Robotics and 3D Printing with Voodoo Manufacturing

by Y Combinator10/26/2017

Oliver Ortlieb and Max Friefeld are two of Voodoo Manufacturing’s four cofounders.

Voodoo went through YC in the W17 batch. They bridge the gap between prototype and mass production with 3D printing.

We’re also joined by Daniel Gross, a Partner at YC.



Subscribe

iTunes
Google Play
Stitcher
SoundCloud
RSS


Transcript

Craig Cannon [00:00:00] – Hey this is Craig Cannon, and you’re listening to Y Combinator’s podcast. Today’s guests are Daniel Gross, a partner of YC, and two co-founders of Voodoo Manufacturing, Oliver Ortlieb and Max Friefeld. Voodoo went through YC in the winter ’17 batch, and they do 3D printing. They’re basically trying to bridge the gap between prototypes and mass production. This episode has two parts. The first is about 3D printing and robotics and the second part is advice from Oliver and Max after having gone through YC, and that part starts about 35 minutes in. And a quick reminder before we get started, if you haven’t subscribed or rated the podcast yet, it would be awesome if you did. Alright, here we go. The easiest way to put it is manufacturing using 3D printers would create a net positive of jobs, and will it be moved back to the US?

Max Friefeld [00:00:49] – Great question. It will in the US. It’s a net positive for places that have seen manufacturing jobs leave their country. I would say China is probably going to be in trouble. I read a blog post recently about how Foxconn is looking to automate 30% of their workforce in the next three years. They have a million employees, so that’s 300,000 people that might lose their jobs due to automation. The hope, obviously, is that things are growing, right, it’s not a zero sum game.

Craig Cannon [00:01:24] – And so why are they losing jobs?

Max Friefeld [00:01:25] – They’re going to be replaced by robotics.

Oliver Ortlieb [00:01:28] – They extend the low cost option, essentially, and now robots are the low cost option. You can deploy them anywhere, so they’re getting beaten at their own game, essentially.

Daniel Gross [00:01:39] – And help us understand why, in particular, this is happening today. How are the robots of today any better in terms of their technology or capabilities than the robots we’ve seen for a long time?

Max Friefeld [00:01:50] – Yeah, I think there’s a few trends that are converging. Robots are becoming a lot cheaper. Just focusing on the hardware side first, we actually use a collaborative robot, UR10 it’s called, and this is a $35,000 robot. Depending on your perspective, that’s either really expensive or really cheap. It’s cheap if you’re used to spending money on $150,000 robots. The other advantage of the collaborative bots is that you can just plop them into a work environment, and you don’t have to spend a ton of money on safety equipment. Robots are really dangerous, because if you’re working around them, it’ll just kill you, honestly, if you’re in the wrong place at the wrong time. These are designed to accommodate that. They have fore-sensing and stuff like that so they don’t hurt people. This started as a research trend in Boston, with this company called Rethink Robotics, and they released their first arm in like 2011, so, not only did the price come down, but they also kind of got easier to integrate into the workplace. You can actually train them now, just by taking an arm and dragging it to a place, and saying like, move here, and then like, move here.

Craig Cannon [00:03:04] – Well, go into that a little bit deeper, because we talked about it when you were giving me the tour. I’m very curious about how you actually program an arm.

Max Friefeld [00:03:10] – There are so many different ways. The software side of it, I think, is the other interesting part of the industry, which is pretty young. There’s, Ollie can probably talk to this even more, but there’s kind of like software of old with robotics, which is very rigid and structured and focused around safety, and then there’s software of new, which is all five years old or less, and most of that has been developed around this one system called ROS, Robot Operating System, and ROS is really exciting for the industry because it’s really accessible, you can program it, it’s open source. But, it obviously has a lot of things to be improved before it’s quite ready for highly structured environments where reliability is key. Why don’t, you can probably talk a little bit more about what we’ve been planning on doing here.

Oliver Ortlieb [00:04:03] – The cooperative arms are really nice, because even before you get to the the level of we need to integrate ROS to add vision or something else like that, the UR10 is actually at the point where you just pop open a socket and send it some ASCII commands and you have it moving around immediately. The interface there is really easy, and actually we’ve been able to treat most of our robotics projects to this point just as software projects. We’re definitely interesting in bringing people in with robotics automation experience who can sort of handle or think about the corner cases more, and things like recovering from failure, but as far as getting a prototype together, it’s basically just a software project.

Craig Cannon [00:04:45] – Okay. And so to make this more clear, how are you guys actually integrating robots? Because on the surface, you’re doing 3D printing, and those are confined boxes, why do you need arms?

Max Friefeld [00:04:55] – 3D printers are robots, I’m just going to say, by definition.

Craig Cannon [00:04:58] – Okay.

Max Friefeld [00:04:59] – But we’re definitely adding more to the mix.

Oliver Ortlieb [00:05:02] – The cool thing, for us, it’s actually a pretty straightforward application of the arm, and it’s something that people have been using robotics for for awhile, which is just your classic machine tending. Basically you have some high-value piece of equipment that’s churning out these jobs for you, you want to be able to run it 24/7, so you just have an arm sit there, and your machine is so high-value that every second you can keep it producing is worth money to your company. We actually have a slightly more exaggerated case of that, where we have, 160 3D printers, so it’s not one arm tending one machine, like I think you tend to see in the traditional deployment of this technology, we have one arm tending 50 or 100 machines, so it’s very valuable to us, essentially, to be able to do something like this with the arm.

Craig Cannon [00:05:59] – Okay, so why have you guys opted for the machines, the 3D printers that you’ve gone with?

Max Friefeld [00:06:04] – We use MakerBot Replicator 2s, we use them because we know they work. And we actually used to work at MakerBot.

Craig Cannon [00:06:16] – Oh, okay.

Max Friefeld [00:06:17] – The manufacturer of the printer, and so when we were there, we obviously got a lot of experience with these. They’re kind of older technology at this point, so we’re definitely looking to integrate the next thing pretty soon.

Craig Cannon [00:06:29] – And, so people understand the landscape, there’s a whole range of 3D printers. In the past, when did it start, like 20 years ago? People started doing 3D printing?

Max Friefeld [00:06:40] – ’83, I think, was the first patent around 3D printing? In the past 30 years, expensive printers, $100,000 printers, the cheapest printer was maybe like $20,000 printers, have been the name of the game. And these are all used for prototypes. Starting at around 2007, those original patents expired, and people started developing desktop printers, which is what really spurred the 3D printing explosion of like 2012, and fall from maybe, it was kind of this curve of like, oh my God, we’re really excited about 3D printing, everything is going to be 3D printed! And then the reality set in, there’s a lot of work that needs to be done before we’re there. Cheap printers, which is the desktop printers that we use, are just low-end commoditized hardware. The technology was invented, you know, a few decades ago, and people just figured out how to make them really cheaply and scale, and that’s what we’re built on.

Craig Cannon [00:07:36] – And you guys opted into those cheaper printers, just so you could have more jobs running simultaneously?

Oliver Ortlieb [00:07:41] – It’s the Google Amazon method to reliability, essentially, and it sort of flips the current or old style 3D service bureau on its head. Very valuable machines creating very valuable parts in small quantities. For us, it’s all about scale and cheap parts.

Daniel Gross [00:08:03] – Going back to robotics for a second, if I’m listening to this podcast today, and I’m kind of interested in getting into robotics, but I don’t really know where to start, what should I do?

Max Friefeld [00:08:16] – That’s a great question. There are really cheap robotic arms out there that you would start with. Craig and I were actually talking about this earlier, you can buy, like a 5-axis arm. It’s called a Servos. There’s a few different ones, I actually haven’t heard of that one. They use servos, they can lift a few hundred grams, and you can just like put on your desk, and have it move your pen from one point to the other. And you can actually, obviously, add more complex tasks on top of that. Pick up one of those, and then start using one of the open source libraries out there. Programming a robotic arm is a lot easier than maybe it was a long time ago, because they handle all the complicated stuff, like how do you get

Max Friefeld [00:09:07] – from point A to point B, you can kind of just tell it, move to these coordinates in xyz, and then something you tweak it a little bit for how it moves there, but the details are handled right now by software that’s been developed over the last 10-20 years.

Daniel Gross [00:09:21] – And when you guys are programming your larger, cooler arm, are you doing any of it in simulation? Are there any good software packages to be had there?

Oliver Ortlieb [00:09:31] – We had an intern working on that project over the summer, and I think just at the end, he discovered some simulation tools, and yes, basically, you can get a simulation that shows your environment, shows, actually this is one of the built-in features to ROS, it’s just a module that you can include, and it’ll let you simulate and just sort of show you exactly where the bot is, or if you’re running simulations, where you think it is. So the tooling around this, I think, has been improving, and actually today is in a, you know, it has a lot of things that we could use, but it’s in a pretty good spot today.

Max Friefeld [00:10:13] – That’s also an excellent point to bring up, because that is going to be the future of robotics. Simulating, so that the robot can teach itself how to do things, and basically, humans can train robots to do more objective-oriented tasks, and then it can figure out the details for you. Our goal is to eventually build a system for, let’s say assembling a product that we’re making on an assembly line. We would love an arm that is dextrous enough to replicate actions that human hands can do, and maybe just have a video camera, and have a human assemble one, and then the arm will teach itself how to do that from a video. That would be amazing, that is what we want to build, I think that requires simulation, is what we’re working towards.

Daniel Gross [00:11:02] – Interesting. And does anyone manufacture today anything that resembles human dexterity, in terms of fingers?

Max Friefeld [00:11:09] – There’s definitely some robot arms that are meant to grab objects, and they have fingers. They’re slow, they’re not the best. There’s probably two movements when it comes to, like, grippers in the robotics world. There’s the hard, rigid, one or two axis actuators, and then there’s the soft grippers. The hard actuators, you design to do one task, and like, you optimize it for that, and it can do that over and over again, and it’s cheap and simple. The soft actuators, which are not as mature, are kind of designed to pick up objects that are weird like this without breaking it. A lot of them even work with, like, air sacs that just kind of like close around things. And those really aren’t dextrous enough, or precise enough. I haven’t seen anything yet that is as good as a human hand. Which is kind of weird.

Craig Cannon [00:12:07] – Following the line of training yourself to learn robotics, where are the areas that you see developing right now that are super interesting and high-value to companies, where people can start contributing to open-source projects?

Max Friefeld [00:12:19] – If I were in undergrad or getting my Master’s right now, I would be learning ROS, and I would be focusing on research that was towards that training problem. This is also a problem that OpenAI is working on, and has published some cool videos and papers about. If you’re really just looking to get a job, there’s kind of two flavors of robotics right now, there’s the research side, which is what we’re talking about, and then there’s the industrial automation side. Industrial automation is very different. Industrial automation is about performing one task over and over again, really efficiently and cost-effectively. Companies, most manufacturing companies will invest hundreds of thousands of dollars building a machine that just does one thing really efficiently. It’s not flexible, is the main difference. To do that, you need to be a mechanical engineer, an electrical engineer, and have a little bit of software knowledge, and then you can become an industrial automation expert, and go work for Toyota, automating their next assembly line, or Tesla, maybe a little bit cooler. There are two paths. There’s the research path, which is probably more interesting for the intellectual folks who like to think about problems that haven’t been solved yet, and then there’s the industrial automation path, which is a little bit more practical, a little bit more, like, classical engineer, I’m just going to build stuff.

Craig Cannon [00:13:47] – And so Daniel, what’s come across your desk with the AI Grants related to robotics?

Daniel Gross [00:13:53] – Just some back color, AI Grant is a kind of decentralized research lab, where people can fill out an application in about five minutes, and within a couple of weeks, we let them know if we’re giving them a grant that has both cash and $20,000 in GPU training credits. And we’ve gotten almost a thousand applications for the second round of AI Grant. Many of them focus on, of the ones involved in robotics, they’re almost all entirely in simulation. The advantage to simulation is you could do this from, you know, your pajamas at home. You don’t have to buy anything. And they all, they’re all very similar to the work OpenAI is trying to do, where you want to give an agent a video feed, like a human would get, and try to get it to learn to do something. All of the approaches are, as far as I could tell, somewhat similar to what general research is doing, so have yet to come across something incredibly different and weird, but I’m eager to find it and help that person get their paper published, if they’re listening to this podcast now, and have a really weird idea.

Craig Cannon [00:15:11] – What are the approaches that people are using? AI is obviously of interest to many people, but robotics specifically within AI… How training within a simulation even works, could you just walk through that for people who want to understand?

Daniel Gross [00:15:27] – Yeah, and I’m sure there are folks that are significantly more expert in this than me, but the rough idea is, even beyond a simulation in reality, what you’re doing is, you set up a camera, and you use, you know, about the 2014, ’15 technology that we’ve uncovered as a species, which is we’ve managed to teach computers how to look inside images, to get a general understanding of what the agent is looking at. And then traditionally you’ll have a human perform a task, so you know, you’re moving a ball from one hand to another, or you’re moving a ball from one point to another, and then, there’s kind of two scenarios here, there’s one where totally on a freeform basis, the robot arm will try to do the same exact thing, and it’s just trying to emulate what the human hand saw. There’s another approach, which is that the robot arm has some type of goal function that it’s looking to maximize, so it kind of flails around, and it realizes, okay, randomly moving around in space, bad idea. Because that’s not increasing my goal function. You know, touching this ball, seemingly slightly better, moving this ball across the coordinates to the right place, even better. In particular, this is done with things where you could measure very clearly the objective function, like unscrewing a cap from a bottle. And then you basically try to mash those two things together into an end-to-end system. As far as I can tell, a truly end-to-end system that works in the physical world, I don’t think has been demonstrated by anyone yet, but I may be wrong.

Daniel Gross [00:17:13] – There was one gnarly unnamed research lab that had just a goal-maximized robot arm, I think with a bottle cap. And this thing wasn’t even given any image priors. It was just told, try to unscrew the cap. And so the arm, which was quite strong, is flailing around, trying to unscrew the cap, and it’s not smart, it’s just trying to unscrew the cap, you know, don’t blame it. And then apparently, so the story is told, it grabs onto a researcher’s arm, and it’s like, oh let me try squeezing!

Craig Cannon [00:17:48] – And unscrews.

Daniel Gross [00:17:49] – Okay, that’s not working! And so no harm was done, but I think it’s an interesting example to how AI, true AI, could actually be, not evil, pretty harmless, it’s just doing this other thing, where it just wants to get more paperclips. And then by accident destroys us all.

Max Friefeld [00:18:05] – Yeah, hopefully the goal function isn’t take over the world, launch nuclear missiles.

Daniel Gross [00:18:11] – My point is, even with a simple goal function, like get more energy to run faster, you could accidentally cause a lot of damage.

Max Friefeld [00:18:19] – Protect humans from themselves. Kill them all.

Craig Cannon [00:18:23] – Who’s putting in the bounds right now, obviously OpenAI is doing a ton of research around this, but when you hit a level of intelligence maybe slightly past unscrewing someone’s arm, what constraints are people trying to create in AI right now?

Daniel Gross [00:18:40] – Boy, it’s an active area of research and tweeting of Elon Musk. So, I think there’s a lot of people working on safe AI. I am personally pretty skeptical on the regulation approach that people keep on proposing, because it’s not clear to me how you could make it global, and I don’t think that, you know, all the smart people are in America, I think they’re spread across the world, so I think regulation would just mean that America’s last to the game. I think there’s folks working on kind of fail safe or dead man switches, something that would stop an algorithm before it starts hurting other humans. There’s people working on algorithms that are cooperative in their style, so OpenAI and DeepMind actually published a paper on this, where the training is done with a human in the loop who is actively making suggestions, and the hypothesis, I believe, if that would always happen, at least the human could direct the AI to be like, that’s bad, let’s stop hurting my arm, type of thing. And I’m sure there’s plenty other approaches that I’m not even aware of. I am personally, I haven’t heard, it’s quite fascinating when there’s an area like this, there’s an open problem, which is the AGI problem, we’re all aware of, we’re all talking about, you know over dinner, oh my gosh, what’s going to happen. And no one’s really thought of a clean answer, and it’s kind of made me wonder if there is one at all. Cause it’s one of those things where I feel like if there was an obvious one, it would have come up by now,

Daniel Gross [00:20:11] – there’s lot of smart people thinking about it. And so I’ve kind of resigned myself to believe that from AGI, you know, killing us, it’s not clear to me that there is a smart answer. Depressingly, if you go online, you read stories about what we do to other species we consider less intelligent, we’re not that moral. There’s a great slash very sad Reddit thread about how people abuse characters from The Sims.

Craig Cannon [00:20:32] – Oh no.

Daniel Gross [00:20:34] – And, you know, we all did this, you know, you build the house, or you put ’em in the pool, and then you take out, yeah, the ladder, and we all thought it was hilarious. But you think about it, and you’re like, that’s kind of the same thing if you superimpose it onto a super smart AI to us. Obviously, we crush ants without giving it another thought. I’m skeptical that we will find a way to fix that. I am optimistic that that’s not going to be what kills us as a species, though.

Craig Cannon [00:20:58] – Okay.

Daniel Gross [00:20:59] – Because I think there are far more threatening things that are closer. For example, the use of AI by a human to create, some type of nefarious damage. What’s happening is, there’s this AI hype right now, and so more and more industries are moving their infrastructure to kind of non-deterministic machine learning-based stuff. Cars are going to be using machine learning to drive, the electricity grid will manage itself using machine learning. As a result it’ll be more susceptible to attacks from that same thing, as opposed to being manned by a human. So if the power grid is now constantly being regulated by an algorithm instead of a human, could that algorithm be interrupted by another, smarter algorithm in a way that would’ve been easier than to interrupt the human? I think so, so I suspect war will actually not be from AGI, war and damage and carnage will be from weaponized, dumb AI.

Craig Cannon [00:21:53] – To concretely connect the dots, your robots currently right now are fixed to the ground, but they are going to start moving at some point, right?

Max Friefeld [00:22:04] – Yeah, they’re going to be put on mobile bases, not dissimilar from an Amazon warehouse.

Craig Cannon [00:22:09] – Right, have you begun that training process?

Max Friefeld [00:22:11] – No, we’ve gotten a few demos, but the hardware’s actually really early. When Amazon acquired Kiva, they just took that out of the market, and so right around now is when people are starting to release mobile bases again, but they’re all in beta. I think there are a few that you can actually buy. So no, we haven’t gotten started on the moving problem.

Daniel Gross [00:22:33] – Wow, so Amazon’s acquisition of Kiva slowed down innovation?

Craig Cannon [00:22:37] – For the rest of the world.

Max Friefeld [00:22:37] – By five years.

Daniel Gross [00:22:38] – Wow.

Max Friefeld [00:22:39] – Something like that, it took a long time for everyone else to catch up.

Craig Cannon [00:22:42] – What about the automated… Have you been to a shipping container port ever?

Max Friefeld [00:22:47] – I grew up in Long Beach, California, there’s a massive one there, but I haven’t like, gotten a tour.

Craig Cannon [00:22:53] – They’re basically trailers that are autonomous. I wonder, I don’t know who is producing those, but I don’t think it was Kiva.

Max Friefeld [00:23:00] – No, Kiva just does warehouse robotics.

Craig Cannon [00:23:03] – Yeah, that’s what I thought.

Max Friefeld [00:23:04] – Which is all, pick and pack, and stuff like that.

Craig Cannon [00:23:06] – Okay.

Max Friefeld [00:23:07] – I think that’s also, the shipyard example is an interesting, it’s an interesting example that shows how robotics scale in a way that humans don’t. You can build a little mini shipping yard and train it on that, or do it digitally, right, do it totally as a simulation, and then roll it out, and start moving around, like multi-ton objects very quickly.

Craig Cannon [00:23:28] – Yeah.

Max Friefeld [00:23:29] – That’s just a process that doesn’t scale. It scales way more efficiently than a human process does, which is a little bit scary.

Craig Cannon [00:23:37] – What challenges do you guys see, if you wanted to scale to a 100,000 square foot warehouse?

Max Friefeld [00:23:43] – For us, right now, what we’re building is this footprint unit factory, which has 160 printers, and we’ll have like two arms moving around, doing machine tending. And then as parts come off of the printers, they have to be post-processed, QC’d, packed, and shipped out. I think when we automate each of those steps, they each scale very easily. Most of the challenge is getting the first part of it automated. Everything beyond that is can you buy the hardware quickly enough? And traditional software scaling issues, many of which are solved. So going from a 1,800 square foot factory to a 100,000 square foot factory is, kind of like, how many more arms do we need, and how many AWS servers do we need to send up? It’s an engineering problem that actually doesn’t seem that hard.

Craig Cannon [00:24:38] – Do you agree, as CTO?

Oliver Ortlieb [00:24:40] – I mean…

Max Friefeld [00:24:42] – Optimism, yeah.

Oliver Ortlieb [00:24:43] – There are probably economies of scale that we could realize in a 100,000 square foot space that we obviously can’t get here, but yeah, I mean, the easy answer to scaling would just be pound this footprint out as many times as we can into that 100,000 square foot space. And we know we will keep costs basically where they are today. But, then we can slowly merge cells, reorganize as we need to. So this is sort of our worst case scenario right now, but we could in theory pop these open around the country if we needed to.

Max Friefeld [00:25:18] – It’s very different from scaling a factory traditionally, and this is a contentious topic, but scaling a factory of robots is so much easier than scaling a factory of people. Hiring people is difficult, as everyone knows. Hiring people who are good, hiring people who show up to work on time, et cetera, hiring people in a different country if your company doesn’t speak that language is difficult, it requires an on the ground expert. In the theoretical fully-automated factory case, none of those problems are real. If you can run a 100,000 square foot factory, or a 100,000 machine factory with 200 people, it’s a lot easier than 1,000 people running a factory. That’s kind of the idea that Voodoo’s built on, is that it’ll be a lot easier to scale as we build out this footprint.

Craig Cannon [00:26:13] – Right, but you’re also factoring in being really close to many people who are interested in manufacturing, right? I heard on one podcast that you offer pickup for certain folks?

Max Friefeld [00:26:23] – Yeah.

Craig Cannon [00:26:24] – Is that still the case?

Max Friefeld [00:26:25] – Yeah, we’re in Brooklyn, we’re right outside of Manhattan.

Craig Cannon [00:26:28] – So do you have a vision of just many factories? Or would you rather one giant one?

Max Friefeld [00:26:34] – We think of it kind of like server farms. It’s semi-local. So, maybe there’s not a server farm right down the street from everyone in the world, but there might be one within one-day delivery distance, or potentially in pickup distance, for larger customers. It’s actually really common in manufacturing for a company like Toyota to contract with somebody who’s going to make a sub component of one of their cars, and then require them to set up a factory right next to the Toyota factory, with a conveyor belt literally moving the parts from one to the other.

Craig Cannon [00:27:10] – Oh my gosh.

Max Friefeld [00:27:11] – It’s kind of a crazy thing that they can just say, we’re going to work together, but you’re going to build this factory right next to us, we’re not going to build it, you are. That’s also a lot easier with a robotic factory. And you can customize it, so it’s just the right size.

Daniel Gross [00:27:25] – On the topic of Toyota, how do you see the future of the materials that you can use in 3D printing evolving? Today, I think when people think of 3D printing, they think of, like, basically little plastic parts? Are you always going to be stuck with that, or is there a strategy where that kind of changes and expands?

Oliver Ortlieb [00:27:46] – Our vision for Voodoo is definitely to move past just 3D printing. But even within 3D printing, you know, it’s been around since the 80s, I think Boeing is 3D printing titanium parts for all of their planes now. 3D printing as a technology, if you consider the entire range, is a very capable technology, and you know, generally technology trickles down to the low end, that’s what we’ll see here. Voodoo is building an infrastructure for onboarding digital manufacturing technologies, so it doesn’t matter if it’s 3D printing, if it’s capable of doing plastic, metal, whatever, you know, we’re interested in anything that takes a digital file as input, and outputs a physical object at the end. That’s how we’re going to handle the quality and material questions that come up.

Max Friefeld [00:28:36] – We got a question from Twitter about the future of 3D printing, and this point is probably the biggest barrier to a 3D printer being in everyone’s home. 3D printers just right now, don’t have the material capabilities or the multi-material capabilities to produce most of the objects in your house, right? Most things are a combination of plastic components, multi-step plastic components, electrical, you know PCB’s, wiring, soldering, batteries, all of this stuff which is currently assembled. It seems like people don’t think about when they say the future’s going to be a 3D printer in my house making my stuff. Maybe like one percent of the objects you own could be made by a 3D printer right now because they’re just a single material. Eventually we’re going to hit that future where it’s like the Star Trek Replicator. It’s not going to happen with any of the current technologies out there. Until then, digital manufacturing is still a thing. Digital file of physical product, but it’s going to be a multi-step process, which I think benefits from economies of scale from a centralized facility of course.

Craig Cannon [00:29:49] – Yeah, what about even just as literal objects go? Take just the plastic on this chair.

Max Friefeld [00:29:54] – Yeah.

Craig Cannon [00:29:55] – What does the time frame look like to where we can just have 3D printers making at least this part for us?

Max Friefeld [00:30:01] – Yeah. 3D printing right now is not super cost effective for big stuff.

Craig Cannon [00:30:06] – Big meaning?

Max Friefeld [00:30:07] – Big meaning anything larger than like a loaf of bread.

Craig Cannon [00:30:10] – Okay.

Max Friefeld [00:30:11] – When you get bigger than that, it just gets prohibitively expensive, and so printing that chair might cost a few hundred dollars. There are big printers out there, so people are definitely working on that problem. But you kind of have to balance quality with speed, and so I think the next wave of 3D printer parts in the world and digital manufactured parts are going to be consumer packaged goods, like smaller things that fit. Kind of again in this like loaf of bread sized shape. Yeah. You’re not going to just start printing tables and houses and stuff without significant improvements in the technology.

Craig Cannon [00:30:52] – Okay, and it’s just expensive because it’s a large amount of plastic?

Max Friefeld [00:30:57] – It really comes down to the extrusion rate of the machine.

Craig Cannon [00:30:59] – Yeah.

Max Friefeld [00:31:00] – Let’s say that that chair is either injection-molded or roto-molded, different molding technologies. You have like a physical mold, which costs probably tens of thousands of dollars to make, and you have a machine that very rapidly injects hot plastic into it. The rate of injection of those machines is probably like a thousand times faster than your average 3D printer. Based on that fact alone, you know, you can just take the cost of the machine time and the injection rate and approximate how much it costs to build something, and large objects just are too expensive. The ratios don’t work out yet.

Craig Cannon [00:31:41] – So there’s one handful of questions from Twitter. One is about IP in China. Do you remember this one?

Max Friefeld [00:31:47] – Yeah.

Craig Cannon [00:31:48] – Wyatt Sanders asks, how do you deal with intellectual property in China? How hard is it finding support for new hardware as opposed to software?

Max Friefeld [00:31:57] – I’m trying to understand the question exactly. I read it and I have a few questions about the question. But I’ll do my best since this isn’t an in-person question. Honestly I think with IP, unless you’re incredibly restrictive about like what gets out, basically anything that isn’t a well-guarded trade secret, somebody else will have. The most interesting I guess story I heard is when you’re manufacturing in China, IP is thought about very differently than it is here. I think there’s kind of this belief that like things are just shared, and so that’s why like everybody gets the sneak peek of an iPhone, because in reality,

Craig Cannon [00:32:45] – I was going to ask you about that.

Max Friefeld [00:32:46] – Culturally, it’s not as big of a deal to just kind of like share stuff I feel like. It’s very different in China than it is in the US. If you’re getting something made in China, I wouldn’t really count on IP being a big factor, and even if you can keep something totally under wraps, like let’s say you’re building a top-of-the-line drone or something, once it’s out in the market, somebody in China will buy one and then in like two or three months just replicate it. It’s not very difficult for them to do that with hardware. Hardware is becoming easier and easier to replicate over time and so it’s getting closer to software. Hardware IP will actually work a little bit more like software IP, which is difficult to protect.

Oliver Ortlieb [00:33:38] – Actually an interesting to talk about Voodoo again a little bit and where we fit into this. One of the benefits of digital manufacturing is that you’re not rebuilding your molds all the time when you’re reiterating on your product. Actually, the avenue for manufacturing companies protecting their products could be just iterating more rapidly and using companies like us to really stay ahead of the market.

Craig Cannon [00:34:03] – How did Apple think about it?

Daniel Gross [00:34:06] – You’d have to ask them.

Craig Cannon [00:34:08] – No comment. Alright. Next question. Enrique asks, I don’t know if you would even have an answer for this but, 3D printed solar panels? Have you heard of this?

Max Friefeld [00:34:22] – No, I didn’t even see that question.

Craig Cannon [00:34:24] – It came in on Facebook.

Max Friefeld [00:34:27] – Man, I don’t know anything about 3D solar panels.

Oliver Ortlieb [00:34:29] – Do they exist? Yeah, what? Do we have any context for this question?

Craig Cannon [00:34:32] – The context is, and I’m adding a word here. What has been the progress regarding 3D solar panels, which I assume means 3D printed solar panels?

Max Friefeld [00:34:40] – Got it. My guess is minimal. A lot of like processes that you use in making chips or dealing with silicone wafers have the same names as 3D printing processes, so there’s like stereo lithography is a common way of getting parts 3D printed. Basically, that just means you have like a laser that’s etching some pattern into something. I can imagine a world where you can make solar panels with 3D printers, but I actually don’t know anything about it.

Craig Cannon [00:35:15] – Alright, next question then. Bit Bit (@searchforbetter) asks, what will 3D printing, will it ever be incorporated into our educational system the way other technologies have?

Max Friefeld [00:35:24] – Definitely. It’s a pretty simple answer. Yeah, I mean I actually, the first time I used a 3D printer was in high school.

Craig Cannon [00:35:31] – Oh really?

Max Friefeld [00:35:32] – Yeah.

Craig Cannon [00:35:33] – Where’d you go to high school?

Max Friefeld [00:35:33] – I went to this really nerdy school in California called The California Academy of Math and Science.

Craig Cannon [00:35:39] – Man.

Max Friefeld [00:35:40] – It was a public school, but they got sponsorships from like engineering companies to help us get things like 3D printers, so yeah. I thin it’s definitely moving in that direction, and that, when I was in high school, the $2,000 printer didn’t exist, so it was a big deal to have this like $30,000 machine in the room ya know, that we could use. Machines are cheap enough. They cost the same amount as a computer. It’ll just follow the same trend like computers started ending up in classrooms in the 90’s right?

Craig Cannon [00:36:12] – Yeah.

Max Friefeld [00:36:14] – If today is the 90’s of computers, in 20 years, I think kids will probably be using them to make all sorts of things.

Craig Cannon [00:36:24] – Okay. I did have a couple questions about just your experience in YC. Were you guys in this group? Were you in Daniel’s group?

Max Friefeld [00:36:33] – Yes.

Craig Cannon [00:36:33] – Yeah.

Max Friefeld [00:36:34] – Yes we were.

Craig Cannon [00:36:35] – Now I can connect the dots. We can talk about Daniel off the record, but what was your experience like at YC being a hardware company? Do you find it was any different?

Max Friefeld [00:36:46] – Yes, I mean we’re not quite a hardware company is the other side of it, so we I think before we went into YC, we didn’t quite know how to position ourselves with investors or other people, ’cause we’re not a hardware company and we’re not a software company, exclusively. I think being in YC and talking to Daniel and talking to Sam, basically helped us find ourselves as a robotics company. That’s really what it comes down to, which is the intersection of hardware and software. If we succeed in the future, we will have contributed significant gains to the world of robotics and automation and manufacturing.

Craig Cannon [00:37:31] – Do you remember what they were like in the beginning?

Daniel Gross [00:37:35] – Yeah. I mean pretty formidable. I think they were more formidable by the end though, and I think YC, the hope is that YC gives founders a lot of things, but one that often people discount, because I think it’s a little bit hard to capture in a couple of bullet points you can read on the internet, is it really changes founders mentality and the way they think of just how big their company could be, and I’m not saying that you guys were meek when you came in. You were already great. You had already had a great business, but I’ve found that with the Voodoo founders and a lot of other folks, it hopefully takes it to the next level, in terms of figuring out how to position their company when talking to investors internally and frankly to themselves.

Craig Cannon [00:38:21] – A question we get often times is what is it like being a solo founder in YC? You guys are kind of on the opposite end of the spectrum with four, correct?

Max Friefeld [00:38:29] – Correct.

Craig Cannon [00:38:30] – What’s that experience been like?

Max Friefeld [00:38:33] – We had a, it was actually really helpful for us, because we’re based in New York and YC’s obviously in California. We kind of had this dilemma because we can’t just move the whole company there when we get into YC. We have a factory. We’ve got people, so it was really helpful to kind of break up going to YC on different weeks so that we could focus on building things here, but also make sure that somebody was there talking to partners, getting advice, building a company. Otherwise, you know, I think our founding story kind of goes back to the fact that we’ve known each other for a long time. We founded a previous company together, sold that company, and the typical pitfalls that you would find with a four-person team, we’ve been lucky to avoid. Usually with four people, I’d imagine you’d run into like serious issues of like, who owns what, or two people might not get along, or there could be so many different problems because your graph gets bigger. We are used to working together, so luckily, we haven’t had any issues.

Craig Cannon [00:39:42] – Interesting.

Max Friefeld [00:39:43] – Oliver, maybe you can confirm?

Daniel Gross [00:39:45] – I would just agree with what Max is saying. I think it’s important to sort of carve out roles for everybody in making sure that people are getting input to the things that they’re really bringing value to. But at a certain point, you know, not everybody is involved with every decision day-to-day.

Craig Cannon [00:40:03] – Right.

Max Friefeld [00:40:04] – On a large founding team, I think that sometimes hurts feelings or rubs people the wrong way, but for us, I think we’ve done a pretty good job of defining what everyone’s role is on the team and making the hard decisions together, but also making sure that we’re not taking things personally when hey, somebody had a meeting, a decision was made. It’s I think, everybody buys in to the idea that we’re a team and we’re doing this together and everybody’s got a part to play there, even if it’s not always exactly in the direction they want to go.

Daniel Gross [00:40:41] – Just for folks that are listening to this and are maybe single founders wondering how to meet co-founders, how did you guys originally all meet?

Oliver Ortlieb [00:40:49] – We have a very funny story I would say.

Max Friefeld [00:40:51] – Go for it.

Oliver Ortlieb [00:40:52] – Alright no sorry, go on.

Max Friefeld [00:40:55] – Voodoo has four founders. Three of us went to college together. Oliver was actually at the grade above me, and Jon who is one of our other founders. Jon and I actually kind of decided we wanted to start a company while we were juniors. We had internships that summer that we like didn’t, so we dropped them, started a company. Oliver had graduated already and had a job at Teradata, and we convinced him to quit his job and move back to Claremont, which is where we went to school, and live in a house with two people who were still in school.

Oliver Ortlieb [00:41:33] – I was doing training with Teradata. I had been there for six weeks or something like that, and I had to go quit basically during a training program.

Craig Cannon [00:41:41] – Nice.

Oliver Ortlieb [00:41:42] – That was difficult, and then I had to move basically across the street from the college I had just graduated from, so a bit humbling at the time, but clearly sitting here, I think the right decision.

Craig Cannon [00:41:55] – Whoa, what was that pitch like, because that’s another common question. How do I convince someone to join my team?

Max Friefeld [00:42:01] – That’s a really good question. I’m interested in hearing Oliver’s perspective because the other weird thing about our history is, we weren’t like best friends before we started our first company. I kind of knew Jon, and Jon kind of knew Oliver, because we went to a small school and like, I was in a class with Jon.

Craig Cannon [00:42:20] – Did you know Oliver at all?

Max Friefeld [00:42:22] – No.

Oliver Ortlieb [00:42:23] – I don’t think we’d met until you guys were pitching me on joining.

Max Friefeld [00:42:27] – There was a little bit of risk, right, which is why I think we got really lucky.

Craig Cannon [00:42:32] – Yeah.

Max Friefeld [00:42:32] – That our personalities were compatible. We went to a really tech-focused school, so just coming from the same environment was a good starting point, but to get back to your question. The pitch was, do you want to start a company and you know, selling him on the future of 3D printing, which at the time was really easy, ’cause it was everywhere. I remember one night we had to drive down to San Diego and get dinner with Oliver’s parents and convince them that it was also great.

Oliver Ortlieb [00:43:01] – Very old-fashioned, yeah.

Max Friefeld [00:43:03] – Almost like we were dating, right? I don’t know what happened from your side.

Oliver Ortlieb [00:43:09] – Yeah, I mean the pitch was basically, do you want to do fun stuff or not? And I wasn’t feeling super fulfilled in the position I was at. I tend to be a very cautious and a cautious person. I like to think of worst case scenarios and stuff like that. This was a very out-of-character decision for me I think to just sort of bail on a job and jump in, but it worked out. A lot of it comes down to timing, right? It’s just they got me at the right time. I was not their first choice actually, which I think is an important thing to bring up. Just because, for a solo founder out there who’s trying to find somebody to make it work, it doesn’t have to be your first choice. It doesn’t have to be your second choice. There’s somebody out there that you can probably build this company with.

Craig Cannon [00:44:06] – What were the technical capabilities of the other people on the team when they were trying to bring you on?

Oliver Ortlieb [00:44:11] – It was just the two of them. They’re both engineering, or they had their engineering degrees, or were getting them at the time.

Craig Cannon [00:44:19] – Okay.

Oliver Ortlieb [00:44:20] – Really they were looking for a, someone with a software background to join the team and sort of take on that part of the company. So that’s the role I came on for. That’s the role I filled basically ever since, is sort of owning the software side of the product.

Craig Cannon [00:44:40] – Do you have any pro tips Daniel?

Daniel Gross [00:44:43] – On what, on finding co-founders?

Craig Cannon [00:44:45] – On finding and then convincing the co-founder?

Daniel Gross [00:44:47] – I was the single founder at some point during YC, and then I found someone and became a very good friend of mine to be my co-founder, but we’ve… It was a risky decision because I met him after YC. His name is Robby and we spent maybe a total of 48-72 hours together, coding together, talking, before we made the jump. So we kind of really lucked out in that way.

Max Friefeld [00:45:18] – It sounds like a theme.

Daniel Gross [00:45:20] – Yeah, the one nonintuitive thing or semi-nonintuitive thing that I would encourage, I always encourage people to think about is, as it turned out for my interaction with Robby, the fact that we were destined to be really good friends mattered more than the fact that we both worked well together. So what I mean in particular is, as it turns out, we had the same shared taste in music with fairly similar political beliefs. We had fairly similar hobbies. We like the same video games, and that context I’ve found matters quite a bit, because you end up spending a lot of your waking hours with this person, and some of those hours are quite stressful. It’s useful to have this shared substrate of you know oh, you both like @deadmau5 or something to fall back on, versus just shooting for someone who’s technically competent.

Max Friefeld [00:46:09] – I would agree with that point wholeheartedly. Oliver and Jon are two of my best friends and they weren’t before we started our first company five or six years ago.

Craig Cannon [00:46:20] – Yeah.

Max Friefeld [00:46:20] – And a lot of that is, we go to concerts together. We go to sports games.

Oliver Ortlieb [00:46:23] – We were roommates for four or five years.

Max Friefeld [00:46:26] – Yeah, I still live with Jon.

Oliver Ortlieb [00:46:27] – Oliver moved out because he has a girlfriend. So good for him. That, I think that was key to us, because we actually just enjoyed spending all of that time together.

Daniel Gross [00:46:34] – Yep. So I mean, I think one way to encapsulate this, is if you’re looking for a co-founder maybe more, better to start looking, I know this sounds kind of lonely, but start looking for someone who could be a really good friend. And then your second filter is if they’re technical or not, as opposed to the opposite.

Craig Cannon [00:46:50] – Yeah, that’s a great point. In closing, you guys have been out of YC for awhile now.

Max Friefeld [00:46:56] – Six months.

Craig Cannon [00:46:57] – Yeah. What’s your advice for other YC companies. I know the during YC experience is very different than the post YC experience, building your company. What have you learned since YC that you’d like to share?

Max Friefeld [00:47:10] – Everybody has a slightly different experience after YC. For us, obviously closing fundraising was something that took effort from me and Jon. That was two months after YC. We had actually kind of pulled the round together, but doing the documentation and paperwork took a long time, because we priced the round instead of doing SAFEs. So if you could do SAFEs, do SAFEs. That’s advice number one. Beyond that the hardest thing that I think we’ve encountered since YC is keeping up the momentum. YC is just great at giving you like this excuse for everyone to work really hard, like nobody took vacation and all that stuff. There was a little bit of like a post YC hangover for the company, because we’d all pushed ourselves really hard, and then we had to like kick everyone back into gear right after we closed fundraising and just start blazing ahead. If I were to do it again, I would be, I mean it’s easy to say this in retrospect, but I would be thinking about that during YC and right afterwards, like while you as a founder are really distracted with fundraising? Maybe the one thing you should focus on is keeping everyone else motivated and moving, even if you’re going to be kind of busy and out of the loop. That’s my experience.

Oliver Ortlieb [00:48:30] – I guess we’re also a little bit larger than probably the average YC company is…

Max Friefeld [00:48:34] – That’s true. We were 16 people in YC.

Daniel Gross [00:48:37] – Yeah, I mean there definitely are some even larger companies, but you guys were on the larger end. One idea on that, that I found really useful for me when running my company is, the obvious forcing function you can give yourself is just PR and media. You just tell the media you’re going to loan something in two months, and then you can use that as a way to rally the team. That works, but you can’t do that all the time. So the other thing that we used to do that was quite useful is we used to set up much like YC did these demo days, where we’d invite everyone’s friends, family, and most importantly significant others to demo what they’ve built. People would sign up ahead of time and say, oh you know, by next month’s demo day, I’m going to have this, and it turns out when you like tell your significant other that you’re going to have built a thing by a date, you really care, ’cause you don’t want to look like a fool. And so that provides kind of a synthetic internal forcing function, and you can make it like a fun event. Invite people over to the office, kind of humanize the company to all the folks that don’t work there. I’ve found quite a bit of that instilled pressure in the team, that we wouldn’t of had otherwise.

Craig Cannon [00:49:48] – Okay, cool. This was great. Thanks guys.

Max Friefeld [00:49:51] – Thanks.

Daniel Gross [00:49:52] – Thank you so much! Please change the future of manufacturing in America for the better.

Max Friefeld [00:49:56] – We’re workin’ on it.

Craig Cannon [00:49:57] – Alright, thanks for listening. If you have some time, please leave us a rating and review, and if you want to watch the video or read the transcript, those are both at blog.ycombinator.com. See you next time.

Author

  • Y Combinator

    Y Combinator created a new model for funding early stage startups. Twice a year we invest a small amount of money ($150k) in a large number of startups (recently 200). The startups move to Silicon