was, I went step by step waiting to find the barrier. I was like, all right, this, this, maybe this one. And I was like, I have no idea what I'm doing. I was like, seriously, I've never done anything like this in my life. And I was just like, oh, happy go. Lucky. You know, it's what you were saying that before, which is like, as soon as you step out of the gravity, well, of, of, of modern technology, because all this low-hanging fruit, it's embarrassing, but it's like, I should not be allowed to be the person that finds this, but now I have no business.
(laughter)
It's there it's so much, I mean, if, if, if people are listening to us and take one thing away, it's like, get out there guys. Like, there's so much stuff that you can and combine, it's almost like the whole, like physical world has been abandoned too. Like everybody's gathered online, like two or three, like water wells, and like drinking from that. And everything else has been like, “Hey, no idea. It's never been done before. So I guess it can't be done.”
MATT:
Yeah. Yeah. It's, it's, uh, it's, it's a, it's a huge problem. I think it's a big issue with the, you know, with, with like the, not just the capital allocation system, but also at the talent allocation system. I think a lot of people are, like, certain ideas about how things are done or pounded into people's heads in university. And, um, and even before that, and it's really…
ALEX:
But, Matt, do you have a PhD to say that,?
MATT:
Right, exactly. (laughter)
ALEX:
Are you qualified to have such an opinion, sir?
(laughter)
MATT:
Pardon me, sir. How dare you critique this thing? (mockingly)
Uh, you, you, you, in education, you need to be in the education field to speak quite credibly about this. Um, uh, the, uh, uh, yeah, I mean, the, I think escaping that epistemic, um, trap is, is among the most critical cultural tasks of, you know, of the next of the next couple of decades. Uh, because if we do continue to restrict technological advancement to a priesthood of people with PhDs and, and PI standing, and labs and, and, and traditional funding sources, we're in big civilizational trouble because we're reaching the inflection points of a ton of different global trends that, um, you know, that absent technological advancement will probably kill us.
ALEX:
Scientism is the biggest threat to science.
You know, I love writing these threads on Twitter, like that are sort of, you know, kind of anti-academic or whatever. And somebody is like, well, you don't have a PhD. So I was like, well, actually I do. So… f*** off.
I don't put it, don't put it on my, on my, on my bio, right? So I could possibly have one if I'm not like putting it in your, so, you know, it's like, look me up on Google scholar. I probably have more citations than your professors. So I don't know, just like… go away.
Speaker 12: (02:10:41)
Balaji… Balaji would deal with that. I think a little bit towards the beginning, apologies, Balaji Srinivasan. For those of you who aren’t familiar with him, is this guy who's mostly famous for working in the crypto world, but his academic pedigree is in, is in, uh, genetics and, uh, you know, uh, biology and, uh, he's, uh, he's, he's, uh, among the most qualified people in the world to speak on that top, on that subject.
Uh, he just doesn't flaunt it and the guy would get into these slap fights with people, uh, you know, with people with a master's in public health, who with very strong opinions on how, on, how we should and how we should, uh, on how we should, uh, uh, worry about the flu circa March of 2020, would consistently win these arguments with these people. Uh, if the thing that you were looking for was, uh, you know, was, was actual rigor of thought. Uh, but if the thing you were looking for was, uh, you know, authority to comment on public health matters, uh, you know, he, he didn't, he didn't have that.
And, uh, we were, you know, we all suffered as a result. Um, uh, yeah. Oh, scientism as a scientism, credentialism, and just the general idea that, uh, the things ought to be restricted. Ivory towers is cancerous.
ALEX:
I, I seriously, I mean, I haven't really thought of it that way, but I almost feel like, yeah, I need to redouble my efforts, uh, to sort of fighting against that one specific thing. I mean, it's, I've seen it so much. I see it in the stuff that I'm getting into. You may be aware of it. Maybe I'll, I won't say too much, it just not to drag us out of this gravity, whatever, but, uh, you know, it's just like, um, people with like credentials just absolutely continually just attacking you, but it's like, yeah.
There's like the field that is the most sort of under attack right now with the virus. It's like a hundred people and their students and their mentors, you know, and their funders, it's like, yeah, you all agree? We know, because like you, can't not because like, and then when everybody sort of criticizes it, it's like, oh, you know, but you don't have, you know, in the field, I'm not in like virology, but in, in, in genetics.
And yes, I can read a stupid study and tell you that it doesn't make any sense. You know, it's not, I don't need like your hyper-specialized, um, degree to, to tell you that what you're saying doesn't make any sense. Right. Um, but, um, the, the, the, the gatekeeping is, is fractal. You know, it's not like in science, out of science.
No, no, no. It's like in this particular field of science, and then it's like, yeah, but you're a virologist, but your thesis is in ebola and, not the coronavirus is therefore, um, you know, you don't, you don't have standing. And it's like, okay. I mean, anyway...
Yeah. It's, it's, it's, um, we gotta do something about that. I think I hadn't really thought it through, uh, this is kind of drawing for me right now, how important, um, it is to I call... I wrote it and it didn't even click for me, but now it all makes sense. I called, uh, you know, siloization as firebreaks, but for, for, for human flourishing, you know, like, yeah, right.
You know, we have one good idea show up in one field and it doesn't spread because nobody in that field talks to somebody in the other field. So it's like, okay, well, good. We got it in our little fields. We got this good idea. And maybe somebody will take it somewhere else for a hundred years.
MATT:
Yeah. My hope is that my hope is the preprint servers alone are enough to push us out of the, at least the worst parts of that. Um, you know, I wouldn't have the level of CS knowledge and capability that I do without things like archive, and without the ability to watch the, you know, watch the research frontier move in real time.
Um, I think that the accessibility of, of researchers has gotten, uh, categorically better ever since the internet emerged. Um, you know, you can actually find people when you need to, um, but there's still this institutional structure, you know, there's still the fact that, uh, even though I'm a taxpayer in the state of Michigan, I still need a University of Michigan ID to go walk in and do anything in their library, let alone their letter, let alone their materials, science lab.
And that's a, you know, it's, it's, it's frustrating. It would be nice to know it'd be, it would be nice to, uh, move these things past the, you know, sort of cult/guilt phase and into something that's a little more accessible. But, uh, you know, I, I hate to say it, monasteries or Lindy?
(laughter)

Which would you give up: GitHub or all scientific papers ever?

ALEX:
I have a question for Dival. Um, you have to, if you had to choose which one to give up, um, all scientific papers ever or GitHub, uh, which would it be?
DIVAL:
Hmmm… that's tough.
ALEX:
(laughter) I don’t have an easy answer. Omega shows up to you, and he's like, “I'm going to destroy either every scientific publisher and all accountant, or, you know, GitHub and every open source project.”
DIVAL:
Right. Um, I mean, I think, uh, do I have to keep the ones I already downloaded?
(laughter)
If I already have them printed, we can't delete them right? Yeah. Uh, I think GitHub is more important. Yeah.
I mean, maybe it's a good thing for all the, the papers to get deleted, right?
(laughter)
MATT:
A significant fraction of papers are on GitHub anyways right?
DIVAL:
Half of them are bulls***, probably more.
MATT:
The ones, the ones where you can actually go run the code and reproduce it, they're all in GitHub. The ones where you can't run the code and reproduce it, those are, those are in the, a little, The Lancet.
ALEX:
Right, right.
DIVAL:
Yeah. Yeah. And maybe it's a good thing to... this is the way you nuke the peer review system.
ALEX:
Oh my God. I can't, I can't even, what the new gatekeepers that are gonna emerge out of GitHub. I just noticed this, this ability for systems to just reproduce themselves and you sort of lands it's, it's, it's, uh, it's fascinating. It, uh, yeah. It's, it's um, yeah. You know you become the revolution and then you win and then you get busy rebuilding the thing that you're rebelled against.
DIVAL:
Yeah. Yeah. It's pretty common in Silicon valley, anti-academic, but then we have YCombinator and like On Deck and like, no, we created our own credentials. I guess they're better than like the previous thing, but we didn't get rid of credentials.
ALEX:
Silicon valley is amazing because it's like, anti-regulation, unless the regulation is regulating us into being locked in everywhere, in which case uh, that's fine. Uh, I mean, I think, you know, or, or, or Google being anti-patent right until it wasn't, um, and whatnot, like it's just changing the side of the system you're on changes how you feel about the system, I guess.
DIVAL:
Right, right. Yeah. I mean, I think what's interesting is that we're seeing this huge erosion of religion, especially among like the among gen Z people, but then at the same time, like spirituality, uh, astrology is becoming extremely popular as, as a result. So it's like, we didn't get rid of that feeling, it just went somewhere else.
ALEX:
Matt. Well, while you were gone, we changed completely the topic into religion. Sorry, we didn't, we didn't ask you about this, but, um, it's, it is fascinating that, I mean, this was actually a big, a big question in the new atheist circles. Right? Like, do we need another thing or is it like, is the whole concept of religion bad, or is the particular systems bad, or was bad being sort of let's call them unnecessary?
Like, is it, um, is it something that is fundamentally needed? Or something that is, um, yeah... So astrology, like, I don't know, kids these days. Uh, I guess Dival, you’re the relevant authority-- is astrology on the comeback?
DIVAL:
Yeah. It's very popular now.
ALEX:
I have no words. Can we take it and use it somehow to promote spacefaring?
MATT:
You can use it to promote crystal healing...and… uh...
ALEX:
Not exactly where I was going with this, but, you know...
MATT:
It's really bad. It's, it's, uh, you know, I'm about 10 years older than you are Dival, and and I think about 10 years younger than you are Alex, so yeah. Um, uh, it is, it's proliferated it up into like my bracket of millennials, for sure.
Um, there's a lot of, a lot of people who do actually make day-to-day decisions on the basis of what their stupid f** star chart says. In fact, I've started to see wars between fundamentalist, Christians and, uh, like wicca people begin to proliferate on Instagram. Uh, there was a, there was a very, very funny, uh, cat fight that happened, uh, some weeks ago, uh, that I was witness to between a group of people in the comments section who are convinced that, uh, astrology is this concrete and useful thing in their day-to-day lives. And people were convinced that, uh, astrology is bad because it's demonic possession.
ALEX:
Right. Right. Of course it was like, I was, I was, I was hearing the, just to, I mean, if you're going to get canceled, might as well do it properly.
Uh, I was hearing the, the, the, the, um, the, the part that the podcast between, uh, Coleman Hughes and, um, I don't remember, uh, Charles Murray, right. That's the one. And I was hearing them talk, which was like a fantastic conversation, by the way. I really, really enjoyed it. Uh, I feel Coleman is really coming into his own and is getting quite more assertive, which is like, I was just like, I haven't heard a conversation that satisfying in a while, but there was a point in that conversation where they were like, you know, um, Coleman was sort of telling Murray that, like, the stuff that Marie's talking about should not be talked about because it has certain negative, um, potential effects.
And then Murray was telling, uh, Coleman that, um, it should be talked about because the alternative is worse. And I was just kinda sitting there screaming at the screen. It's like, “No, truth is truth! We should actually just, it doesn't matter what we do. We don't to make a consequentialist argument just say what is true.”
If it's true, it's true. And we should say it. And if it's not true, we should not say it. And I don't know why you're trying to contort an argument about what, you know, that, that there is a good, that there's something else that's better or what, but it was like watching too, you know, the blind leading the blind, basically down the, down a cliff. And I was just like sitting there, just kind of like, why is nobody saying the obvious? You guys know what you’re talking about? Oh, it was, it was so frustrating.

On the dangers of creating and fixating on information hazards, how can it limit technological development? Can the hazards exist in an environment where information is plentifully available?

MATT:
There are very few functional downsides to having a higher definition view of reality than other people. Yeah.
ALEX:
The true centralization that you can have is reality. Right? That is the one thing that we all share that is replicable. And I can send you over there and I go, go say like, do this experiment and take the answer and use it. Right? And I know what it's going to be because I've done the same experiment that I don't need to send you the answer.
I can just refer you to the source and you can figure it out. Like everything else needs to be constructed and supported artificially. And therefore is going to be a huge drain to like, oh no, no, we don't talk about that. We don't. And if I bet you, if you look into the past for knowledge, that was not, um, uh, sort of widely available, um, and what people were saying about what would happen is that knowledge to become widely available, was at nothing to do with what actually happened when it did become widely available.
You know, like, um, how, like encryption was not, you know, how, you know, how, uh, the UK, uh, gave like, enigma machines to all its colonies to like, if their communications, because they knew that they could just crack it like trivially. Right?
So that, that whole ruse, because the whole Bletchley Park, what they had done, there was not well known until the seventies pretty much. Um, they were just giving out their African colonies, like, and they come up, you know, like, but okay, then communication became a thing. And then all of a sudden you have e-banking right? With, with encryption.
It's like, I don't think people were thinking about e-banking when they're like, should we, you know, make, you know, encryption something that people can use casually or not like they had, they were like, oh, will our opponents be able to share messages that we can crack, but not in banking. Right. So, so, so, so yeah, like that, I find it so arrogant when people are like, oh no, if you, you know, this conversation makes it out to the plebs, like they're gonna, uh, you know, do this and this it's like, that's one thing that you can think about and maybe it'll happen, maybe not, but also infinitely other things. So, I don't know.
MATT:
Yeah. Idea the idea of info hazards, I think, is a very frustrating and, and, and, uh, uh, counter productive development. I think there are a lot of people who do genuinely believe in info hazards. They think that, uh, they think that there are elements of reality that you are better off not knowing. And I think that the, that leaves the, that leaves a question unexamined, which is why are you better off not knowing that?
Is because you have a lesser degree of fitness for purpose in reality, as a result, because that's its own issue, you know?
ALEX:
To me, like, it's just like, we should just give up on the idea that certain things can be restrained. But, um, now that you mentioned it as info hazards, um, I'm reminded that I should probably do the work to, to prove conclusively one way or another, if it's a thing, because that is the beast that must be defeated, right? The basilisk that must be taken down....because that is the valid argument.
I guess the, the fully, the fully formed, uh, version of this argument is info hazards, right? So do they exist or not?
And intuitively, I feel like you, even if they do, how are you going to constrain them? Like, I don't understand what the idea is. Right? It's I guess not, I'm just going to assume that this is an argument near and dear to your heart, but it's like guns basically. Like you restrict them, you keep them out of the guns of good people. Not just out of the hands of bad people.
MATT:
I've been asked to speak less on that subject, but, um…
ALEX:
Okay, so you are, um, yeah, I mean, I'm not into that world. You're just my stand-in, smart person who has interesting ideas about that. So I'm just putting you in a tough spot, but you can also change the subject.
MATT:
No, I'll, I'll, I'll comment, I'll comment, I'll comment, obliquely and vaguely on it. Uh, uh, I mean, ultimately, knowledge of how to act appropriately in a situation requires you to use less firepower to an effect, to affect a change on that situation. And this is true in a military sense. It's true in a markets sense. It's true in a lot of different senses.
Um, when you have a greater level of knowledge about something, um, that actually trades off with, with the amount of destruction you need to, to cause to effect a change in that thing, or the amount of construction you have to do to build a lever that moves it, um, you know, in a more elegant way. And I think that, uh, I think that restrictionism around this sort of stuff is very, very fundamentally misguided.
It doesn't take into account the fact that, um, that while humans are not always rational, they tend to be more rational when they have better information about things, right? Human beings have a high degree of bounded rationality that is, um, you know, that that is largely a function of how aware they are of whether or not something that's going to affect them in a negative or positive way. And, uh, and how immediate it'll do that.
Um, so the more information you proliferate, the more, uh, the more rational people's decisions will, are likely to become in aggregate. You'll have outliers, you'll have exceptions to that, but in aggregate, that's then what has happened throughout history. Extrapolate my view on firearms from the that and you have a general idea.
ALEX:
I guess the problem, what I feel a lot of the time is a problem for people is that they are not willing to go through an intermediate period where the knowledge is spreading, uh, and, and the asymmetries that are available to people during that time. They're like, “oh, you know, this sucks.”
Like, I mean, you know, I don't have actually a strong opinion on Brexit, but like the worst kind of argument for Brexit is like, “it's going to be uncomfortable for a little while.” It's like, okay. But if they are right, then that doesn't matter. And if they're wrong, then that doesn't matter. So why are you making that argument?
You know, like, it's, it's kind of like the uh, you know, at balena for instance, I tend to worry that... ...the problem that worries me the least is a problem that gets smaller over time.
It could be pretty bad. Right. But if it's something that, like, the dynamic of it is the sort of thing that like, it will, it will self-correct, and it will not kill us, I'm like, okay. We have an old version of the operating system out there that is getting replaced every day, that doesn't work very well, and it's causing high support burden. Am I going to worry too much about it? No, because all I have to do is wait, and then it'll get replaced, and then I don't have a problem.
If I have a new version of the operating system, that's kind of causing the high support volume, then, you know, I need to run as fast as humanly possible to get that thing out of production. But if it's an old version, you know, whatever, it'll, it'll burn itself out, we'll be fine if we're not dead yet, if we made it this far, we're probably gonna make it to the next level, you know, like that, that sort of thing. And, and, um, I think people are not patient with that kind of like, we don't have the rationality training as a civilization. I don't know where to ascribe this to... but yeah, we can't just say like, well, this thing? You can wait it out. You're fine.
MATT:
Well, so, and I think that this is, I think that this is a huge issue when it comes to all of that sort of stuff that people like to restrict, whether it's, you know, whether it's, um, things that can propel objects rapidly over distances, uh, or whether it's, uh, uh-- and I'm talking about, I'm talking about engines in this, in this context. Um, uh, rockets and all sorts of other stuff like that. You're the ability to send something at high speed from point A to point B, uh, is something that, as it was proliferating, caused all sorts of imbalances. And once it proliferated, reached, uh, a, uh, reached an equilibrium, that was the case with the proliferation of firearms. That was the case with the proliferation of trains, rockets, you know, what, what have you.
And I think that, that also applies, um, you know, that applies, uh, in sort of, sort of a, a metaphorical sense also to information. It applies also to, um, you know, models of analyzing information.
Uh, you know, when you, you know, when you, uh, start to send that sort of thing out there, you create a competitive environment in which, uh, you know, in, in which the adoption of the, uh, of the new thing becomes necessary to survive. Um, and I think that that is a virtuous constraint, uh, to put on human civilization. We need, we may, you know, to call back to our earlier discussion about warfare, we might not need warfare but we need to, we do need some sort of virtuous constraint, whether it's, uh, you know, whether it's competitive or existential. Without a tiger chasing us, we have no reason to run faster.
ALEX:
Dival, this as your time, man, we've all stopped talking. This is your time. I'm Greek. I can just jump in, you know what I mean?
(laughter)
DIVAL:
Every time you guys talk, I'm just learning and processing. And then like, I take a bunch of threads that are... I'm super ADHD. So it's like, uh, yeah. But, um...

Are we on a slippery slope toward a world of maximizing pleasure and reducing pain and how does this inhibit technological advancement?

I do think that space is going to be a forcing function for improving technology. Like that's pretty clear, and it has been pretty clear, um, like with, I remember, like, I always see this like stupid Luddite argument on Twitter from time to time about like, uh, “oh, why are we putting resources in this space when there's so many problems on the ground?” And it's just like the most, like, it always drives me insane when I see that.
I mean, those people are just people that will just be ignored…
ALEX:
But, uh, like, well, yeah, but sometimes they get, um, you know, the Greta Thunbergs of the world get that, uh, coverage. Right. And, and, and to people, uh, to make arguments and that, um, I mean, you, you want to believe that it'll all wash away and not matter, but like, you know, it's hard to sometimes some days it's easier than others.
DIVAL:
Yeah, yeah, yeah. Maybe we can keep those people in VR land.
ALEX:
And that's the other thing we've kind of hinted at this a couple of times it was like, to me, like, what I notice with AI is, but honestly, with people as well, is that as soon as we have an opportunity, what we choose to do is Wirehead basically, like, what is the easiest way to like, um, for, for, for the listeners, this is actually a pretty, uh, unusual concept.
It's the idea that if you could take your pleasure center and just sort of connect it to at will, uh, satisfaction, right? Like you just find whatever is making your brain feel happy and you connect it to a button you're pressing the button basically. Um, uh, you know, that it feels like we're inching into that. Like as much as we can. And that's, I guess “VR land” in a way.
Um, but intuitively makes sense that like, that cannot, it can only go so far right. Until somebody is going to keep the lights on. So I remember Elon, uh, talking on [Joe] Rogan, and he's like, “if we don't make stuff, there is no stuff.” Like, he was really upset about this. And it felt like he was coming from somewhere deep where, you know, he's been, if he feels like he's talking to people that don't understand that, like things happen in order for things to exist, things have to be made and somebody's got to make them and, you know, you know, like make stuff. And that means that that stuff will be made. And it's sort of like, you know, it was a very extremely basic argument that, that this was not coming across to whoever.
MATT:
Well. Right. I mean, that's, and that's, that's, that's, uh, that's an important question about the future of humanity, like is our, is our purpose to, to maximize pleasure and minimize pain, or is our purpose to, uh, do something different from that? Yeah. Is our purpose to, you know, as our purpose to continue propagating Terragen life. I mean, I don't, that's above my pay grade, you know, coming up with a robust answer to that is above my pay grade by quite, you know, by, by a couple of rungs. But, um, I already know what the answer to that for me is, and I intend to execute on that, so…
ALEX:
Yeah, there was a, there was a, this thing shared on Twitter about like the top 10 things to do to be happy. And I was like, “that's amazing. I do none of these things.”
(laughter)
MATT:
Manage your sleep cycle?
ALEX:
Yeah. Be nice to people...
Um, well, you know, eat well, exercise. I'm like, oh my God, this is, this is getting worse and worse. Um, the, the, uh, but I remember having made a, like a conscious choice that like, if it comes down between happiness and, you know, uh, making progress, I will consciously choose progress.
The weird thing is I'm pretty happy, in retrospect. And I think happiness is the sort of thing that ultimately my lesson from that is happiness is this sort of thing you can't, um, consciously pursue. If you pursue it the way we are, or maybe “VR land” with solve that, or maybe it won't actually, which is interesting. Um, but you know, it's the sort of thing that you can't get up in the morning and say like, I will be happy today. Right? It's the sort of thing that, that, that you get through not wanting, not chasing after it.
Um, in a way, if you look at the richest people in the world today, they did not, um, take a money-maximizing strategy in their lives. Uh, necessarily they didn't say like, oh, you know, it wasn't like Gates said, like, what is the most, the fastest way to make money? And as a kid in the 80s, he was like, microcomputers right. Basic compilers.
He just did a thing he thought was like, awesome. And that ended up, you know, linking into something that was extremely valuable. Uh, but you, you don't… yeah...I find that several extremely valuable things that cannot be perceived directly, which is maybe this salvation to the Wireheading problem because yeah. It, it, it doesn't, doesn't look good for humanity.
DIVAL:
Going back to the space thing, like, I think like, you know, colonizing spaces, like three, great for Earth, like, you know, like you, you want a place to test like all your terraforming technology, right? You don't want to use Earth for that. Like Mars is going to be perfect.
Um, so I think like, it's probably like you want to solve climate change, you go to Mars, like that's, that's the answer.
ALEX:
I mean, honestly, I feel like Bezos doesn't get what he deserves in terms of recognition for his visions of like, you know, the Earth being zoned for light industrial. That is what is it called? Like, um, that's his, his thing with the O'Neill cylinders and whatnot, where he wants to sort of create these, uh, these off-world worlds, I guess, and then the Earth retained for, you know, its ecosystem, which is going to be, you know, let's face it, it's going to be hard to replicate the Earth's ecosystem.
MATT:
In terms of a counterpoint to Kazinksy, you could do a lot worse than Gerard K, O’Neill, uh, uh, the High Frontier, um, which comes with, uh, lots of cool pictures of things. Like what Alex just described, uh, you know, O'Neill Cylinders, Bernal Spheres, lots of other playgrounds for mankind.
DIVAL:
What's actually interesting is that Bezos actually studied directly under O'Neill.
MATT:
Really? At Princeton?
DIVAL:
Yeah. Yeah.
ALEX:
Whoa. Is this like a long con basically from Bezos? Is that, is he like the mastermind we didn't know we needed?
DVIAL:
Bezos was very interested in space travel before Amazon. It's pretty interesting to think about.
ALEX:
Does anybody know why Blue Origin is looking like they're doing nothing? Are they really doing nothing or are they doing a ton of stuff, but they're not telling anybody about,?
DIVAL:
I honestly don't know. I don't know.
MATT:
I think, I think it's possible that all the best engineers got scooped up by space X and are running in that they're running in that verdict and all that's entirely possible that that's the case. Their, uh, their, um, moon launch or their moon access system that they proposed was just a fundamentally worse proposal than what starship is capable of.
ALEX:
They're getting a lot of flack from a lot of those space YouTubers, because yeah. They're like, yeah, what you propose is fundamentally unsafe. Um, like not only you will not win, you should not win, uh, sort of like, you know, anger, uh, which, you know, you never really see that community is always very open and friendly when except those guys.
MATT:
When you’ve made Scott Manley mad, you've, you've, you've really gone down a very, very, very, a very dark path.
(laughter)

What are the go-to YouTube channels for people who want to self-teach mechanical engineering, science, and how the world generally works?

ALEX:
Um, okay. I think, um, maybe we wrap it up, but before we do, um, maybe, I don't know, like, you know, there's a standard question that people ask and I don't know which one to, uh, settle on, but I'll try this one. Uh, what are some YouTube channels that everybody should be watching?
MATT:
As Andrew was saying in the chat here, the balena YouTube channel.
ALEX:
Not an acceptable answer!
MATT:
Grant Sanderson's 3Blue1Brown is reinventing math education. I give him 15 bucks every time he posts, you should too. Um, I think that's excellent.
The group of people, uh, Wendover Productions, PolyMatter um, and a few others, uh, all have kind of a, uh, internet rationalist, um, view of current events. And I think if you want to replace reading The Economist with something that is both more informationally dense, and, uh, maybe a little more keyed into, um, you know, ground truth, uh, watching an hour's worth of those YouTube videos once every two weeks, uh, you could do a lot worse than that.
Um, I'm also a huge fan of, uh, I mean, I, I honestly, honestly, um, I know I mentioned earlier, but people like Scott Manley, just individually individual, uh, people who categorize or who collect, categorize, uh, distill lots of information and then distribute it well, um, those folks are, those folks are, uh, in whatever niche stand up and usually the best sources of information.
And, uh, they're either on YouTube or on Twitter, most of the time. That would be my answer.
ALEX:
We gave Dival quite a bit of time to…
DIVAL:
Yeah. Um, there's one YouTube channel. I really like it's called The Efficient Engineer. Um, they make like videos on, uh, like they go really into detail on the theory of like material science, like you might, they do. They had an excellent video on the stress strain curve. Um, they go into explaining like how FDA works and like, understanding like tortion. They do videos basically on fluid dynamics and, uh, solid mechanics mainly. Um, but it's like really well animated, very well explained, very concise. And I found that that was very helpful for me. Um, especially like, if you're trying to self teach yourself like mechanical engineering, like a lot of the theory, like it makes it digestible. Um, I'd say it's almost like Grant Sanderson-like, like the equivalent of that for like, uh, materials engineering and that kind of thing.
Um, and then the other one I like is Teaching Tech. Um, he does a lot of videos on like making 3d printers. I saw a video recently he did on Klipper. Um, a lot of it's like beginner stuff, but like, I guess if someone's listening to this, they might enjoy that channel a lot, especially if they're enjoying that channel. So I found that both of them are very helpful.
MATT:
Teching Tech does a great job of distilling stuff. That's an awesome channel.
ALEX:
There's, there's, there's, uh, I think in general, the 3D printing YouTube community is just completely insane. Like so many people that I've, I've learned so much, because again, like, I'm very much closer to that beginner. I've watched a lot of Teaching Tech, and, you know, Thomas Lander and like, just so many of those guys who just like, they'll just take the topic and they'll just break it down to bits right? For you.
Um, uh, and just like you end up knowing the thing, like they don't, I don't know, there's like a weird sort of meritocracy on YouTube that is rare elsewhere, because these people live and die by the views. So they can't, you know, fake it, uh, not for too long anyway.
Um, like, um, oh yeah. What was the, um, Thought Emporium I saw, have you guys seen that guy? Um, he did one thing where he literally grew human neurons and connected them to a circuit, uh, to do something. And it was like, I was like, okay, this does qualify as a maker project. I guess.
(laughter)
I, you know, the aesthetic is all the same, but the topic is, um, we out there...
MATT:
Synthetic spider silk using yeast?
ALEX:
Oh, wow. Yeah. I'm telling you, uh, that, that channel is just one of the most, uh, one of the most insane things I've ever seen. Uh, it's just, um, everything, all of it is just madness. Yeah. But it's, it's, it's, it's broadly a part of the same circle, but also just brings a twist that is just... it’s the thing that we were talking about. Basically, he goes out into the wild and it brings back things that you're like, okay. That, like, I consider myself as somebody who like goes out there and that is out there for me.
MATT:
That's awesome.
ALEX:
All right, folks, I think we've made it, uh, we, we, we covered it. We covered so many topics. I don't even know what the title is going to be about. Uh, I, I, I'm going to be thinking about this conversation for a while. Uh, this is fantastic.
MATT:
Yeah. It was a ton of fun. Thanks for bringing us together. I, uh, I can think of far worse ways to spend Friday evenings than, uh, than, than, uh, shooting the s*** with a couple of guys who were smart as you two. Seriously, this is awesome.
DIVAL:
Yeah. Yeah. Definitely learned a lot. I'm going to listen to this a couple of times to process it all.
(laughter)
ALEX:
Alright, thanks guys.
--- END EPISODE ---
by Andrew NhemAndrew is the Content Strategy Lead / Product Builder at balena, and enjoys tinkering on web content, building stuff, music, hydroponics, and homeschooling

Share this post