Embedded - 322: Learn Assembly Code

Episode Date: February 28, 2020

Ramiro Montes De Oca spoke with us about modular electronics, chiplets, and his company aThing.io athing.io Chiplets Project Tinkertoy (movie) is a 1953 US Navy project on automated manufacturing of modular electronics. Ramiro mentioned his accelerator: CoFoundersLab Accelerator

Welcome to Embedded. I'm Alicia White alongside Christopher White. We're going to talk today with Ramiro Montes de Oca about modular electronics. Hi, Ramiro. Welcome. Hi, Christopher. Hi, Alicia. It's an honor to be here. I'm a big fan of the podcast, so thank you for having me. We're excited to talk to you. Could you tell us a bit about yourself? out from mechanical engineer to start what became at that time Web 1.0 at the end of the 90s.

So basically I did a lot of coding. And when the bubble burst, I came to US. And I started working on broadcast engineer at that time. And then I slowly switched to doing some hardware engineering and broadcasting to finally switch to consumer electronics and manufacturing in China, basically. And now you're the CEO and founder of a small company. Yeah, it's called A-Thing.io. It's motionless electronics.

Which we'll be talking more about. But first, lightning round. Are you ready? Yeah. I love this part. Okay. What's your favorite component ever? Oh.

Resistors. Favorite processing core? I work a lot with ARM and microchips. So, basically going both ways. LEGO blocks for Tinker Toys. Yeah, LEGOs.

Smallest circuit that is big enough to be interesting. Smallest. Nine deep of freedom, like, yeah, accelerometer, gyroscope. Is there a place in Argentina that everyone should visit? Oh, let me think. Mendoza. What is Mendoza?

Mendoza is on the west side of Argentina. It's where the big mountains are. There's a lot of wine over there. It's similar to California, actually. If you could teach a college course, what would you want to teach? So I love manufacturing. So probably a subset of lean manufacturing, which is manufacturing on demand. I really think that that's the future of manufacturing.

And do you have a tip everyone should know? Learn assembly code. Really? Oh, wow. Yeah, one day it's going to save your life. Okay, so we mentioned a thing.io, and it's modular electronics. But what, that could mean anything. What is it really?

What does it do? Who is it for? Yeah, so a thing is like Lego for electronics. It's not a toy. It's a professional electronics for rapid prototyping and rapid manufacturing, too. So basically, we separate each part of the system into modules. They're the design of a penny. And they can be assembled by hand, so there is no need to solder.

And so we have processors, power management, memory, antennas, all kinds of sensors, peripherals. So all the modules are pre-certified and ready for market. We can talk about that later and so another way to explain is like arduino but on the size of a penny tiny arduinos and you said that they they don't require soldering so they all have connectors that go between them, yeah, so they're connectors similar to, they're not similar, it's what we use for cell phones, basically. They're very small connectors,

not necessarily user-friendly, but the idea here is not to be a system that is for playing, it's more like a final product. So once it's connected up, you're not expected to reconfigure it and change things around? You know, the specs for those are a Hi-Ros connector. They're very popular.

And yeah, I mean, they say this like 50 times. But yeah. But who hasn't broken one of those? Exactly. It's how you handle it. So yeah, I wouldn't trust that. You said that it was for rapid prototyping, but then it sounded like you just said it was also for shipping products because of the cell phone connectors.

Yeah, so here's the deal. It's both. And the only difference between doing final products and rapid prototyping would be the connector. There is an adapter to just make it easier to connect or disconnect. But the final product, I mean, you can use the final product and actually ship in as a product. And to be fair, when we spoke over email,

I did say I wasn't, I'm not 100% on board with this idea because I'm accustomed to custom boards being lower power and cheaper and just generally more designed for their purpose. And so modular electronics, I'm like, well, that's going to be more expensive. It's going to be more power hungry it's probably going to be bigger can you convince me this is a good way to go yeah so yes yes yes and yes I mean you are right on that but so yeah

we pay a lot of attention on power management that's something that I'm not really worried about So we pay a lot of attention on power management. That's something that I'm not really worried about. But, I mean, in your case, rapid prototyping is going to be an easy sell. I mean, it's nothing different than playing with Arduino or Raspberry Pi. Okay? So I don't think it's going to be difficult to sell you on that side now the rapid manufacturing in your case is not going to be easy to sell i can give you one example how modular system

are making big noise and accelerating everything and one is a WordPress, for example. Okay. So WordPress. Wait a minute. WordPress? WordPress for web development. The blogging platform. It's not only that. Blogging, you can do a whole website.

It's a huge database that you can manipulate, right? I'm giving you a hard time. Our websites in WordPress, our business websites in WordPress. Sorry. So, I mean, I was a web developer and working on PHP. And I can tell you that what I can do in one day,

it might take me maybe one month on PHP, right? So, I mean, modular is a really good way to go. Obviously, there's going to be a trade-off, especially security, okay? If you don't update. update and but it's the same way that um you try you grow a company or a product right once you use that to make it grow then you're gonna go to the alternative basically okay um and i i was asked recently about the challenges I faced entering embedded systems.

And the challenge that hit me the hardest was the complete lack of dev boards. And so I can completely understand how modular electronics would be useful for prototyping and even small scale deployments. Right. One of the biggest problems with manufacturers sometimes is they don't release a developer report, right? And I remember back, I think it was the 9-dip of Freedom. It was an SD chip. I don't remember exactly the model number, but

when SparkFun released it, that chip became the most popular chip ever for 90 off. And I have the feeling at the same time manufacturers don't get it. That we need

to, like not everybody has a QFN, you know, you can't do a QFN on a breadboard, basically. Right. So, yeah, developer

boards are really important, and every time I go to conventions and whatnot, and talking to cheap manufacturers, the first thing I ask is, do you have a developer board? And, you know, you'd be surprised, but sometimes they don't. And that's odd now. I remember when having a developer board, I mean, that was borrowing one from some person you met who you probably took out for a

beer and they loaned you a $10,000 development board. Yeah. Or the FAA would give you one for a little while, but you wouldn't get to keep it. Yeah. And for some stupid little chip, right? It wasn't for a $10,000 chip. Oh, no, it for like accelerometers and gyros it was it was yeah but there is a difference between dev boards and modular electronics isn't there uh there isn't there's i mean it's a fine line at this point and, so we're going to talk about chiplets a little bit later now, but it comes down to how you take that technology that now is so small and applying to your project, right? So now we have better options, like smaller options, smaller packages.

And we had this huge developer board, which makes no sense. So basically what we did is just make it smaller. What about manufacturing quality? That's been one of the differences between dev boards and things that should be used in shipping products. In fact, some of the ST dev boards specifically say they should not be OEMed for products. Yeah, I remember your comment. I don't remember which episode, but yeah, exactly.

You want to buy 20 or 40 boards, and sometimes you can't, right? What we're doing basically is a system and package, okay? So the quality control is good. We work on silicon level, and we also work on chiplets, which is basically like QFNs or small packages. So, I mean, quality control is going to be good. And there's also

more requirements that we can over-engineer, like shielding that sometimes you don't need, but why not? And ESD, I don't know. So, not every board has all this stuff.

So what does this look like? If I come to you and say, okay, I have a product idea. Let's say I want to make a Bluetooth-connected accelerometer with a light, and I don't really care which microcontroller, but some microcontroller.

What would that look like? So at this point, everything that you mentioned, we have it. So it's not going to be a black box. So you're going to be able to actually replicate that on a PCB if you want to. Okay. So the same way you use developer words and then take it to your own project I mean you can do it

it's like bluetooth modules or wifi you have to if you want to take it to market you have to pay a lot of money on certification so this is going to be pre-certified and it's one way to use the system have to pay a lot of money on certification so this is going to be pre-certified and

it's one way to use the system and but you can keep going you can just you know uh use the power modules for example we have all this lt uh you can use webcam uh you already mentioned that bluetooth and wi-fi and processors and i mean, you just keep using it. So is there like a common bus that everything connects to? Okay. I should send you the common bus so you have a better idea. But yeah, we have from basically there are 45 pins available. And I can describe it a little bit.

Yeah, yeah, please. pins available. And I can describe it a little bit. So basically the main chip or the main module is going to be locally controlled by a chip that is going to be connected to I2C. So

with that chip you're going to do power management, you're going to do some reprogramming to the local module. And yeah, some other local functions. And basically that main chip is going to control the flow of the bus. It's going to tell you when to talk, when to shut up, basically. So if I look on your website, I see you have

kits, developer kits. And they're for different themed things. You have one that's a multi-tracker that has a GPS and a dog wearing a collar. Doesn't come with a dog. Doesn't come with a dog. Doesn't come with a dog.

And then a smart toy, a teddy bear with lots of pieces, and then a smart flashlight, which seems to have a speaker and Wi-Fi. So I guess that is a smart flashlight. None of my flashlights do that. The sun is out. Turn me off. Yeah. And so you have, and these show different modules. The teddy bear has a wifi, Bluetooth, a camera, that's frightening, a microphone, a speaker, SD card. And I don't see what that one is. The screen won't blow up that much, but I assume that there's a microcontroller in there.

Am I going to write code on the microcontroller? And do you already have drivers for everything else? Yeah, so basically you're going to use the same tool chain that whatever the microcontroller is going to be on that system. So we have from 8-bits, which is going to be an Atmel tour compatible with the Arduino's project.

So even if you have an Arduino project, it's going to work on that. 32-bits, we have a microchip PIC32 MX and STM32. And the most expensive, that's where

that's a different story, is going to be compatible with Raspberry Pi, basically. Oh, okay. So you're going to use the tool chain that you usually use. Okay. And then I can see you have flex circuit going from board to board, and that must be your 45-pin bus. Yes.

Okay. But I don't see where I can just buy modules. Can I only get kits? Are you actually shipping kits? I know this is kind of new. Okay. Yeah. At the moment, we're recording a lot of things that are going on in China,

and no, we're not shipping yet. Okay. And given the state of the world, you're not quite sure when you're going to ship? Or how close are you to shipping once things become unstuck? Yeah. Yeah, I wish, I didn't wish that the problem was that, but the problem started way before, like a year ago with the trade war. And we started actually changing manufacturing basically a year ago.

So, yeah. You're not alone. Yeah. I mean, everything is on the same boat. Everybody knows what's going on, but it's not easy right now to launch a product in the middle of this

situation, but what is going to happen? On your website, it says, and on the boxes, it mentions iOS and Android. And you have BLE modules. Do you have programs? Yeah, that's one of my biggest, let's say, frustrations. It's, like, I worked with embedded products

for a long time and it's not easy to actually work with iOS specifically. We understand. I mean, Android, it's easy, but then you have

security problems, but iOS you have security problems. But iOS is tricky. And so basically we hire a programmer specifically to... So the whole code is going to be open source. And so you have the whole project open, so you can actually modify your iOS product and app and do whatever you want, basically. So it's probably going to be a good product to actually learn iOS and embed it.

And it says AWS partner on here as well. Are you helping people understand how to send data to the cloud? Yeah. So one of the products specifically the Teddy Bear, there's a conversational chatbot and it's very fun. It's like Alexa,

but it's one of the plugins called Lex from AWS. And so you're going to be able to actually create a chatbot. And it's just like in the movie

AI, I guess. That's going to be fun. Okay, so you're going to going to be fun. Okay. So you're going to have 100% open source code. And you've said that these boards, you're helping people go from many small boards to one single board. How are you going to make money?

Well, you know, ideally, I don't want to make like a single board. Ideally, the idea is to make an open source code so you can actually do your own project on this by the way some of the chips it's not that we want to close it but some of the chips there's not going to be a way to maybe buy it

outside our system and that's because some of the chipless that we're going to use or the combination is not going to be available. So it's not that I want to close the system. It's just like we're doing custom chips too. Okay. Sounds like you're selling hardware.

Yeah. Yeah. No, but, you know, another frustration that I have when Raspberry Pi came out is that, okay, I want to buy the chip. Like, I want to put it on my own system. And they never gave us the chance to actually get the chip and make your own whatever right um so that's the beginning of why we start this company actually i think part of it too is i mean was it a realization that there's only so many we we joke often on the show we don't want to work on another bluetooth connected whatever but but there's only so many combinations of things and people We joke often on the show, we don't want to work on another Bluetooth-connected whatever.

But there's only so many combinations of things. And people make a lot of custom products with custom PCBs, but they're all kind of the same, right? There's a microcontroller, and there's a this and a that and these sensors. And it's a grab bag of maybe six common things. So to get from zero to having something that looks almost like a product, it seems like you either go the route of Arduino, where you've got lots of giant breakout boards or, you know, Spark front boards and things, which are fine,

but you're not going to ship that. This seems like bridging a little bit the gap. Is that kind of what you're trying to do? That's pretty much it. And keep in mind that these chips are going to get smaller and smaller. Right. So it's going to happen. I mean, and it's better to start at least building libraries right now. So when it actually gets small enough to just, you know, don't even think about it.

You just make it on something like this. And by the way, we're not the only one doing this. So, yeah, this is going to be more common than you think. What's already more common than I would have imagined? People are shipping dev boards. Yes, yeah, yeah. Yeah, like a Raspberry Pi, Pi Zero, yeah. Yeah, like a Raspberry Pi Zero, yeah. Well, there's the Pi Zeros, but there's also the sensor boards.

Sometimes people do buy them in huge quantities. Well, on the Wi-Fi modules, I mean, there's plenty of products that are ESP8266s. Oh, no, no, don't talk about that crap. Sorry for saying. They're the Regato boards. Those are pre-certified BLE modules. Right. There are a bunch of those.

Oh, you mentioned pre-certified, and I wanted to go back to the certification because I don't have a lot of experience with electronic certification, but if you certify each of the modules, do you have to certify the combination, or does the combination inherit that goodness? You can use it the

way it's shipped. If you buy a pre-certified microchip Bluetooth module, you just solder it onto your system. The only difference is that our module you're not going to solder it by hand.

There's no difference. The only thing that that our modules, you're not going to solder this by hand. Right. But there's no difference. Okay. The only thing that you can do is pre-program the chip. I see. Which a lot of times you can't do anyway. Right. I mean, some people will do it for you. I know, actually, Microchip did it a lot.

I remember Microchip did it for you. Yeah, in Atmel. Yeah, Yeah. On your website, you also mentioned Kickstarter and Indiegogo support. What do you have in mind for that? Whatever good project we think is going to be enough for us to get involved,

we're going to get involved. So if you bring us a project that makes sense and we have to make a new module for that, we might support you on the campaign. Okay. But it's only for new modules, basically. Ah, okay, new modules.

Okay, so you've mentioned chiplets a couple of times. Can you explain what these are? Yeah, so chiplets are a solution for big chip manufacturers right now where the cost of making a processor, it's doubling the price, and it just makes no sense for them. So basically, Moore's law says that every two years, capacity doubles. There's also a second law that says that the cost every four years doubles. And that's what's happening right now. So it takes a lot of money and a lot of R&D to make new chips on smaller packages.

And probably 80% of the chip doesn't make sense to make it so small. So the solution is to separate each block of a chip and make chiplets. Okay. So you have a processor core and you have memory cores on different technology and different packages. I think AMD already released a chiplet-based chip, and IBM and Intel is also working on that. Now, the problem is that each company is working on their own solutions. And the defense industry through DARPA created

a project to actually try to unify the technology so you're going to be able to mix and match different companies on the same chip. Is this putting two dyes together?

Or two dyes in a package? Or two packages on a bigger module with wire links? Yes and yes. Or is this just IP macros? Yeah, it's everything. And at the same time, having the chance to actually separate those chips on different packages. Okay. So, I mean, it's really interesting for us because we're going to be able to actually ourselves build different chips. So the idea between the chiplet is how you communicate on the physical layer. And that's what, you know, a hero without a cape,

which is the Linux Foundation is trying to unify here. So they created the Chips Alliance. I think the last one to join was Intel. And so they're creating a protocol so they can communicate each other. And you can combine it either on a physical level or, so the circumple level or separate. Okay. So that's where the whole chiplet revolution is going to change a lot on manufacturing basically. This doesn't sound new to me. I mean, we've had the ARM Cortex-M3, but with these peripherals only and this memory,

and I want this specific arrangement that may not be available, and then they'll do it using this multi-chip, multi-module method? Yeah, that's always going to happen. That option is always going to be there, right? I mean, silicon level is always there. But for somebody like me who might be making a low-volume product, that's not an option because it would cost millions of dollars. Well, not necessarily.

And that's the point. So when you create the chiplets and you unify the protocol, and that's where the Linux Foundation is going to save us, basically. If that really happens the way it's planned, yes, you're going to be able to actually mix and match and make really cool stuff. But the price has been an issue. I've been in a couple of companies, two that have made their own chips Really cool stuff. But the price has been an issue.

I've been in a couple of companies, two that have made their own chips and two that have realized that's way, way too expensive unless you're shipping in million quantities. Yeah. How are they, I mean, silicon at these small sizes, it's a tough process. How are, how is, okay, let me step back for a second. You're talking about making modular electronics because we have these pretty often blocks, but we have our occasional need for a particular block that we make for our particular toy, a particular product that we make for our particular toy or a particular product. Sorry, I'm looking at teddy bears and talking about products. And then we're talking about custom or semi-custom chips.

And I see how they're both modular, but it seems like the modular electronics is making things cheaper and smaller, and custom chips has always been an area that's really, really expensive. It's one of those deep water, don't do this unless you're really sure areas. How are these joining in your company? So a lot of the technology that is going to start being used on chiplets is the technology that we're using. So chiplets are those small chips that have already been combined with chiplets. It's already been used on custom ICs, right?

Is it that... I think what he's saying is that the chiplet method allows you to combine what normally you would have to do on a die with... It's like another step of modular, right? You order, okay, I want my chip to do this, but it's really a chip with five dies that are interconnected using the standard method that they just pulled off a shelf

and so you're not really ordering custom silicon you're getting a custom processor thingy that's a combination of a bunch of silicon that they already had so it's not that you're saying spin up a whole process to to make me a wafer with all this stuff on it. No, no. So the idea is you're going to be able to buy VCA packages, basically. Right. It's kind of like a multi-chip module. But stacked dies are really tough because of the precision requirements. And MCMs are really tough because of those tiny, tiny wires between them.

Yeah. So at that point, I mean, that integration has to be done for someone else and we're going to do that. Okay. So that's where, you know, motionless electronics are going to make more sense and to actually buy that chip. Okay. This idea of modularity for both chips and boards has been around for a long time. You sent me a video about the military talking about it in the 50s.

What was that? Pallets of vacuum tube that you connect together? Yeah. Yeah. You should put that on the show notes. It's very interesting. Project Tinker Toy. I see precursor.

Okay. 1953 U.S. Navy automated manufacturing of modular electronics. Yeah. It's interesting. It's not new. And that's the secret. I mean, modules of transaction in the defense industry has been used since the 50s and maybe more. It's a different way to think system engineering.

And actually, it's a really good way to keep secrets. Because, you know, at one point, you're going to be working on one part of the project, but you don't know what's going on on the other part. Right? So, you have a lot of teams working on modules, and they don't know exactly what their final product is going to be. So that's probably one of the main reasons, but the other reason is like it's really easy to maintain, to replace parts, basically,

versus one giant PCB board, right? Yeah, because you can replace the broken module instead of replacing the whole thing. As long as the system is designed to not, well, I would call it spaghetti code if it was in code. If it was a nice, truly modular system, you can test and replace modules, but

in the cases where you, like when the processor has to be able to control the power subsystem, that's where, that's an area that breaks down modularity in some of the boards I've worked on is how chips come up and how the power system has to sleep or be able to turn off other components so that power goes low. How do you figure out what should be a system-level concern and what should be purely on module? How do you do API design? Yeah, so basically that will be managed by the main chip. And yeah, that's interesting how to control power management. That's the most important part. I've been working on coin cell operated products.

So I think a lot about power management. And that's one of the things that, you know, when you work on coin cells, you're pretty much looking for every bit configuration, whatever clock is running that you have to run, whatever. Everything. You're going to go look one by one on each chip configuration, basically. So it's not easy. That is true.

No other advice on that one? I mean, I totally agree with you, and I'm not sure what advice I'd give. No, basically, yeah, no. Just work with your oscilloscope power. You know, check the power all the time and run loops

and see what's wasting power, what clock doesn't have to be run, what process it's running unnecessarily. Yeah, it's a lot. That all makes sense when your processor has access to everything around it. But when you have a truly modular system, you start losing the interactions that you would have maybe on a single board.

Yeah, not necessarily, because actually you have more control. Are you talking about backdoor connections between the processor and maybe something that you can't reach in a modular fashion? I'm just thinking all the times I've i've had to ask an electrical engineer can you give me access to the power so i can turn the sd card or the microphone off when i need to sleep so it doesn't continue to take battery and how that's always one of those things that it requires yet another pin and in a big system you end up with a lot of yet another pins. Yeah, so that's basically the solution here. So you have a main controller on each module taking care of that.

So there's a bus controller on each module. Not necessarily a bus controller on each module? Not necessarily a bus controller. You don't have an FPGA, like, maxing everything. No. Sure. You have a chip that is going to tell the local module when to turn on and when to turn off, basically. So it's not just that the module is just the accelerometer with some pins that go to the 45 pin. That goes to another, there's another chip on there that controls that. And you control also through that chip via ISA Square some process that maybe doesn't have to be tuned off, but you can tell the accelerometer, okay, it should go to low mode.

Okay, so that's interesting. So does that mean when I write my code on the main processor, does it look like I'm talking to an ST accelerometer? Or do you provide another API kind of abstraction in the hardware? Yeah. Yeah, no. So you have two APIs, basically. You can pass through a lot of stuff. But there's going to be some specific for the module.

Okay. That's not bad. That's interesting. No, it is interesting. But now you're going back to, okay, that other little thing. That's not bad. It's interesting. No, it is interesting. But now you're going back to, okay, that other little thing now takes power. Well, there's another chip that controls that one. Yeah.

I mean, it's turtles all the way down. What are those little controller chips? Like tiny picks? They're microchips. Yeah. Okay. So they're probably fine and they're and they're yeah and they're very low low power so yeah by itself the chip is is really good on that

i mean there's there's power and cost you do make the penny the penny argument for for size is really a good one um it just, it makes a nice visual. But I'm always like, oh, well, now there's another layer of complexity. But sometimes in order to get to rapid prototypes, you have to add a little more complexity. Yeah. Yeah, it's the beginning at least, right? I mean, you're going from, which one is the smallest form right now? It's, let me think.

It's not Arduino. It's going to be. The trinkets. I mean. Yeah. I'm sure there are smaller ones out there. There was a red board or something that was very small

but when I think of I want something small to put in I think you can get M0 trinkets now so you can get really small ones right but it's just you have the pin headers

and you have to reconnect it gets big at the end so It's just, you know, then you have the big hitters and you have to reconnect. I mean, it gets big at the end. So, I mean, yeah, it's not going to be the best of everything, but it's the beginning at best. difference between the professional rapid prototyping and manufacturing market and the hobby and maker markets are you looking at both or mostly the former i think the difference is the big difference is that this is going to be pre-certified and you're going to have more temperature control and weather vibration,

ESD, EMI shielding. So it's going to be more towards professional products than, yeah, hobby. And you've been working on this for three years? More like four and a half, but back and forth. And yeah, the last, I would say, yeah, three years.

Almost full time, yeah. How are you funded? You're in L.A., right? So you're out of the Silicon Valley funding circle. Yeah. I'm out of the big dinner parties. Yeah. Yeah, I mean, L mean, LA is not bad.

We have a lot of investors here. Right now, I'm part of Accelerator. It's a co-founder's lab. It's okay. It's good. We have a lot of coaching. And, I mean, there's a lot of people with money here, too. The dynamic is different.

Here in LA, we have more like a defense industry. I mean, you have it here. You have it in Silicon Valley, too. But they're different investors, basically. Oh, basically. Okay. Being in an accelerator, given you have so much experience in manufacturing, it must be kind of odd, I think, because many of the people there have ideas, but no manufacturing experience. Yeah, that happens.

And actually, I'm offering my service to all those guys, and we're having long conversations. And it's actually, it's funny because there's a couple of projects that would be perfect for anything. So I try to convince them, but at the same time, yeah, it's amazing. I forgot how difficult it is when you have an idea and you think you know it. There's a lot of things that people don't think about. Yeah, the whole process of actually designing for manufacturing

is one of the biggest ones. And then there are the engineers who are like, well, once I build it and it's awesome, everyone will buy it, completely forgetting about that step of people need to know about it, which is marketing, which sometimes gets a bad name, which sometimes it deserves, but more often we just like to pretend it doesn't. Yeah, it's, no, there's a lot of things that you have to think of. You know, I've been the last year just focusing on the accelerator

and actually the business part, not even electronics. It's just the business part is what is consuming my life right now. But I had some experience. I didn't mention it, but I had a dot-com back in the 2000s. So I know how it works, and I know the challenges on fundraising. So, yeah, there's a lot of things you have to think of on top of your great idea. And it's tough where you are now, which is, it sounds like you're really far along in the design process, but a little stuck on manufacturing. And so you don't have a launch date. You know, sometimes

it's just the universe. Yeah. I need this time there is a lot of things that we have to work on specifically the business part and you know I'm not desperate

with money right now my wife would say otherwise but yeah it's different when it's different when you're not desperate for money and you have investors waiting for people. It's weird, but sometimes you have investors waiting for people that is desperate and we don't have that. One of the things on your website

is about being a green company what does that mean to you yeah it could be like for some people marketing but if you go to China once you get the idea

how green manufacturing is a disaster you get the idea how current manufacturing is a disaster I really think that modules electronic is going to solve a lot of those problems especially reusing modules

so if you have an environment where you have different products and you can recycle those modules, probably we're going to save a lot of I don't know, problems in the earth. There's a lot of thrown away PCBs

in the world. And that's a lot of bad chemicals. It's a lot of bad chemicals. It's a lot. And it's just actually the process of actually getting those raw materials is from the beginning, it starts not good. So one of the other things you mentioned

as we talked over email was that you attended CES as an exhibitor. What was that like? Yeah, that was interesting. So, you know, I've been going to CES for more than 20 years. And, yeah, this is the first time I go as an exhibitor. I actually used to go, like, I ask you if you go to CES. To me, not really.

But it's a really good place to get clients. Actually, when I got a couple of clients over at CS years ago, just talking to them and offering my service, it's the feedback that you have with all the engineers. And basically, I would say 80% of the people who came to my booth and actually a lot of them might be your listeners too

so I want to thank them there was a lot of thinking on the process and probably I also save a lot of money on not doing certain things in the future and actually just small tweaks on the product.

And so it's just that, you know, talking to engineers who, by the way, they love the product. So it's just that, you know, talking to engineers who, by the way, they love the product. So just that, you know, that was priceless. Did you get to walk around at all? No, nothing. Like, nothing. No.

No. This is the first year, and actually, so when you're an exhibitor, you have a dashboard, and it tells you who's booking your place. And I have a lot of, like, head engineers of the biggest companies, And one of them was from Arduino. And that was cool. I met Maximo from Arduino. And the funny part is I made the CEO of Arduino take the picture. I mean, it's more tiresome than actually walking on sea. It's a lot of talking and standing, not even sitting all day. So yeah, it's a lot.

Has it changed much in the 20 years you've gone? I think the Eureka Park, which actually that's where we were, it got better than other years. And there is a new section where I couldn't do it this year, but you have a supply.

So basically all the Chinese companies doing PCV manufacturing and whatnot. Everything, right? That's a nice boot. It's not a boot, it's just a big tent outside CES. One of the things I forgot to ask about with power and cost and size is security. If we get a lot of modules that are the same or similar, don't they become attacks for...

I mean, this is why monoculture is bad, we want lots of different possibilities out there, but a modular electronic system gives us a few common attack points. How are you dealing with that? I mean, yes, it's going to impact security, but I think that there is bad practices and bad hardware and one example of bad hardware is ESP32 where you can grab the

username and password of your network on plain text that's bad hardware and bad practices would be Phil-Hugh doing, leaving open the over-the-air updates. I mean, why? So I think those are the main problems when you design a system.

But I don't know if actually Felix Hugh, people actually hack inside the chip, and it's one of the most popular IoT products that we have in the market. But it's just that stupid. I don't know, they want to eliminate friction, but why not, you know, put a button, you know? And no, they don't want to do that. Okay, well, there you go. Well, I mean, a lot of our human rights are difficult to get to.

Yeah, but when you're talking about security, I don't know. I wouldn't do that. It's always a battle between security and friction. Yes. And probably it's marketing, not engineers. All right. Well, Ramiro, it's been really great to talk to you. Do you have any thoughts you would like to leave us with?

I love the show. Thank you. It's, you know, one of the, there's a lot of insight on what other engineers think about everything. And I just want to thank you. Thank you for the opportunity. That is very kind of you. And I'm so happy you could join us. Our guest has been Ramiro Montes de Oca, CEO and founder of a thing.io. You can find links in the show notes.

Thanks for real. Thank you guys. Thank you also to Christopher for producing and co-hosting. Thank you to Nathan Tuck for making it sound like I knew what shiplets were. And thank you for listening. You can always contact us at show at embedded.fm or hit the contact link on embedded.fm. Now a quote to leave you with from Patrick Rothfuss, the name of the wind. It's like everyone tells a story about themselves inside their own head. Always. All the time. That story makes you what you are. We build ourselves out of that story.

Embedded is an independently produced radio show that focuses on the many aspects of engineering. It is a production of Logical Elegance, an embedded software consulting company in California. If there are advertisements in the show, we did not put them there and do not receive money from them. At this time, our sponsors are Logical Elegance and listeners like you.