The Crystal Ball on AI with Juan Aparicio Ojea (Linkedin Live)

Starting the Linkedin Lives up for the 2024 season, Nikki, Ali G, and guest host Courtney talk with robotics and AI expert, Juan Aparicio Ojea.

Juan is the co-founder of a start up and spent many years working with robotics. They discuss how AI can be brought more into robotics, the lack in current AI technology when it comes to automating certain jobs, and a look into the future of technology.

Transcript

[00:00:00] Hey, welcome to the first linkedin live automation ladies episode of 2024. happy new year, everyone. And, yeah, if you're here, if you showed up, If you were earlier than us, especially, please tell us who you are in the comments, where you're coming from, whether the weather is crazy, where you are, we've had deep freezes all over the place, I just got back from the A3 business forum in Florida, so I managed to escape most of the freeze here in Houston, but was stuck on the runway for de icing for a while, and then I Finally made it back.

 I think i've kind of got my voice back to normal talking Straight for like over 12 I don't know how many hours that is from like 7 a. m to 3 a. m. But yeah, I'm thankfully not sick, but Allie, how are you? We've, we have Allie on today and I'm super stoked. I am surviving, the lack of [00:01:00] sun and also very busy and wrapping up a new project that, involves a lot of people and, yeah.

Going through those like growing pains and started to look at like small group, insurance so I can insure people and stuff and so I'm like got my big kid pants on. That's how I'm doing. Yeah, everybody was asking about you at forum. Oh, is Ali here? Is Ali here? Next year, you'll catch her. It looked really fun.

Yeah. Oh, hey, Ted. Good to see you. Were you, I guess he was at, there was, now that I'm seeing like some of the posts from A3, I'm like, dang it. How was I in the same room with you? And we didn't get a chance to talk. So, hey, Ted, good to see you. Thanks for joining us. We'll have to catch you next time then, if not earlier than Automate in May.

 And I'm sure you're doing some big things at IMTS as well. So, look out for a future [00:02:00] collab with Ted and the, Festo education, Festo didactic arm. Well, I would like to welcome our special guest Juan, long time no see, as in like yesterday. How are you? Yesterday, I am also recovering from three days, even for me, four days, non stop talking.

Yeah. With 750 very alike fellows in the automation world that love automation like me and we can talk forever. So Yesterday I was like, oh, I lost my voice. I have to tell Nikki. Yeah, we are lucky enough. I woke up today and I get like 90 percent of still, which is okay. So hopefully the 10 percent of the voice that I lost, the spine doesn't get in their way.

Okay. And we are hoping for Courtney to join us. You mentioned that she would. She is actually in Florida right now. She texted me yesterday while I was heading to the airport and asked if I was still in Orlando because [00:03:00] she was en route on a flight to Orlando. Hey, welcome. I'm so glad to see you.

Talking about the devil. I was just talking about how you, uh, how we crossed paths like ships in the night yesterday, to and from Orlando. Yep. So what are you up to? I'm here for another day in Orlando with United Robotics Group, just, doing a company get together. Oh, okay. Because I know I missed you.

We'll, we'll get together. Well, say hello to Greg for us, please. Will do. And, uh, I know our friend Sammy Birch is still over there as well. I'm going to try to run into her tonight, if I can. And then I have to give a shout out to our good friend Allie Walker at Grey Solutions. Somehow, we locked eyes across the room like four times over the event, and texted each other, and still did not pick up.

I saw the alligator. That was from last year. Oh! [00:04:00] Unfortunately there were no alligators at the party this year. At least not that I saw. Maybe fortunately for the alligator, right? How does it feel like being hung from one hand to another? That seems to be a common theme for me and Ali to be at the same conference, in the same place, in the same room, and yet be so busy that we don't end up connecting with each other.

So, that's another 1 where we're going to have to really make an F, make sure that we see each other at automate. But, yeah, to get to the topic at hand, so we're going to be talking about robotics a bit today and it looks like. You've got your company logo on there. Automation. Did I get that? Resave automation.

That's correct, right? So, I think the thing that got a lot of people excited was like, what the heck is Juan doing, right? Setting up the expectations. I don't know how much I will say in the next hour, [00:05:00] but what I can tell is that We have been busy building a new company in the robotics and automation space for the last month, and the team is growing, and we're very excited to launch in the next few weeks, months, who knows?

It was the worst kept secret, right? So I was also like working in the conference with this. Okay. Okay. Well, you don't want to say a name, but I can kind of read it and I don't know that company. Uh, so one thing I can tell is that we are called Reshape Automation. And, my co founder is Carlos Vanegas.

He is, we know each other. We come long time, but he's an ex out of this bright and rapid. And we are in this because we really want to help SMBs to automate, and we want to reshape their automation journey, right? The way that they discover automation, that they design automation, and we're going to talk about the problems, I think, and also crowdsource what people think are the main problems of why people [00:06:00] We only deployed 30,000 robots last year.

Wow. It used to be 40,000. And I was making the case, this is too little. And then I saw the numbers yesterday and it's actually last year it was a 30% dip. Ouch. Only 30,000 robots. So I wanna really dig into what are the problems on why we don't deploy faster? Because there's a correlation, right? The more automation we have.

The less we are dependent on labor shortages, the more we can grow as a country and also manufacturing install base. So we are here to really dig deep into this problem and offer a solution, right? And we want to hear from customers as much as possible. Also, the experts like you on why is that the case, right?

What is happening and what is needed in this industry to accelerate in terms of automation. Okay, so a lot to unpack there. I have, but I do want to, I think that we should, I'll try to keep our format a little bit consistent. [00:07:00] Can we back up? I know we know you or you've been around, right? For so long that I kind of feel dumb asking the question, but I think there's probably people in our audience that may not know you.

So I'll start with our standard question at automation. Ladies are only standard question, which is, can you tell us a little bit about the backstory one? How did you get interested in automation and robotics and how did you get to where you're now starting your own company? Um, doing all the things that you just said.

Perfect. Great question to start. And so the accent people can probably tell is from Spain. So yeah, That's where I grew up, then started in Sweden, and I was a telecommunications engineer by training. So I know 5G, now 6G, how do we do an antenna here and there, which was interesting, right? At some point, what brought me to the U.

S. was actually the first, very first, indicators of projects in the autonomous car [00:08:00] industry. It was actually communicating the infrastructure for traffic controllers that Siemens manufactured in the U. S. Should we eat cops right to enhance them to make them better? For example, predicting where an ambulance is coming, right?

Sometimes you don't come and discuss in the middle are horrible. That is what brought me to the U. S. I started working at Siemens, in Princeton, and those were the first projects we're talking about. More than a decade ago. An autonomous cars was not even a hype back then, right? But it was like this was going to happen around the corner.

So, yeah, they are coming next year. Like we're talking with them with the person responsible at Google. They were the ones pioneering that technology. And it looks so near future. And here we are. And still that we are not at the point of wide adoption, but we have come a long way now. As I was growing Siemens, I got exposed to automation.

I was like, Oh, wow, there is a lot of potential here. Like [00:09:00] telecommunications, 5G, 4G, a slight difference. We are optimizing. We are just before it was okay to send text. Now people were excited to send pictures around. So you need more bandwidth. Now is tiktoks in the future is going to be your own avatar in three days, and you just need more bandwidth, but it's an industry that is very mature and really have embraced all the advances in software.

You go to automation, you're still sending, like, exchanging memory blocks using kind of the same way that you were programming 20, 30 years ago, right? There hasn't been really a revolution. In the last decades, like the ICT revolution, this information communication technology, all this industry 4. 0 is still a moving target, but the reality is that, yeah, we do want to program, we'll use one of the CMOS PLCs that Ali has in the back, right?

And it's ladder logic and it's, uh, 6 11 31, right? [00:10:00] Languages kind of. Uh, so I saw a lot of potential. I saw that this industry is really ripe for innovation, and I was lucky to back then. I'm talking about 2013 and 14. My boss at the time gave me a task of Okay, fine. Design the factory of the future.

That was my task. So it wasn't Princeton in the Siemens campus facility. That really opened me them. The eyes and my brain into this world of robotics and automation. And then I after that, I just wanted to deploy robots. That was my purpose in life. So continuing Siemens, my last role was the head of advanced manufacturing automation for Siemens technology.

That's a long way to say a lot of robots and AI. That's what got me excited. Um, why I moved to the West coast to really answer the question. Deep learning, gonna make a dent in the world of robotics or not? Or are we gonna still, or just another pass? I still remember. [00:11:00] We were working with Berkeley, and that was the first implementation of really AI into the world of grasping.

Grasping used to be something that the holy grail of robotics. Now there are many companies touching that, but still for a robot to grasp reliably an object that it hasn't seen in its life. It was a difficult topic, and it's still challenging to do it reliably 100 percent of the time. But I had a chance to also be exposed, be part of the Inception of the Arm Institute, the Advanced Robotics and Manufacturing Institute in Pittsburgh, being part of the initial forming team and the initial technical advisory committee.

And, and through Arm, I, we did all sorts of crazy robot projects, like Printing 3D structures with aerospace companies for space applications, 3D printing, or not 3D printing, but, assembling with robots, uh, the uniforms for the Navy in Puerto Rico. We did projects where literally we did [00:12:00]bracing with a collaborative robot.

It's like, why did you use a cobalt if you want to put a flame in the head of the robot? Well, because it was cool and we could, right? Yeah. But did someone do that? We did that. Yeah, we had Final coverage of robot and it was it's a bracing operation that requires two persons so we were able to do it with a robot on a person, but definitely applications not collaborative.

So, but the robot was, and we can talk about why we chose a robot, a cobot then, right? But back then I also was, I started to advise different robotic startups that were forming and came coming out of Berkeley through the Skydeck that is the accelerator of UC Berkeley. Okay. I was working with the startups, but not at a startup.

So finally, I did say after a decade of Siemens and leaving an amazing group behind. Continues to do amazing things. I decided, Okay, now it's the time to try this [00:13:00] startup drug and see if it gets me higher. Now, I joined Friday Robotics exactly as their VP of product cycle. It's you to solve a problem of democratizing robotics.

Oh, I saw that back then that it was Extremely hard to there was not really an, like a developer ecosystem in automation with what's a pity, right? All the industries have really take over because it made it so easy for a company to be born to just like, what's up with two guys in a garage and became a billion dollar company tried to do that in robotics.

So ready was, um, addressing that problem, creating an operating system. Horizontal across robotics and automation. Um, amazing time there. We were able to release 4. 0. 5. Now fast forward, like NVIDIA invested in them, Rockwell, etc. Mature company. They just released an announcement of a partnership with Toyota and NVIDIA on creating [00:14:00] applications in the Omniverse.

Great company. And just check out the news release. That's pretty cool. And then after that, I went to Rapid Robotics, which is a great company out of San Francisco. That was in my backyard, so it was easier to see robots, right? Because going to Columbus every time to Ohio, where headquarters already was.

Yeah. Take some time to get into the plane and go all over the country. And when you work with robots, you want to be close to them. So Robots as a Service company with Rapid have a great time deploying robots all over the U. S. Um, and then over the summer, they decided, okay, this is the time to now is my time to get to be the crazy founder because I thought, okay, these founders have an easy life, right when I go, I'm going to do the same.

No, I'm kidding. So we, uh, put together some ideas, kind of look at all the topics. And as much as I like to say that this company [00:15:00] was founded in the love of robotics and it was actually not love what founded it was the hate of the problem, the hate of labor shortage, right? Like every factory that I go, they are running a 70 percent capacity because they cannot find people.

And that's a problem that automation seems to be the clear answer. To the problem, but where are the robots? Why do we deploy 30, 000? So, got together to with Carlos, and decided to find reset, but an amazing initial team, raise some funding. We are going to announce that in a few weeks to And get up and running, got up and running.

And now all hands on deck to see for first platform to make it very easy to manufacturers to again discover. It's on the discover, design, procure and then, deploy automation. That's what we want. We want to touch everything that cannot be just one single silo. We think that this is a fundamental [00:16:00] problem that if we want to solve it, it has to be Through a holistic approach.

So I took I think 15 minutes for the intro. No, that's perfect though. Sometimes. Yeah, there's I appreciate some fidelity in the introduction because I noted down like three different things that now I want to talk about, but. And I should also try to let Allie and Courtney say something if they want to before I start talking and take up the next 15 minutes all by myself.

Oh, I'm just saying hi back to some people. Yeah, and, well, I guess unless you want to answer this question, I was asking myself if somebody in the audience didn't so. What can I help with? Thanks for joining us. Hi, Naomi. So we want to enable actually the manufacturers to choose the right tool for the job, right?

So, yeah, I'm not going to be married with just one robot brand or one solution. It's more about being a [00:17:00] horizontal and not have a hammer and be looking for nails. That's one thing that I really like on it. Like that was a prerequisite on when making a company is do a company that if you're successful, all the industry successful, it's not just you as a company.

And also the same thing. Sometimes you going back to the example of the bracing application, putting a flame on a cobalt, maybe the cobalt was not the best choice, right? If I had a industrial robot in my tool belt. So we are really going all across the spectrum from cobots to high end big industrial robots and ideas that expose those capabilities to SMBs.

And the word automation has been also chosen on purpose. It's not research robotics. It's automation because Automation is also more than robotics, so it can be also beautiful and 3D axes or linear actuators and conveyors and [00:18:00] other amazing things that can also solve the problems of the customer, right?

Sometimes we think that there are tons of videos. Do you guys consider yourself more of a service provider then in that case? Yes, like kind of without hinting too much. Someone that can help a company. I can't help get companies started, but also go all the way through their automation journey and help them design the right solution or pick the right solution on.

And then in the future, also be the partner when they need to deploy either with the ploy or with our partners. That's them. That's the ultimate goal. And again, back to the automation. There are a lot of problems that like a simple automation can solve. It doesn't have to be a lot of videos out there with a six degree of freedom robot that you think, okay, you have three too many.

This is too complicated. You are just moving the one thing up and down, my friend. There are a lot of solutions out there. Yes, this could have been a gantry. That is my [00:19:00] favorite type of video. I think, there are some folks out there in, Lincoln and highlighting those. And that would be my favorite YouTube channel.

This could have been a gantry. Yeah, you should make that. Let's add that to the list of things we'll do when we one day have all the time in the world, which will be never.

Yeah, maybe I just look for that. Yeah. Yeah. Yeah. Yeah. I mean, I think at this point, somebody starting a startup and isn't thinking about it as a platform business, I think is, you know, maybe not well, maybe not across all industries. Right? But at this point, having that end to end. And an open some sort of ecosystem where you're not, like, you said, trying to be a hammer looking for a nail.

I think those are the businesses of the future because we are at a point where we have the ability using partnerships connections data. [00:20:00] To not have to live in a silo in any way. Right? We should not be. And so I obviously draw a little parallel between what you the way you describe the business and quote beam in the sense that we also want to be an enabler and not somebody that shoes somebody into a silo there has to be and in my background as a sales engineer for 2 different brands.

You know, directly at the time it was, you drink the Kool Aid and you go, we're the best for the best. And then you realize the best in a lot of cases, depending on the application, but not always. And I think everybody, you know, now that everybody has this easy access to too much information, what you need is some sort of help to curate that information and shoot you in the right direction.

Exactly. Because right now we have a ton of information, but most of it, you end up running into a brick wall when you actually want to do anything with that information. I don't know about you guys, but I've been seeing really cool concept videos of factories of the future and cities of the future and cars of the future for like 10 years now.

[00:21:00] And then you're like, Oh, this is so cool. And then you realize it doesn't actually exist. Yeah, 100%. And back to being a kind of ecosystem or an enabler, a catalyst, choose your favorite objective. Yeah. The worst thing is going to a factory. You see that they have. Real needs. And you are not the one because you have the comments.

Okay. No, I cannot help you. Or you, this is one that is at least honest. The other one is yes, we can help you knowing that no way, like this is like a straight. And what I love, we love is so I've been a factor and say, okay, look, there are a hundred things that, this industry as a whole, the automation industry can do for you.

We can probably do one, but let me introduce you to 99 other persons. So at least be your broker to help you with this amazing industry, right? The amount, the breadth of capabilities from the companies that we were at Automate is just incredible, right? And back to the issue at hand is like, with all of [00:22:00] that, but we only deployed 30, 000 robots.

How do we make this pie bigger, right? It's not like eating each other's pie because the pie is small and we have too many. So let's make this bigger. That's what, hopefully. We as a company, at least I'm trying to enable, we want to enable that we happen to be a company, as a vehicle to enable that, because we think it's the best vehicle for that.

But the goal is really to help automation thrive in this company. That's the goal. And I think you're absolutely right. The pie is not too small for all of us. It's not, get your piece of it. There, there's some sort of bottleneck preventing people from, you know, getting to it. It is. There's so much untapped potential.

I would almost like to say, you know, a limitless. I think we're really limiting ourselves by somehow trying to compete for the. Available slots of someone that's in the market to buy a robot today, because they already know, and they already have justified that they need 1 or any nation, right? [00:23:00] We help businesses actually get to a point where they can.

Do it, and then there's unboundless opportunity, I think, for all of us. And I know, Ali, like, you've seen that in your business and Courtney the same. I know you guys operate this way, but which is if I don't have the right tool for this problem, someone in my network does. No, no, no, so it seems to me like, you're like, you're.

Obviously, they're to serve the end user. Of automation and help them select the right tools and you know get their company to receive the benefits of automation Do you see your company also? uh taking business from like the manufacturers saying hey, I've got a deployment or something that You know, I have a I don't want to say I have a difficult customer, but I have a challenging situation would you see yourself taking business from the other end, you know, or it's like the manufacturer's pushing into the market The manufacturing that will be the OEMs.

So, so the mm-Hmm. , the robot vendors source. Yeah. I think ultimately, again, [00:24:00] without opening all the, layers of the cake here, but at the high level, what will be very useful in the, what we think will be very useful in this ecosystem. It's to take care of both sides, right? You cannot just go to, okay, I'm going to focus on the manufacturers, but I don't know.

I don't have any tech or I have all this tech in the world, but I don't know what the manufacturers need. There should be something in the middle. Or someone in the middle? That's someone the problem. I think there's a question on somewhere that what is the most exciting thing about the company, right?

Uh, Thomas, thank you for that question. But it links well to what we are talking about is that if we want that to a scale, there should be some level of even Automating automation on some level of being able to sell serve yourself at both ends of the vendor side and on the manufacturer side to discover what is the right solution for you or even what is [00:25:00] the, the best problem to tackle, right?

And answering the question now is using AI to do that, right? That's what we are most excited about. Why this company probably wouldn't have been able to be. In the market one year ago are the advances on a very specific, branch of artificial intelligence that is the large language models, our multimedia or large multimedia models that are not necessarily, having capabilities to the robot themselves, but all the process that go into, into an automation journey.

That's what gets us excited, right? Being on the wave of technology that can make an impact is not the sake of tech for tech. But, it's great to use AI somewhere else than in the robot itself, because when you use it in the robot itself, it's so hard because it has to be 99. 999 times the size for a lot of cases.

So , you have to design around your corner cases and that's pretty hard. That's why, yeah, even though [00:26:00] computer vision has been around for a while, even in robotics, it's not widely spread because there are always this edge case that, oh, wow, like when we were doing. Picking from supermarket items at Seaman.

What if the milk is open? Well, your robot, everything is full of milk. And yeah, cleaning milk out of a robot is not a funny thing. But it happens, sometimes things have spills. And the robot wouldn't know, right? The robot will just continue with its life. It has been trained to so, long way to say that.

Yeah, I'm very excited about that. But using it probably in a different way. That's what has it been used so far? Yeah, I think in its current state is a huge enabler when it's built into products properly that people can use that would otherwise not be able to utilize these models to build their own use case or anything like that.

I like to quote usually. My favorite speaker is Rodney Brooks. Okay. Um, Rodney is amazing, right? He was in a conference, the [00:27:00] first conference in robotic learning, every single robot is excited about robot and AI. In Google. It was held at Google, but many companies went there. He was the keynote speaker.

First slide. AI is overhyped. Machine learning is overhyped. Robotics is overhyped. So just cooling down. Like, okay, I have cold water for everyone. But one thing that he said that is true is that every single successful deployment out there. Every single successful deployment out there of AI of machine learning in the wild has two key ingredients, and you need one of the two.

Either there is a human in the loop, or the cost of failure is very small. Yeah. Beam picking, for example. Doesn't matter, right? You pick, you fail, then you pick again. As far as you are fast enough. So, that's enough, right? That's why they are very good companies like Cobadian, Ambi, Osaro, et cetera.

Tackling that problem very well with AI, but in [00:28:00] interlogistics, right, you don't see much in manufacturing because if you pick it wrong, you put it in a machine and it's wrong. Okay, forget about that. The placing is equally important than the picking. And we are not there in terms of technology yet. And the other part of it is a human in the loop, right?

Discover also there were a lot of autonomous cars that in reality they had two drivers in the back. Uh, that's happening now with humanoids. You see videos that are super cool and then you realize they are teleoperated. And I am a big fan of teleoperation, but that's the same. I was going to say, teleoperation shouldn't be a hidden thing to pretend you're doing some sort of autonomous.

It, I mean, there's a huge value in and of itself. Exactly, exactly. I don't this is the this will be your highlight that you'll do the operation correctly. And these two ingredients. So, uh, that's also why we, you know, in our end is more than the 1st case, actually, that fits that there is a human in the loop or that the cost of.

Messing it up. It's not that [00:29:00] big. So that's why it's so popular, right? It's not that it's always accurate. It actually elucidates quite a lot, but it's helpful and that's enough. If you have a competent human in the loop that can evaluate the answer and check it before doing something with it, it becomes an accelerator for one person.

 Yeah, and it's just, it's strange. I guess the parallels I actually used to work with, electromagnetic simulation software that was used in antenna design for optimizing antennas and things like that. And then I ended up working for a startup doing machine learning and AI, application and supply chain.

And I had an office in Berkeley, and I had a few interns from UC Berkeley, in 2017. And I was speaking at a few AI conferences at the time, and it was very much, you know, a lot of pilots going on, um, not a lot of things autonomous in the wild actually running any kind of neural network and producing result without a human in the loop.

[00:30:00] Um. For successful deployments anyway, and the company that I was with algo was a startup at the time. They're a little bit later stage now, but they've built their entire business off integrating services into their platform. Using the platform to make their services better and more competitive than Mm-Hmm.

you know, not using it. And then the, both the customer and the service team is using that same platform, but they're not relying on a hundred percent on ai. Uh mm-Hmm. , it's really architecting a system that has, it can have traditional machine learning, it can have deterministic, you know, features it, and it also, has the ability of people to do the tasks that make more sense.

And I think that really to me is anyway, kind of the future of this is we utilize the AI tools to help us be better. Some applications over time can become completely autonomous. And I agree that I like the systems view, right? Because when we think about [00:31:00] AI is all about algorithms. Right. And that's what people get.

You don't write so many papers on systems in AI that doesn't get all the headlight highlights, but the algorithms, yeah, I'm better than these other algorithms. So therefore I get some visibility into my work. When it goes into the real world, you realize, okay, algorithms is just one small part. Actually, the previous one was okay.

It's just. How do I look at it from a systems point of view because it's non deterministic. What do I do with this? Doesn't go my way, right? Although I architected a system that can tackle failure modes, which is very different than programming. Traditionally a PLC, for example, right? You want to be very deterministic.

You have to do this in 100 milliseconds. You better program it well, and it has to work. And now what if I tell you that this logic piece of logic was learned? Completely from a I right? That's a completely different paradigm, that we still don't have the very good tools to look at it from the system's [00:32:00] point of view.

We have good tools for creating algorithms. And I think what the industry is now also focusing on more companies are gonna succeed on. Creating the pipeline and the tools for looking at all like very domain specific as a system. Yeah, so that's actually brings up a relevant question from Terry Bray, who we saw again, but he was at that forum with us as well over the last couple of days.

Curious to hear a bit more about your platform, enabling applications already solved to be connected to potential customers. So I think in this case, it might be talking about, you know, not necessarily robots. Right. But automation systems. How can you because right now, I think we have a silo of multiple different systems, integrators or distributors that do value add that solve a problem for someone locally and then never sell that system again because we find that what we do is not that repeatable.

Right? But it could be maybe there's somebody 3 or 4 or 5 states away that actually has that exact application. But how do we know if that is actually something that can be replicated? [00:33:00]Exactly. Yeah, I mean, that's one of the points that I believe. Will unlock higher adoption of automation. If you look back into the what are the fears, why manufacturers will adopt a solution is always the well, I am unique, like this course that we have in the industry that everything is custom.

So, if everything is custom, I cannot reference 10 other customers because yes, I just did it for that 1, but I don't know if it's going to work. Now, let's talk about palletizers, for example, like a box is a box, a pallet is a pallet. There may be like every application, like a layer of between 10 and 20 percent that is unique, but the rest would be custom.

So if someone is really looking into, should I automate palletizing or not? The answer is, yeah, there are like, I can show you a hundred other examples, of other. Applications and advertising happens to be very popular, but the same thing can happen to your point Nikki on we did this dispensing project [00:34:00] and we just did one.

But, what about the hundreds of others that I don't know? Right? So, gaining that visibility of what is the art of the possible. I think that's definitely a scope to what we are addressing here. It is you can really enter the. Uh, this automation journey. If you want, I see that as a journey without even knowing what does to automate where you should end up on this kind of panel or path is knowing.

Okay. These are all the applications that I could automate. This is my ranking is the first one that I'm going to choose because. Both is it the ROI for me and the risk is low, right? It's highly automatable and low risk. These two things can vary So we want to be that kind of Can use a word platform or market or just choose your favorite, word for that.

But yeah, someone that [00:35:00] can say, okay, this application actually will make sense for me. And this was developed, deployed all these times. I get some assurance and you're going to take care of that. This thin layer of, integration also is. Taking care of and I don't have to even worry about that. So, back to why we don't deploy that many robots.

That is one of the reasons, right? It is. Cost is definitely 1, the expertise needed. And the 3rd, 1 is who's going to maintain this, right? Like, who's going to program it and who's going to maintain it. But overall, all of these 3 things is a lack of. Sometimes understanding how much automation cost and how to calculate that.

Why this is very interesting to me. Always there. You've got a factory and it can be one or the other, right? It's either, oh, I thought this was going to cost like, Hey, I mean, like, how is a robot going to cost 5, 000 in which one? And another one of the thing, oh, this was going to cost 1 million, but it's like a 10th of that and kind of [00:36:00] settle the expectations of cost and capabilities.

It's like, what is possible reliably. Um, at what cost? And it's usually a continuum, right? One of the things that we want to be the or that one of the questions that we should address quickly in a sales process is a little bit on the education part of, the equation. Like, are you going to Would you want to input your parts on a bin like you do right now, or are you willing to put them in a grid?

Yeah, it's different, right? We can do both like that. Very cool solutions nowadays to. Pick from a being, pick and place. Now you need a camera. You need the right lighting. All of this has a price. You somehow can, give me this in a different input, then the robot is going to work. This is going to work fantastically.

Right? So, one of the things that I like is let's keep this as simple as possible and it's a little bit counterintuitive because humans, we [00:37:00] are very good at both sensing and seeing and touching and applying logic. But automation has a different set of skills. A robot is fast, it's precise. Without vision it doesn't even know what it's doing.

Sometimes, right, it doesn't, it cannot adapt flexibly. Right. When you add vision, nowadays you can get also very accurate there. But it comes with a cost, right, and nothing is for free. Yeah, and I think in some cases when people try to automate one process or solve one problem, there are upstream and downstream sort of implications that aren't necessarily looked at, especially also when you're looking at the R.

Y. of the process and all of that, it's hard to look at it holistically without some way. And I think I probably is the way to be able to open that up and. Courtney, like you've done a lot of robotics and you get, like it makes it more apparent that you have issues downstream or upstream [00:38:00] because now this process is working so consistently.

So I'm sorry about having to take care of some other part of the process in addition to the thing you just saw. 100%. I think there is a very cool, quote from Bill Gates that I like that when you add automation to a Like non optimal process. It highlights the non optimal part of it, right? When you hide when you are an automation, an optimal process highlights the optimal parts.

So when you are the robot as 100 percent as you said, Courtney is like, okay, you're gonna you're gonna have so some dirty laundry here. You're gonna see that things were probably you. You thought that your process was under control, but every part is not just One micron indifference. No, it's like one millimeter.

Well, what are you talking about? Now? The robot doesn't work. Okay. Yeah, your boxes. Every box is different size. Okay, what is happening with your boss box? How did you put boxes and all of these [00:39:00] things? One of the things that was always fun when I was doing like machine vision for quality inspection, Greg would always point out when I was working with him that You have installed a camera system and maybe a rejector so you can kick off the bad parts.

But you're not actually fixing the problem. You're just keeping the bad parts for making it out to your customers. So a lot of people don't realize that when they install a quality inspection system that now they're basically putting a big spotlight on their other production problems.

It's very important to start with the process. Just identifying the problem basically. Yeah, I think one of my favorite source of quotes lately is a friend from LinkedIn, Chris Sturge, you sorry if I'm saying your last name wrong. I think even chimed in on your post earlier today about a number of robots.

But I think something like if you, you know, an automation system, if you have a bad process is really just a very [00:40:00] expensive quality. Inspection problem or process, whatever, because you end up exactly basically highlighting your quality problems with that. You're not going to get your intended use out of the automation if all it does now is catch your bad.

Parts or your bad process upstream. That was one of the reasons I think why I couldn't stay in machine vision for too long is because I kept moving further and further like upstream as to again, I would like you said, Courtney, I'm like, well, you can see before we kick this part that there's a trend. How do I tie that trend back far enough and know what's causing it to fix the problem before any part ever needs to get kicked out for quality?

So the systems that actually collect that data and send it back to the mothership are of real value. I'm seeing companies that do that now. Like it's not just a machine vision system, but it's collecting data and sending it back like production data and sending it back to the actual production [00:41:00]machines.

Because that's what's actually valuable. If you set a customer up, you know, with their factory of the future, do you guys continue to support that customer as they make changes or add more automation? Or is it more of a process of getting them to understand their own process and then, continue to add automation themselves?

Yes, and I think that ties to some of the previous discussion of can we just be like one provider of a small value? Yeah. At this point in time, it will have to be end to end, right? We cannot just let the, hey, this is my problem. Okay. I'll throw it over the fence. Now. It's your problem. We want to also take care of you all the way.

And if a risk that this road is not going to work as you expected is part of your decisions, why you don't automate, we have to remove it out. Right? So, but all the end to end means owning the [00:42:00] design all the way to the discovery, the design and deployment and all the way to, okay, I'm gonna maintain this for you or help you do that.

Ideally. Is not also one off like we are, an integrated or disappear or as a service company is all the spectrum, right? And we don't want to do it alone. That's your part. So there may be partners with for a specific case. They take it part of the value chain. Yeah, and we work with them.

There are some times where we will want to play that. That role without stretching us to think, right? Every startup has a set of resources and we have to be focused on those set of capabilities. Only the market will tell us. Okay, you have now the right to grow because you have graduated or stay in your lane and get better there as you grow.

But yeah, I think overall, even if it's we do it or we as an industry do it, uh, robots needs to be right. Cheaper, easier to deploy, easier to program, easier to maintain. [00:43:00] And any tool that goes and touches those, I think the entire industry will benefit. So, I just pulled up this question because it was very similar to what you just said, Courtney, but will you provide educational tools, tutorials to help guide these small.

Uh, oh, I guess this settles the question. I don't know. Do we call them? S. M. E. S. or S. M. B. S. we're just having this. Myself. Yeah, but to help them if they don't have any experience. Yeah. So I think a huge part of this, just as a community, as an industry, we need to get our market educated. I was having a discussion actually.

And if you listen to the podcast, the episode that will drop today, I just need to release it after we're done with this. I got to write the show notes and stuff, but I talked with Kathy Wren and Lauren Vandermark from FlexLine Automation. A couple days ago, while we were in Florida, and we were talking about this as well as it relates to safety.

I mean, we just have so many frontiers of our industry that the [00:44:00] public. And therefore our customers that are in manufacturing just don't know and it's not helpful either. When we have the media telling people that robots are attacking when it's a, you know, it's a tag out. Yes, we have to educate the world about what the heck this is and how it works in addition to actually providing the tools and services and all of that.

100 percent that I think we are a unique juncture right now where technologies can really be useful for this purpose. And is that before there were kind of general purpose courses material out there. Yeah, may or may not speak to your problem. One thing that, genia excels is to talk to your problems to customize the entire corpus of automation.

In this case, safety, anything, you name it to your needs. So that if you don't have any A. M. R. S. Maybe that's not the [00:45:00] use case for you, right? Or there is a very specific part of the safety standard where Okay, you cannot if you have a mobile manipulator. Well, I don't care about that part, but I care about the other ones.

What I create something that applies to my use case and technology can do that today. So going back to Ryan's question, that was a great question is, I think without education, all the pieces of the puzzle fall. You cannot just say, hey, this is the solution that you need without you arriving to that conclusion by Okay, with education, with reading materials, kind of becoming, you don't need to become, have a PhD, but you need to have some expertise on why is this the right solution for my case, and why does it mean to, why, even though I have a cohort, I may still need a fence, these kind of things, and what does it mean to, and safety is a very good one, right?

What does it mean in terms for me for safety? For training my [00:46:00] employees. So this, new piece of equipment doesn't become a hazard, but it's actually a multiplier of efficiency. That's what it should be. Good question, Ryan. Thank you for that. Yeah. Another good question from Thomas over here. What is your approach to ethical AI use and data protection? I don't know if you want to get into, that right now, but that's something as we use AI, right? As we use more data, everybody's concerned with privacy. I mean, it's similar to cybersecurity, right? Like the more we do have this, the more it becomes something we have to focus on.

And in the article about the robot attacking also made a broad stroke statement as we get more automated facilities, we're going to see more robot accidents. And it's like, yes, you are just by default. If there's more robots, there will be more robot accidents. But like, take a look at the number of industrial forklift accidents and tell me we shouldn't have forklifts.

More cars to the freeway. There will be more freeway accidents because there's just more cars. Yeah. Yeah. Which is why [00:47:00] as we've been driving cars for. You know, decades now, we have more safety features in cars. As we've learned, we have to focus on these things growth and, you know, protection at the same time.

But yeah, I'll let you tackle the question if you want to the important thing on that topic that you mentioned on the safety aspect of that is that the systems can be designed and should be designed with safety in mind in the 1st place. And that requires some procedures that need to be controlled. And that's why you see more automation in factories than in the roads, because that's an uncompletely uncontrollable environment.

Like, if you ask one of the engineers, uh, autonomous cars company, what type of edge cases are they working right now? I mean, yeah, a person, working backward. Following with a wheelchair and I gets in the way of the car, like it's really super edge cases that in manufacturing, fortunately we can have a higher control on at least saying, okay, this is my operational, how I wanna run this unit.

Everything [00:48:00] around that we need to put, the mechanisms to keep it out, to keep it safe. Now back to the question from Thomas, which is a great question. I think as an industry we are definitely. We should be looking into this. As you say, Nikki, and that's one of also the reasons why I'm now inside a three season for advancing automation.

That is, there used to be just three pillars robotics, matching vision. And then all the other control side of things on the machine controls, and then there is the fourth pillar, like a new. Organization a new tech board on a I recently since since last year, but I'm also I've been honored to be there since the very beginning on that board, right?

Exactly. And that is something that we have to address also as an industry in terms of, okay, at least give enough guidance and recommendation of what does it mean to what data is needed? What does it mean to use to have a [00:49:00] safe application of a facility? So, those are topics in mind of us as an industry.

Definitely. There are also this is also not a black and white. There's an spectrum of applications when you are collecting. I don't know, facial data from. Uh, from people. It's not the case for us, but there are companies obviously doing that as part of their computer vision algorithms. Then you need to be able to make sure that this is anonymous, right?

That's a very important part of thing. What is and I think to me, the most important thing is that the customer knows what they are getting into, right? But by using certain algorithms, they are enabling the collection of some data And they have to opt in. I'm doing, I'm doing that. That's not by default.

Hey, by the way, I'm collecting all your process data by because you install that pieces of software that has to be extremely clear of where each part of the data. Where is that used? And, but also from the point of view of the customer understanding that [00:50:00] there are very good technologies out there and people looking into, uh, relaxing those concerns.

That, they may have hopefully answer your question. Thomas. Yeah, we'll see. It remains to be seen like I we're not going to take the mo of kind of the iPhone apps that, you know, the fine print is so long that you're not going to read it. Right. I didn't know that face app has perpetual license to use my face for the rest of my life.

I guess the rest of their existence because I used it. Before yeah, but I did and what am I, am I really going to read the, you know, however many page fine prints there are when you download an app or something. And as we found out, I mean, I'm sure it's not common, but if you've seen, and I know this is not probably not factually accurate, but the, um, super, what, what is it?

A super pumped that the Netflix show about the. Journey of Uber. Um, they even, you know, skirted the rules of the app store for a while. So we can't [00:51:00] necessarily, you know, I think we all have these experiences. Now we realize that we can't trust the big tech companies with all of our data. And, they kind of just throw these terms and conditions at us that we have.

What are we going to read them and say? No, I'm not going to download this app because they found something in the fine print. And I think that that is, an ethical. Consideration. Do you try to bury it? And yes, people will opt, you know, they'll without their real consent. They'll give consent.

Do we want to be listed about what people are giving consent for? And then you really have to give them the choice to opt in rather than opt out. Because just. You know, the rates are if you basically, I don't know, get like 80 percent of people will just continue if they have to opt out versus the other way around.

If they have to opt in. I think 1 more question from Thomas, since he's been very active asking us questions. Thank you for interacting with us so much. Thomas. Do you see most businesses feel if they're not implementing some aspect of integration? That's a good question, right? And I hope I had a crystal ball.

I think it's a [00:52:00] little early, right? It's a little early to know, but I bet the answer will be yes. Um, it's probably like in the early 2000s or late 90s when someone said, okay, replace AI with the Internet. And then there you go, right? Would you fail if you don't have Internet integration? There are companies out there that have been going on forever.

But the reality is that those That use those new tools and the capabilities that they enable. And they do it in the right way. They will see a multiplier. Yeah. And it's a business that True acceleration. Exactly. It's transformational in their capabilities. It's also not a magic wand or silver bullet that will solve all the problems in the world.

But it adds a level of functionality and customization and integration of knowledge that no other technology has really [00:53:00] been able to offer until this point. So, in some sense, there will be winners and losers, and I think the majority of the people adopted in the right way, not for the sake of I use AI and I don't care about the use case, but whenever they have a use case.

They are going to see the gains as we hopefully are part of that process. It's not necessarily failing if they don't use AI, but the ones that are using AI are going to be the big winners. I would say if you make a point to never use anything that uses AI, you will probably eventually fail. Or your business might stay the same until you lose whatever current.

Business environment and customers. You have no, it's not. Yeah, it doesn't. It's not that it's making you fail. It's that if everyone around you is accelerating and you're not able to do that same amount of work or that same quality or that same repeatability, then, yeah, you're eventually going to be irrelevant.

So, yeah, and you can absolutely I've heard a lot of people say that they're not going to use [00:54:00]a I and, Good luck. I think sometimes the distinction is not, you don't need to integrate AI directly into your business. You just need to be using tools that use AI to help your business in the process, right?

Do you use QuickBooks? Do you use Excel? Do you use, you know, all of these tools or softwares or whatever that you use in your business will be integrating AI, the good ones anyway. And so you may be taking advantage of all the AI advancements without ever touching any kind of AI development yourself.

Report that was like 2024 is going to be the year of platform businesses. Everybody thinks it's AI, but it's really the platforms or companies that are using the AI or the application of them. Not the AI itself, right? And, having been in this business or talking about it for a long time, like AGI, a general artificial intelligence that's going to cover and do everything.

I don't, A, we're not really anywhere near there. And B, [00:55:00] that's in my opinion, not what we need anyway. I will agree on that. I usually will say even if we get it, that's not that shouldn't be there. And I think, even the industry has relaxed a little bit. The definition of A. G. I used to be like sentient, artificial beings.

And now is well, some something more intelligence than average person. Okay, that's probably more achievable. So we define if we change the definition, then I'm more on the boat. But otherwise, even without driving to that point, Okay. There are so many other useful technologies that we can brand again as AI, and the same conversation that we are having about AI, replace that with automation and robotics and apply that to manufacturing and say, Those who don't automate or in the bigger definition of automate or improve their processes, they're going to have a hard time surviving in the next five years to the point that even face extinction because labor is not going to get better.

 That's a reality. And then what [00:56:00] miracle there is no other miracle that will. I will save them. So only robots automation can be that force of repeatability, quality growth and all the good stuff that comes with it. That is not just definitely not human replacement. It's everything.

It's more around growth than anything else. But also, like, good luck attracting or the future workforce into a job that has no or digital anything. Right? Like, come in here and push buttons and write things down and read the spreadsheet. Like, those kids are going to be out of there in a hot 2nd. Can I also push you and pounding some railroad stakes?

So I know we're up basically on time. We're at an hour. This has been so much fun. We have a couple more questions. I want to throw up before we head out. And say hi Jason Jones. Thank you for joining us. I'm so sorry about your frozen water pipe Like to me A couple of years ago and [00:57:00] it's a pain in the butt.

 I hope that works out for you. Okay, and isn't too terrible and we really appreciate you coming anyway after dealing with that. And then, Jamie Callahan has a question for one will reshapes project management include processes to keep suppliers and manufacturing companies working together with integrity and good communication.

All this is a good one. So I promise I told with Jamie on the, automate, but I didn't pitch this. So this was not prepared. I promise. But that is also one of the goals. Um, I know that. Pain that I see on, I see myself sometimes, acting as, uh, or like with integrators are working together with them. It is the, okay, how do we even exchange?

I need to exchange pictures of, deployment. How do I do that? Oh yeah. We don't over email. That's not the best way to exchange pictures. Now we need to sign an NDA. What is that document in the entire process? I lost about that. Or when it's worse, once you are deploying something. Are you going to be ready that a [00:58:00] robot shows up in your doorstep by that date?

Are you going to have power? Are you going to have electricity to the part of the factory that requires that? A lot of times that's not the case, right? Because there are long cycle times. So communication is the key. It has to happen. Otherwise there are, it only leads to frustration, to delays. And right now there is no such a tool.

So, that is something that's definitely in a scope and will be great to. Uh, again to make this industry as efficient as possible. So thanks for the question, Jamie. Yeah, it's a great one. And that's one of the things that early early on and before I even joined quote being back when Roman, you know, was talking about it.

We had this meeting in a coffee shop in Seattle, and I was just thinking, you know, if we look at all of the issues or inefficiencies in the process of and we're talking about systems integrators and custom machine builders working with their suppliers, the component suppliers. Because this problem kind of replicates itself on many different levels.

 In the industry, [00:59:00] which is project management and communication with multiple stakeholders. From multiple firms, and there's many stages to this, and oftentimes you end up having to loop back because something changed. And right now, everybody's managing basically all steps of that process through emails, phone calls.

Spreadsheets, um, and then in person meetings and all of those parts are disconnected. And it, you know, if we think about this as a manufacturing process and, and you map this out like a factory, it's ridiculous. Like we would never make anything in this fashion if we were actually making a widget, but we're okay with making the things that make the widgets like this, this whole time.

That is, I guess every country has a way to say that, like, I don't know, in Spain, we say that The house of the metal worker. Did you use like boot spoon or something like that? Oh, yeah. What? The cobbler son is [01:00:00] wearing no shoes.

Great. But yeah, thank you all for joining the great questions. Really lively conversation. We are, we're really super happy to be back. Now when we're in London and have something to sell. Uh, yes, we are not, we don't want to be on this alone, right? This is like, uh, Yeah, it requires a bill. It's So any other feedback like this, , any questions, any ideas on how do we make, how can we, as an industry, improve and automate faster and at a higher rate?

Yeah, we want to hear from you. So this is, we have this together. How do we get in touch with you, Juan? How do customers get in touch with you? I am known to be somehow active on LinkedIn. So LinkedIn is probably the best. The best avenue or one at resaveautomation, this one. So you can also send me an email anytime.

And I will be very happy to jump into our conversation. [01:01:00] I'm always ready to talk about robotics and automation and manufacturing. So it's my softest spot. I was going to say one, we'll have a guest profile on automationladies. io. That will be tied to this episode. Um, and we can go back and update in the future.

If you have any future links, or places to contact you as as the company grows, as you do more things, um, we can try to make sure to keep that up to date so that people can follow you in the journey. That's beautiful Nikki. So you have the answer all the way. Great. Thank you so much. It has been a pleasure being here.

Thank you for inviting me. Likewise. Thank you for coming. Thank you everybody that joined. If you're watching this later, feel free to still jump in the comments and ask your questions. The Automation Lady is as well as Juan. If you, you know, tag us on LinkedIn, we'll enter the comments and, you know, continue the discussion.

The Internet is forever, so this will live out there somewhere. And we don't want it to be a one time thing. We, we [01:02:00] want the broader automation community to connect, work together. And this is a big part of that. We appreciate people like you on that are building things in our world to do that, and part of this community.

So thank you everybody. Um, have a great rest of your day. Thursday, a new episode from automation ladies with Kathy and Lauren from the A3 forum will be dropping a little bit later today, and then this one will be coming out, probably on the audio podcast next week. So thanks, everybody. Have a good night and we'll see you.

Uh, I think we have a live scheduled next week as well. I've got some admin work to do. Bye bye.