Building a Robotics Company in Ohio: Path Robotics and Drive Capital on Midwest Venture, AI, and Manufacturing
Manufacturing Happy Hour
This live episode of Manufacturing Happy Hour, recorded in front of a packed room in Columbus, Ohio, asked whether the Midwest can build world-class technology companies, and what that looks like in practice. Host Chris brought together Andy Lonsberry, CEO and co-founder of Path Robotics; Nima Gard, Path's Director of AI; and Nils Root, who runs marketing at the venture firm Drive Capital. The guests agreed on the main point: Ohio is not "the next Silicon Valley," and it doesn't need to be. Their case is that the region can produce serious technology businesses on its own terms, drawing on its manufacturing base, its talent, and a startup culture that is slowly maturing. Along the way, the Path founders explained how their technology works, why they treat data as their most valuable asset, and how they think about automation and jobs.
"Absolutely not": why Ohio isn't the next Silicon Valley
Chris opened by asking Root directly whether Ohio is the next Silicon Valley. Root first gave context on Drive Capital. He described it as the largest venture capital firm between the coasts. It is about 12 years old and was founded by two Silicon Valley VCs on the thesis that more great technology businesses would be built outside Silicon Valley than inside it over the next 20 years. Drive invests in places such as Columbus, Chicago, Atlanta, Toronto, Denver, and Austin. Root's view is that technology companies will increasingly be built where people and industry already live. Root has been at Drive for six years and handles PR, content, and events.
His answer to the question was "absolutely not." He compared it to calling Atlanta the next Hollywood because films are made there, or calling Montgomery the next Detroit because it builds cars. By his account, Silicon Valley was a special historical case. Building technology companies once required specific infrastructure and a talent pool that could manage it, and both were concentrated within a 10-to-20-mile radius around San Francisco. That changed around 2007–2008 with cloud computing. With an internet connection and a credit card, anyone could rent storage and compute from AWS. What remained scarce outside the coasts, in Root's telling, was venture capital to fund those companies.
So Ohio's opportunity is not to copy Silicon Valley but to become "the best version of Ohio possible." As evidence, Root pointed to Path Robotics and Root Insurance, which he said was started in Drive's office and is now public. He also noted that established tech companies such as Anduril and Intel are coming to the state. He credited Ohio and Columbus with being business-friendly and said he doesn't know whether there will ever be another Silicon Valley, just as there isn't another Hollywood. Chris said he had deliberately set up the question after hearing the "next Silicon Valley" framing elsewhere, and he endorsed Root's reframing.
Path Robotics' origin story: a garage fab shop, humanoid robots, and a $300,000 first customer
Asked why Ohio was the right place to start Path, Lonsberry told the company's story. He was born and raised in Ohio in what he called a standard Ohio manufacturing family. All of his grandparents and great-grandparents worked in manufacturing, and many were welders. His father has spent roughly 40–45 years in manufacturing and turned the family garage into a fabrication shop with a mill and a lathe. Lonsberry started building off-road vehicles there at age seven. When he was ten, his father quit his job to start a company building two- and four-wheel off-road vehicles. For about four years they ran it out of a roughly 5,000-square-foot facility in Warren, Ohio, with just the family and a handful of employees. Lonsberry said the experience showed him firsthand how hard it is to manufacture a product, and how hard it is to find and keep good people to scale a business. That lesson stayed with him.
In college he was studying mechanical engineering when he first saw Boston Dynamics. It made him want to make mechanical systems intelligent. He interned at the Institute for Human and Machine Cognition, which he described as a facility of MIT PhDs working with Boston Dynamics on bipedal robots and exoskeletons. He then spent six years on a PhD in deep reinforcement learning for humanoid robots, essentially teaching robots to walk. He noted the irony that Path doesn't work on humanoids. His brother was doing a PhD in computational neuroscience at the same time. Drawing on their robotics background and their manufacturing upbringing in Ohio, the two decided to build a company around machine intelligence and computer vision for robotics. The goal was to make manufacturing, which Lonsberry called one of the oldest and hardest businesses to scale, easier to grow.
They talked to about a hundred U.S. manufacturers in Ohio. One, Corsa Performance, agreed to pay about $300,000 for what Lonsberry described as the first fully autonomous welding cells. He stressed that these were genuinely autonomous. An operator put in a part and pressed a single "go" button. The system then had to 3D- and 2D-scan a generic area, recognize what should be welded, decide how to weld it, collect data during and after the weld to judge whether it was good or bad, and feed that data into a reinforcement learning algorithm to keep improving.
With that customer in hand, Lonsberry flew to Silicon Valley expecting to be dismissed. He said people there asked where Ohio was, and some assumed Ohioans "ride cows and chew tobacco." Still, the two PhDs found a first investor who gave them seed money. Back in Ohio they met Drive Capital. Lonsberry described Nick, Path's board member from Drive, as relentless, constantly texting and calling to be part of the company. After closing the deal with Drive, the founders moved to Columbus, reincorporated as Path Robotics, and set out to be a machine learning company that makes robots smart enough to help manufacturing.
Not a welding company: welding as the "heartbeat," assembly as the next task
Chris said Path seems to have become more than a welding robot company. Lonsberry agreed that people usually label Path as welding automation, and said that isn't entirely accurate. Welding is the first product. What defines the company, in his account, is its technology stack: using data and machine learning to get systems to learn tasks. He called welding "the heartbeat of the company," the thing Path wants to be the best in the world at. But the ambition is to handle the tasks before and after welding as well, giving manufacturers the labor capacity they need to scale.
The main task Path is working on beyond welding is dexterous assembly: putting parts together, using tools, and fitting physical components. Lonsberry said large OEMs and small and mid-sized manufacturers alike have a big need for this. It also sits directly upstream of welding, because two or more parts have to be assembled before they can be welded. Path is expanding into assembly using the same intelligence stack it built for welding.
Chris asked how Path chose assembly as the next application out of many possibilities. Lonsberry tied the choice to the company's business model and its focus on customers. Path sells robots as a service, so it only gets paid when customers use the systems every day. He contrasted this with purely capex robotics sales. In that model the vendor delivers a system once for one use case, and it's up to the customer to adapt it to the next one. Path, by contrast, has to keep providing value even when customers switch parts. He said that happens constantly in metal manufacturing through change orders, SKU changes, and shifts in demand. So the company asked what would give its existing customers the most additional value. Assembly was the step before welding that most relieved their labor pressure and could feed the welding cells.
The technical approach: one network, world models, and the long tail
Gard described Path as a modern AI robotics company. It doesn't think of itself as doing welding or assembly but as helping manufacturers automate what they need. To do that, tasks have to be broken into abstractions that apply across applications. First there is infrastructure that can handle various tasks and move robots around, whether the job is painting, grinding, or welding. Then there is perception: robots need to see what they are looking at and where they are going. Gard said this is handled by a single network that learns across application types. Finally the robot performs the task itself, which might be welding or dexterous assembly.
The team's current focus, Gard said, is building a world model. He explained why. Data is the bottleneck in robotics, and in his view the reason robotics hasn't had a "GPT moment" is that there is no internet-scale dataset for physical tasks the way there was for text. That matters in manufacturing because of edge cases. Gard described a small number of common problems and a large number of rare ones. He estimated, with uncertainty, that the rare ones add up to maybe 20–30% of situations. Manufacturing also sets a high bar for accuracy, quality, first-pass yield, and cycle time. To handle this long tail, models need to encounter edge cases and learn from the mistakes of other robots.
Gard laid out two ways to get that data. One is simulation. In his account, simulators take a lot of money and time to build, and even then the data isn't accurate enough because of the "sim-to-real gap" between a digital twin and what actually happens on the shop floor. So simulation alone won't be enough. The alternative, which Gard said Path believes is the answer, is a world model. It is trained on real data, learns to predict the next step, and can be controlled. He described it as a neural network acting as a physics-based engine that learns the laws of physics and can be used to perform different tasks. Path sees this as its future "data engine." When Chris summarized this as needing both simulators and next-step models, Gard corrected him: these are two alternative routes, simulation versus training networks on real data.
What's changed recently: from task-specific models to generalization
Chris mentioned that a Path colleague had said the company may be moving faster than Moore's law, and asked what had improved most in the last quarter or six months. Gard described a change in mindset. A year or two earlier, the standard approach was to collect data and train a model for one specific task, and to repeat that for every new task, customer, or part. After GPT, the thinking shifted to training on many tasks at once so the model generalizes. Path has made the same shift. Gard said that on the perception side, the model no longer depends on which customer, state, country, part, material, joint type, temperature, or lighting it faces; it can handle all of them. Similarly, for assembly, it doesn't matter whether the parts are two flat pieces or cylinders; the system can pick them and put them together. He attributed this to moving toward generalization with one large network.
Lonsberry added that the biggest change is the volume of data as more cells go into the field. Every deployed system collects data every day, which he said has an exponential effect. The data spans time zones, lighting conditions, countries, and ambient temperatures from 0°C to around 40°C across North America. Path combines all of it to train one large foundational network. He said the company is growing about 3x per year, and that this figure covers data as well as revenue. He called data "the gold mine that we care about the most right now." Chris highlighted that framing as a concrete version of the common claim that data is a company's most valuable asset.
What excites the team: embodied AI and real deployments
Asked what excites them most, Gard first named the team. He said he gets emotional and gets goosebumps seeing everyone aligned on a mission. Second was the work itself. He said large models for text, image diffusion, and video generation are great, but what excites him is embodied AI: putting machine intelligence into a real, physical robot. He said Path is at the forefront of that.
Lonsberry's answer was deployment in the real world. He said the robotics hype cycle is "through the roof" and that much of what circulates online is hype rather than reality. His rule of thumb is that any demo that hasn't gone to a customer site is three to ten years from being real. He said Path lived through this three to four years ago. The company had an incredible demo, but when it delivered to a customer, "everything breaks." He said anyone who has delivered a real robot knows it's the most painful process. Path went through what he called the pit of despair and came out the other side. It is now scaling real systems aggressively. Lonsberry said he thinks there are maybe two or three companies in the world, adding that he isn't sure, that actually deploy neural networks trained with reinforcement learning to customers who use them every day.
Standing out when everyone claims to be an AI company
Chris noted that every company now calls itself an AI company and asked why an outsider, whether a customer or a prospective employee, should care. For hiring, Lonsberry said people should come meet the team, which he described as electric and motivating and said he learns from daily. For customers, his answer was also "come see it." Send the parts and watch the technology run. He said many demos lead to a signed contract within 30 days.
He credited Heather, Path's chief revenue officer, who had left early to catch a flight. According to Lonsberry, the current sales cycle is about 80 days from first meeting to close, with deal sizes of roughly $5–10 million. Asked what other manufacturers could learn about shortening deal cycles, he joked that the answer is to hire an amazing CRO. He then described her approach. She aligns Path with a prospect's three major strategic initiatives for the next three to five years and shows where Path fits. She also disqualifies prospects faster than anyone he knows, so the team focuses only on companies with a real need that Path can deliver value to. He gave an example. For one prospective customer, she researched their biggest customer, that customer's growth rate, and the prospect's most pressing need. She then walked the prospect through their own strategic initiatives and built a case for how Path could accelerate them. In Lonsberry's words, "it's not about the tech, it's about the value," measured in revenue, growth, and profitability. He admitted the sales process was about five times worse when he ran sales himself.
Chris, a former salesperson, pulled out three lessons: tie the pitch to the customer's strategic initiatives, disqualify quickly instead of chasing poor fits, and have the self-awareness to recognize when someone else does a job better than you.
What other regions can learn: the ecosystem around startups
Chris asked Root what other regions could learn from Columbus. Root said he wasn't sure how actionable his answer was for individuals. His main point was that building an innovation economy takes more than startups. It needs a supporting culture. That includes law firms that know a term sheet doesn't need to be 100 pages or take six months, and landlords who will offer one- or two-year leases because a startup won't sign a five- or ten-year lease. More broadly, people have to understand startup risk. Root said that today in Columbus, if you told your spouse you were joining a venture-backed startup, the reaction would be enthusiasm and a question about which one. Ten or fifteen years ago, he said, the reaction would have been worry about risking your career with kids heading to college. He described this as a cultural shift that takes time and open-mindedness. He also saw it happening across America, not just in Columbus, and said he didn't think Columbus was ten years ahead of, say, Dayton.
He added a distinction Drive uses: scary versus dangerous. Joining Sears in 1995 wasn't scary, but it was dangerous for your career because Sears is gone. Joining Amazon in 2010 was scary, a fast-moving company that would intimidate someone coming from a company like Nationwide, but it wasn't dangerous, because of the experience and skills you gained. Root put Path Robotics in the same category: a startup may feel scary, but it gives people the skills for where the economy is heading.
Are the coasts changing their view of the Midwest?
For the last main question, Chris asked whether attitudes toward the Midwest are changing on the coasts. Root went first and said, "absolutely not." Nobody in San Francisco is asking how Illinois is doing, and he said that's fine. The coasts play their own game, and the Midwest doesn't need to worry about how it's perceived. He said people should be bullish on Columbus and Ohio because a world-class technology company can be built anywhere. Asked why the Midwest needn't worry, he pointed to a growth mindset and the amount of innovation left to do. Social networking has been figured out, he said, but industries such as hospitals, manufacturing, and transportation still need modernizing, and that is where the opportunity lies.
Lonsberry agreed. He said coastal investors still routinely ask Path when it will move to Silicon Valley or the East Coast, and the answer is always that the talent is here. He cited Ohio's population of roughly 22 million, by his estimate, and research institutions such as Case Western, where he and his brother studied, Ohio State, and Carnegie Mellon. He recalled a statistic, which he said he couldn't remember precisely, that more software engineers graduate in the Midwest than on the coasts. He contrasted the cultures. He described coastal engineers as moving from company to company, pointing to LinkedIn profiles listing ex-Twitter, ex-Meta, and ex-Google. By contrast, he said the Columbus talent pool is bought in, teachable, willing to grind, and wants to win, and that "there is no IQ test" showing Ohio below the coasts.
He also made a strategic argument. Path is one of very few AI robotics companies in the Midwest, while many are forming on the coasts, yet their customers will mostly be in the Midwest. Path chose to stay close to where its equipment is deployed so it can respond quickly, work in the weeds with customers, and learn from them. He said he wouldn't build the company anywhere else.
Audience Q&A: other rising startups and what's really advanced in AI
An audience member who has been a supplier to Path asked which other promising companies or industries in Columbus and Ohio deserve attention. Root suggested Lonsberry might know the manufacturing startup scene better. He said Drive constantly looks at companies automating work, whether in manufacturing or back-office healthcare tasks, and that many come out of Columbus. His explanation was that a critical mass of people working on related problems starts to feed off each other, as happened in Silicon Valley. He noted someone in the front row wearing a Ready Robotics sweatshirt, a past Drive investment that no longer exists, but said he couldn't name a current Columbus manufacturing investment on the spot. Lonsberry said he didn't know of other AI robotics companies in Columbus specifically, only one in Pittsburgh a few hours away. Chris joked that Pittsburgh may not like being called Midwestern.
Another attendee asked Gard what real advance, beyond the buzzwords, he had seen in AI over the past 12 to 24 months that Path is applying. Gard repeated the main theme: larger datasets and larger models that generalize across more tasks instead of being specialized. He said Path doesn't plan to stop at welding and assembly. It will keep expanding to the processes before and after, and it doesn't want to reinvent the wheel every time it onboards a new product, task, or customer. The goal is to make that onboarding frictionless.
Automation and jobs: displacement, new roles, and the labor gap
Leslie, from the event's sponsor, asked how humans and robots will keep interacting as AI extends what robots can do, given the common claim that nobody loses their job and workers simply upskill. Gard didn't fully accept that framing. He said it's fair to admit that some jobs will be eliminated by any new technology. He gave the example of the person who fed and cared for horses before cars. He also noted that barbers 200 years ago also pulled teeth, which seems absurd now. He paraphrased an idea he attributed to Steve Jobs: when a new technology arrives, people first apply old problems to it, and only over time learn how to really use it. His conclusion was that there will be some displacement, but new jobs will be created, and that humans are poor at thinking about exponential change but will adapt and create new possibilities.
Chris offered his own view. He said the "GPT moment" shifted the automation story from lower-wage manual jobs to highly paid knowledge work. Automating a $20–30-an-hour job saves that much per worker, while automating a lawyer billing $300–400 an hour is a completely different calculation. He pointed to uploading a contract and asking an AI how it's unfair as the kind of task that changed the conversation, and suggested we're only beginning to see who will be automated. Gard agreed that, from a manufacturing perspective, physical jobs are much harder to automate than legal or medical work. He suggested regulation is one of the main things protecting the medical field. Chris summed it up: the more your job lives on a screen rather than in the physical world, the more exposed you are.
Lonsberry, leaving for a 7 p.m. call, took a different angle. He said the gap between jobs needed and workers available is still very large and growing. In Path's experience, no one at any company it works with has been laid off or replaced by a robot; the robots have always been an addition. He expects that to hold for a long time. He also said there will be plenty of opportunity in learning to work with these systems to become more efficient as the economy and population grow. He said he doesn't see it as "tomorrow it's robots, goodbye humans," but as filling a gap so manufacturers can accelerate. Chris added that he had heard there are around 2 million manufacturing jobs open or about to open as workers retire, and said he sees automation as repurposing jobs, as has happened for centuries.
Closing: building the next workforce
The evening ended with two short segments on the workforce pipeline. Josh from A3, the Association for Advancing Automation, a former K–12 teacher who joined the trade association the previous November, described its education push around the Automate show in Detroit. He said the show will include an education pavilion and an educators day, and that nearly 400 students were expected, with the number still growing. A3 works with programs such as VEX and FIRST to get young people interested in automation and manufacturing, which he called the next workforce.
Finally, Adrisu, founder of a Columbus nonprofit that the tour supported with a portion of its proceeds, described its mission: getting underserved and underrepresented youth interested in STEM careers through hands-on, project-based workshops in schools, libraries, and community centers. Adrisu studied mechanical engineering at Cleveland State and then built and tested rocket engines in the California desert. Speaking with kids at STEM outreach events there showed Adrisu how eager they were to learn. After returning to Ohio to raise a child, Adrisu started the organization to show that interesting career paths exist and that STEM is for anyone who is curious, not only students in wealthy school districts.
This is going to be a very unique episode of Manufacturing Happy Hour. I assume not everyone here has listened to the show and that's A-okay. Normally when we do this it's like an interview podcast, like it's a one-on-one interview. Tonight we're going to have like a rotating cast of characters throughout the evening. So we've got a main squad that's going to be up here for quite a while. First, I want to thank all the people that have made this possible. And what better way to do that than just bringing them right up on stage. Leslie, welcome to Manufacturing Happy Hour.
Thanks, Chris. It's fantastic to be here.
And thank you to CLA for making this whole tour possible. I know there are a lot of CLA folks here. Round of applause for them. So, I think the easiest way to do this, to do a little pitch for CLA, is to ask the typical Manufacturing Happy Hour question. How do you describe what CLA does as if you're having a drink with someone?
Absolutely. And I'm going to raise my glass to some water because this is my third night doing this and my liver is a little tired.
Fair. Yes, it is. We put an aggressive schedule together for this tour. So, this is night three of four and I don't think we've seen a room this full yet. This is pretty cool to have everyone packed in here.
Fantastic. But anyway, I would love to tell you guys all a little bit about CLA. So CLA is a CPA and professional services firm. And what I would tell you is that we exist to create opportunities for our clients, our people, and our communities just like this Columbus. So raise the roof here, Columbus. Woo.
And so what I would also tell you is that we are industry focused and I think that's very unique. So, I have the joy and the privilege of leading our manufacturing industry. That's why we're together on stage and why we've been peanut butter and jelly all week. And when I say professional services firm, we do have the traditional CPA services, so audit and tax, but we also have so much more than that for business owners. We have wealth advisory, we have outsourcing, and then we have digital services. So, we really want to wrap our arms around businesses really from start through succession to really know them and to help them be successful. And that's what we're all about.
Well, cheers to that. Hey, another round of applause for CLA for making this whole tour possible. Thanks, guys.
Thank you so much.
And I want to bring up our local sponsor for tonight, Solace O'Brien. So, ironically, Brian will be speaking on behalf of them. So, same question for you. How do you describe what Solace O'Brien does as if you're having a beverage with someone?
Yeah, thank you. Appreciate that. Solace O'Brien is an, or at least our local chapter, I guess, is an industrial-focused engineering consulting firm. So, we get involved with our clients from concept ideation through design, detailed engineering into construction, startup, commissioning, and then ongoing operations all the way through. We can get involved in that process at any point for any duration. And we're really looking for long-term relationships with our clients. Our first client we started our business with in 1948, and they are still a client and one of our biggest clients here in Ohio.
Well, excellent. Hey, give it up for Brian and Solace O'Brien. Thank you for making tonight possible as well. And I'm going to switch seats here because we've got three individuals coming up after this.
One of the things that we wanted to do here in Columbus, Ohio is, I live in Milwaukee, Wisconsin. So, I'm going to turn the volume up real fast here. I'm hearing some... Let's see. Is that better? Everyone here? Excellent. Perfect. I'm glad we fixed that towards the front end of the conversation. That's always good.
Hey, Columbus has been a city that, funny enough, I actually started my career here when I was a sales, not a sales guy, an engineer at Anheuser-Busch. I spent a lot of time at their metal container can plant. So Columbus has a special place in my heart. But as a Midwesterner, I keep hearing more and more about Columbus on a regular basis, the tech scene that's here. And that's what we're going to be exploring tonight. So I want to welcome up our crew. Let's do this like you're coming out for a game show. Andy Lonsberry, welcome to the stage.
Nima Gard, welcome to the stage. These gentlemen are from Path Robotics, but our third guest, Nils Root, is from Drive Capital. Welcome to the stage, Nils.
And since we were just mic checking on the front end, everyone speak into your mics and we'll just judge based on thumbs up real quick. How's the volume, guys?
Check. Hello. Check. Check. One, two.
Okay, good. We're getting thumbs up. Great. Well, we're here to talk about venture investment and startups in the Midwest. And this is going to be a bit of a celebration of Columbus, Ohio. It's also going to be something that manufacturing leaders across the country, across the world in some cases, listen to. So, we're going to be teaching them things that, hey, what's working well here with your organizations, with Columbus, Ohio, that other people can take back to their regions as well.
So my first question, Nils, you're going to kick us off with this. Is Ohio the next Silicon Valley? And as you answer your first question, I'm going to ask you all this question. So introduce yourself, your role as well, so the audience knows who you are and who you're with.
Sure. So I'll answer that question, but first I'll give a quick intro to Drive Capital because I think that is an important context here. So Drive is the largest venture capital firm between the coasts. We all have an image in our head when we think of venture capitalists. I think most of us think of typically somebody in Silicon Valley, maybe New York, maybe Boston. Those are pretty much the three markets where the vast majority of venture capital, at least traditionally, has existed. And there's various reasons for that and those reasons no longer exist. So more and more you're going to see venture capital dollars and firms live outside of places like Silicon Valley.
Drive is one of the first firms to kind of sniff that out and invest behind that. So, a 12-year-old firm started by two Silicon Valley VCs who realized that more great world-class technology businesses will be built outside of Silicon Valley in the next 20 years than will be built inside of Silicon Valley. And so Drive is just investing behind that thesis in places like Columbus obviously, where we're located, but also places like Chicago, Atlanta, Toronto, Denver, Austin. All the places where people live and industry lives is where you're going to see technology companies being built in the future. Path Robotics being a great example of that.
So, I've been at Drive for six full years now. I am in charge of all things marketing, whether it's PR, content, events, but, you know, just through osmosis I feel like I eat, breathe, sweat Drive Capital. So happy to answer any questions related to the firm and also answer this question, which is: is Ohio the next Silicon Valley? Absolutely not.
I think that would be ridiculous to say. Silicon Valley, it's almost like saying is Atlanta the next Hollywood because they're making a bunch of films there. Like no, of course not. Hollywood existed and it was a very special place for certain reasons. You can make movies in Atlanta now. You can make movies in a lot of places. Is Montgomery the next Detroit because they can make cars there? Like of course not. Detroit was a very special thing that happened.
Silicon Valley was a very special thing that happened, and there's a reason why there was such a small area where so many great technology businesses were built, and it's because you needed a very specific infrastructure and a set of people, a talent that could manage that infrastructure, and that just was located in a 10 to 20 mile radius, you know, San Francisco, Silicon Valley. Well, 2007, 2008 comes around, the advent of cloud compute. You no longer need to be in Silicon Valley to access that. Now, with an internet connection and a credit card, you can have access to AWS and all the storage, all the compute that you can pay for. What's missing is venture capital, the dollars to invest behind that.
So, no, Ohio will not become the next Silicon Valley, but Ohio has a chance, and I think is becoming, like, the best version of Ohio possible, and it's super exciting. I mean, it's exciting startups like Path Robotics, Root Insurance, which is now a public company and was started in the Drive office. You have Silicon Valley stalwarts coming to Ohio, whether it's Anduril or Intel. And I think that trend is only going to continue because Ohio is a place that is very business-friendly. And I think that's a great strategy for the state. I think Columbus is also doing a great job in making it very business-friendly. So, no, we're not the next Silicon Valley. I don't know if there will be another Silicon Valley, just like there's not another Hollywood.
There aren't necessarily right or wrong answers on Manufacturing Happy Hour, but that was certainly the right answer. I kind of set you up for that because I think I might have heard that in an older video where it was mentioned that Ohio could be the next Silicon Valley. I love the thing that you said where Ohio's just going to be the best version of Ohio that it can be. We don't need to be the next Silicon Valley.
Nima and Andy, the next question is for you, and Andy, maybe you want to lead this one off. We were just asking about Ohio in general. Why has this been a great spot for Path Robotics to start? And then we'll get into your story of your organization.
Yeah. So, I'm just going to go into the story because I think it kind of builds back in. So, I'm born and raised in Ohio. I didn't transplant here. I was born here. I grew up in kind of like the standard Ohio family. I don't know if other people can relate to this, but a family member in my immediate family worked in manufacturing. I mean, everybody here has got to understand that. All of my grandparents worked in manufacturing. All of my great-grandparents worked in manufacturing, and a lot of them were welders.
And so, when I was growing up, my dad specifically, he was deep in manufacturing. He's been in manufacturing about 40, 45 years at this point. He converted our garage into a fabrication shop where we had a mill, where we had a lathe, and it was the main thing that drove my mom crazy. I'm sorry. But that's what I grew up in. And so, at the age of seven is when I started basically building off-road vehicles in our garage. And at the age of 10, my dad decided to quit his job and he wanted to start his own company building off-road vehicles.
And so for four years, we had, I think it was about a 5,000 square foot facility in Warren, Ohio, where we were making off-road vehicles, off-road four-wheel vehicles, off-road two-wheel vehicles, and it was literally just my family and a handful of people. And I saw firsthand how difficult it was to truly manufacture, to truly build a product, and on top of that, how difficult it was to find and retain amazing people to help scale your business. And that kind of stuck with me forever.
And so, as much as I was really into mechanical engineering, when I got to college, it was the first time that I had seen Boston Dynamics. And Boston Dynamics really kind of hit home. I wanted to see what could I do with mechanical engineering, but what could I do to make these mechanical systems intelligent? So, I went off to do an internship at what's called the Institute of Human and Machine Cognition. And so, that entire facility was basically just PhDs from MIT that all worked with Boston Dynamics. And we were working on bipedal robots. We were working on exoskeletons.
And so I decided after that, I was absolutely obsessed. I was so amazed by what you could do to actually control these systems that I went and did my PhD in basically deep reinforcement learning for humanoid robotics. It's funny because Path is not in humanoids, but my entire PhD that I did for six years was about how to get robots to learn how to walk.
And after doing that, and actually while doing that, I was doing that with my brother, and my brother was also doing his PhD at the same time. His was on computational neuroscience. And while we were doing this, it was a really good opportunity to think about what we wanted to do next. And we just had such a rich experience being in Columbus, or not in Columbus, being in Ohio, being in manufacturing. We wanted to utilize our skills to see, could we make a company that really focused on machine intelligence, computer vision for robotics, to just make manufacturing better, to make it easier, to make manufacturers be able to accelerate, because it's one of the oldest, hardest businesses to truly scale.
And so that's when we jumped into it. We talked to about a hundred US manufacturers in Ohio and we found one named Corsa Performance, and they were the first group that was willing to give us a little bit of money, about $300,000. And for that $300,000, we had to deliver to them the first ever fully autonomous welding cells. And when I say fully autonomous, these were really autonomous. You put a part in and it was one button that said go, and we would have to 3D and 2D scan just a generic area. We'd have to recognize what we should weld. We'd have to determine how to weld it. We would take data during and after to say if it was good or bad, and we'd use that data in a reinforcement learning algorithm to continuously improve. And that was basically the start of it.
After that, I was really nervous. I flew out to Silicon Valley. This is actually your joke and I love it. But when I went out to Silicon Valley and I said I'm from Ohio, you could tell that everybody looked at me like, where is Ohio? And the people that did know Ohio, there was definitely the thought, don't you just ride cows and chew tobacco all the time? And so we went out there, two PhDs, basically saying, "Look, we have this tech. We have this customer. Do you want to invest?" And I truly thought nobody was going to give us the time of day. But they did. We found a great first investor that believed in us. They gave us our first seed money.
We came back to Ohio. We met Drive Capital. Nick, who's not here, is our board member from Drive Capital. He was relentless, constantly texting, calling me, wanting to be a part of it, and eventually we did the deal with Drive and we were so happy to have done it. We moved to Columbus, and that's really when we reincorporated, became Path Robotics, and went on this journey of how can we be a machine learning company that makes robots smart to help manufacturing.
Excellent answer. I was going to ask you to describe Path Robotics as if we're having a beer with one another, and I've literally gotten halfway through my beer during that answer. So, good, thorough discussion. I'm probably going to need a refill before this podcast is over.
I'm writing down some topics I'm going to bring up later in the conversation, but since we're on the track of Path Robotics right now, by the way, I should have introduced Andy a little better. He's the CEO and one of the co-founders of Path Robotics. So for context there for everyone that might have been curious. But I've been following Path for a little while and I just feel like I'm getting a different vibe from your organization than maybe when I first learned about you, right? It feels like there's more to it than just the welding robots now, for example. So Andy, maybe you want to answer this. Nima, maybe you want to do it from a more technical standpoint as well. Either way, whichever one of you wants to lead off on this one.
So yeah, I can take it for a second. So
Everybody looks at Path and everybody says we're a welding automation company. That's kind of how we get branded and it's not really completely true. Our first product is welding. But our technology stack is really what we work on every day, which is really what defines us as a company. And our technology stack is much more about how do we use data with machine learning to get systems to learn how to do tasks.
And the first task is welding. It's not the last task. We always think about welding as the heartbeat of the company, right? That's going to be the core thing that we want to be the best in the entire world at, but we want to do every task before and every task after welding in manufacturing to be able to give manufacturers the true labor source that they need to continue to scale.
And so we are looking at task before and task after. And right now the main task that we're looking at outside of welding is dexterous assembly. So when you think about manufacturing and you go to large OEMs or you go to SMBs, there's always a massive need for putting things together, using tools, assembling objects, making the actual physical components go in together. And in a lot of cases for us, even on welding, there's always that task before you're welding. If you're welding something, it means two parts or three parts or X number of parts were put together and then they are physically welded together. So, we are already expanding our capabilities, already expanding our task range using the exact same intelligence stack to move from welding to dexterous assembly.
So, I've got one more question for you, then I'm going to hand it off to Nima, because I think this is something that manufacturing leaders here in this room as well as, you know, across the country and across the world can learn from you. So, you started with autonomous welding and then you said, "Hey, dexterous assembly is the next right application for us to look at." What is your advice to manufacturers to pick the thing they focus on right now but know what the next thing is?
What do you mean? Well, just in terms of like a business strategy, like there are probably a lot of different applications you could have picked beyond dexterous assembly. So, why was that the one that you're like, this is where we go to next?
So, why we did it is like we want to be customer obsessed. We want customers to win at the end of the day. I don't know if anyone knows, but our business model is robots as a service. So, we only get paid if our customers are using our systems every day. That puts a lot of pressure on us to actually deliver something they want. Unlike some robotics companies, that's purely capex. You deliver it once, it works for that use case, then it's on the customer to ensure that it goes to the next use case. Well, we have to continuously provide value every single day, even if they want to switch parts.
And everybody knows in metal manufacturing, there's always change. There's always a change order, there's always a SKU change, there's always change based on market demand. And so, for us being customer obsessed, when we really focus on what's going to make our current customers happier, what's going to give them more value, we did the next task that drove the most value to them, which is the assembly side. It's the one step up from welding that allows them to take that next pressure off on labor to be able to feed the welding side that we have.
Excellent answer. Nima, I've got a question for you because you're the director of AI at Path Robotics. And you know, Andy, when I looked at your LinkedIn profile, it says you're obsessed with revitalizing American manufacturing with a strong focus on like AI and machine learning. So, I'm going to tie these things together. Nima, what is Path working on right now?
Thank you very much for picking the most difficult task in manufacturing. So what is Path working on now? As Andy said, we are a modern AI robotics company. The way that we're thinking about this problem is not that we're doing welding or we are doing assembly. We're thinking about we are automating and we are helping manufacturers to do what they really want to do. And we happen to start with assembly and welding.
To do that you obviously have to decompose those tasks into some abstract task that can be done in various forms. First of all, you need an infrastructure that can handle various tasks and then you have to probably be moving robots around, right? It doesn't matter if it's, you know, painting, grinding, welding, it doesn't matter, right? So, we have to come up with that task that enables robots to move. And then we have the machine learning aspects of this. Your robots need to see what they're looking at, where they're going, right? It doesn't really matter what the task is. Again, this is just one network that learns all different types of applications and then the other part is, okay, now that we know where we have to go, we actually have to perform the task. That task happens to be welding or it happens to be dexterous assembly.
So the work that we are doing right now is focus on building a world model. What is a world model? Well, I have to give you a little bit of context. One of the hardest things in robotics is data. Data is the bottleneck right now. The fact that we're not seeing a GPT moment in robotics is because we don't have enough data or internet-scale data as we had for text and language.
If you don't have enough data, well then how do we create very large models or models that can handle edge cases? And as you know, in manufacturing it is very, very important to handle those edge cases. There are very few common issues or problems that you encounter and there are very uncommon or very infrequent problems that come up in manufacturing, and when you add them up, those edge cases will add up to, I don't know, maybe 20, 30%. And manufacturing has a very high bar for accuracy, for quality, for first pass yield, for cycle time. So in order to solve this long-tail problem, we have to train our models so that they can see those edge cases and they can learn from the mistakes of other robots.
There are two solutions to do. Either you work with simulated data, you build your simulators that can take, you know, a lot of money, a lot of time to set up. And even when you set it up, it turns out that the data that is produced from simulators are not that accurate. There is a sim-to-real gap from what you see, you know, on the floor versus what you actually created in your digital twin. So simulators alone are not going to be good enough.
So what else can we do? World models, we believe, are the answer. What they do is they take real data and they learn to predict the next step and you can control them. So you take real data and you create this neural network that is a physics-based engine that learns the laws of physics and you're able to do different tasks with it. So that's what we're working on. That's where, you know, it's going to be our data engine and that's something that we are really excited about.
So just to recap, I want to make sure I got this right. You're training models for edge cases and you mentioned you kind of need two parts of this. You need simulators and then models that learn the next step. Did I capture that correctly?
So there are two ways to do it. One is simulators, right? Another one is actually getting real data and then training a network to work with that.
And one of your colleagues was mentioning earlier that you are very likely moving faster than the speed of Moore's law right now in terms of like how quickly you're advancing. On Manufacturing Happy Hour, I say this is like a leadership podcast disguised as a manufacturing podcast. So, you know, a typical leadership podcast, you wouldn't get into the bits and bytes, but we like getting into the bits and bytes a little bit on this show. So, you know, what has improved the most just in the past quarter, the past six months, just to give industry and the people here an idea of how quickly things are moving?
Yeah, absolutely. So maybe if you look at a year or two ago, the mindset was you get some data and you train a model that is specific for that task. Okay? So, if you have a new task or a new customer or a new part, you have to keep doing that. And then GPT happened and people changed their mindset in a way that, okay, you have this data, you train it on a large number of different tasks and it's able to generalize.
So that has been a shift in our mindset as well. And what we're doing right now is we are training these models that it doesn't really matter, for let's say the perception side, which customer in which state or which country, which part, which material, which joint type, which temperature, lighting, it doesn't matter anymore. It's able to handle all kinds of situations. Similarly for the assembly, it doesn't matter if you're looking at two flat pieces or you're looking at, you know, different parts, maybe cylinders or something like that, we are able to pick them, we are able to put them together. So this is all because we are moving towards those generalizations with one large network.
Andy, go for it. Yeah, I'll just add on to it. Yeah, the big, big change for us is just quantum of data as we are getting more cells in the field. Every single system is collecting data for us every single day. And so it has an exponential effect basically for us. And so as we're getting more data across different time zones, different lightings, different countries, if you think about some cells are running at 0 degrees Celsius, some of the cells are running at, you know, 40 degrees Celsius. There's a lot of different data that's coming across the entire North America for us right now. And we're taking all that data, you know, basically concatenated all together and then training one large network from all of that data. And so we're really pushing for big neural networks, foundational neural networks that are trained across the board. And for us, as we're scaling aggressively, we're doing about 3x growth every year. 3x growth is a combination of not just revenue, but also in data. And data is the biggest thing. It's the gold mine that we care about the most right now.
3x growth is a combination of not just revenue, but of data. I think that is because I keep hearing that data is the most valuable asset in a lot of my interviews. And I think that's a very tangible way of describing that.
Nils, I promise I'm about to work you back into this interview at some point. I actually said everything I had to say with my first answer. Your first answer was really strong, but no, there's more to this. I do want to ask Nima you first and then Andy, what are you most excited for at Path Robotics right now?
There are several things that I'm really excited about. One, working with people. We have an amazing team, brilliant team. I'm, you know, actually getting emotional goosebumps talking about this. Whenever I go to work, you know, it is so exciting to see everybody all aligned with a mission. The second thing is the work that we do. Large models are great, diffusion models and text generation, video generation is great, but the thing that I'm really excited about is embodied AI and embodied intelligence, and we are working on the forefront of that. We've been talking about this for several years now. We are bringing machine intelligence and AI and we're putting it onto a real, tangible, physical robot. So that really gets me excited. Andy?
Yeah, for me I think the biggest thing is just the fact that we are deploying into the real world. I mean, right now the hype cycle on robotics is through the roof. I mean, the amount of things on the internet right now that are just totally based on hype and not reality is crazy. So, for us and for me, the biggest thing, the most exciting thing for me is the fact that we're building real technology that's grounded in machine learning and it's actually going to the field and impacting customers, and we've gone through the entire cycle.
Anytime you see a demo that hasn't actually gone to a customer site, they are three to 10 years away from making that reality. And I lived through that. That was the reality for Path three to four years ago when we had an incredible demo. But when you actually deliver to a customer, everything breaks. And anyone in this room that's actually delivered a real robot knows this. It's the most painful process. We have been through the journey, or the pit of despair, to get back to the fact that we've gotten through the really tough deployment process of making that work, and now we are there and we are scaling aggressively these real systems into the real world. And to me that's one of the most exciting things. I don't think there's, there's probably like maybe two companies in the world, maybe three, I don't know, that are actually deploying real neural networks trained with reinforcement learning that are actually deployed to customers working every single day.
How do you make sure your customers or people that want to work for you understand that? Because, you know, you've evolved into this forward-looking technology company. And I feel like every company out there is whatever the company name is .AI, right? It's like, we're an AI company, we're an AI company. How do you make sure, you know, people know that you do have these real-world applications? Maybe the right way to ask this is like, why would someone outside of Path care about everything you just said? Why would someone want to work with you, for example?
Well, I think, why would they want to work with us? I think they just come and meet the team. The team is extremely exciting, extremely electric, extremely motivating and I learn from them every day. Why other customers would want to? They just come and see it. We're at the point where it's just come see it. Give us your parts. Come see the technology. There's so many times that when we do a demo, we are signing a contract within 30 days. Our current... So, where's Heather? Oh, Heather left. Anyways, she had a flight. She snuck out the back. Oh, Heather, she had deals to close. Yeah. Good. Good. Yeah, under Heather's leadership as our chief revenue officer, our current deal cycle is 80 days from meeting us to closing and that current deal size is about 5 to 10 million.
Well, you know, I have a follow-up question to this. What can manufacturers learn from you about reducing their deal cycle to closer to 80 days? Like, what's the macro lesson here someone else can take away from this? She's on a plane, so I don't know. Question for the chief revenue officer.
Yeah, hire an amazing CRO. Yeah, I mean, so she does it better than everybody. She aligns with companies. So, what's their three major strategic initiatives for the next three to five years and how do we play a part in it? And what she does better than anything is that she disqualifies companies faster than anybody else I know, so that we have hyperfocus on companies that need us, that have a need, they have a want and that we can deliver them something of value. So she always focuses on the top three initiatives, understands their true learnings.
So like one of the best things ever is when she walked us through, we had a prospective customer. She did all the diligence on them. What's their biggest customer? What's their biggest customer's growth rate? What's the biggest need right now for them? And she was able to walk them through all of their own strategic initiatives. And based on that, she was able to build a case of how we can help them accelerate into them. And she's done that meticulously. It's not about the tech, it's about the value that we drive to them. And it's both on revenue, on growth, on profitability. And she's done an amazing job at being able to paint that picture. And I have not, just to be clear, it was about five times worse than that when I was leading sales. So
That's my best shot at doing what she has done.
So I took three big takeaways from that answer. I like how you tried to pass the buck and not answer it because she had left, but you know, you mentioned, hey, she gets involved in their strategic initiatives, right? Three strategic initiatives. Understand how your company fits in. The other thing that sticks out to me as an ex sales guy is that you said to disqualify quickly. I don't think enough people are willing to do that. They'll chase someone that they're like, "Ah, this probably isn't a good fit, but maybe we'll win." Terrible idea. And then gosh, what was the last part? Oh, you said I wasn't as good at it. Heather's way better at it. Right. More companies need leaders with that type of self-awareness. So, love that answer.
We've got about 10 minutes left here before we open it up to a little Q&A. So, let's bring it back to Columbus for a second. Let's bring it back to Nils, who has just been sitting there drinking his beer waiting for his next question so far. Second beer. That's how long we've been taking. No, this is good. What can other regions learn from the success you've had in Columbus? And Nima and Andy, feel free to jump in here as well. But Nils, I think you're the right person to start off this answer.
Yeah, I don't know how actionable this is at the individual level. And I don't know, maybe most folks, and I don't know how we're defining region, as Columbus or as Ohio or as the Midwest or whatever, but you know, one thing that we've learned as a firm is that when you start setting out to create an innovation economy, which has existed in, you know, say Silicon Valley for decades, you don't just need startups, you need a whole kind of culture around that to support those startups. So, you need law firms that understand, hey, you know, we're going to sign a term sheet here and it doesn't need to be 100 pages and it doesn't need to take six months. That's not how this works. You need landlords who understand that we need one to two-year leases. Like, this startup is not going to sign a 5 or 10 year lease. There's an understanding of risk, of what startup means.
I think right now in Columbus, if you were to tell your spouse, I'm going to go work at a venture-backed startup, they'd be like, "Right on. Let's go. Which one is it?" I think 10 years ago, 15 years ago, they would have said, "But honey, you know, we've got kids that are going to be in college in 15 years. How could we possibly risk your career like that?" So, there's a cultural shift that needs to happen that just takes time and kind of open-mindedness about that. But I don't know how actionable that is. It's just a matter of time, but that's what Drive has been going through on the Columbus level. But, you know, I think there's a cultural shift that's happening across America. So, it's not like Columbus is 10 years ahead of Dayton or something like that. I do think a lot of these markets have made a lot of progress alongside Columbus in the past 10, 20 years, and that's only going to continue.
Well, a couple things I liked about that answer were just some of the very tactical, pragmatic aspects, right? Law firms that get it, landlords that get it. And it's also, you know, we talked about people earlier in this conversation. We talk about people in pretty much every interview we do on this show. Now, you have people in the region that understand going to a venture-backed startup isn't, you know, a massive risk in your mid-40s when you have a family and kids. I thought those were all excellent points.
Yeah, we like to talk about what's scary versus dangerous. You know, was it scary to join Sears in 1995? No. Was it dangerous for your career? Actually, yeah. You're screwed. Sears is no longer in business. Is it scary to join Amazon in 2010? Yeah. That's a fast-moving beast of a startup that's going to scare the bejesus out of you if you've been working at Nationwide or wherever before that. Was it dangerous? No. Cuz you had an incredible experience and you learned a lot of valuable skills. Same for a Path Robotics right now. So something might be scary that's a startup, but it's not dangerous. It's going to give you all of the tools and skills that you need for where the economy is going. And I think understanding the distinction between what might be scary versus actually dangerous is important.
So last question, and I want each of you to answer this one. Andy, we'll start with you and then we'll go down the line. One of the things that stuck out very early in this conversation, Nils, when you mentioned that you were out there and people are — no, I'm sorry, Andy, it was you that mentioned that you were out in California and people are like, where is Ohio? Right? Like the attitude around, hey, the Midwest is a great place to do a startup, to build a company, was just not there. Is the attitude around the Midwest changing on the coasts? Are people looking at this region differently than they did 10 or even five or three years ago? Andy, maybe you start this off and then we'll literally just go straight down the line.
Feel like you should start this off. Okay. All right. We'll go back to... I'm coming in hot and saying absolutely not. It's not changing. I mean, if you live in San Francisco, nobody is like, "How's Illinois doing?" Like, they don't give a... And you know what? That's fine. They're playing their game. They're playing their market. It's fine. But we don't need to worry about how they perceive us. I think we just need to play our own game. And I think we should all be very bullish on the futures of Columbus and Ohio and the fact that you can build a world-class technology company literally anywhere.
And why don't we need to worry about them anymore? What would be your answer to that? I've got some of my own answers to that, but I'd love to hear your take.
I mean, just growth mindset. There's so much innovation that is yet to be done in so many industries. Like, yeah, we've figured out social networking, but there's so much that needs modernized across industries and across verticals. So, you know, I don't think we need to be concerned about what Silicon Valley is doing. They're going to do their thing. We need to worry about hospitals and manufacturing and transportation. There's so much work to be done here and there's so much upside and so much opportunity that, you know, I think we can just play our own game and do very, very well.
I like it. By the way, to everyone that listens to this after the fact from California, New York, we love you guys as well, but we're just trying to learn some things here as well. So, kudos to all of you. Andy, Nima, anything you want to add to that?
Yeah, I mean, I probably kind of echo it a little bit. We still have that same issue, like, you know, raising capital from the coast. Everyone still asks, "When are you going to move to Silicon Valley? When are you going to move to the East Coast?" And everyone keeps coming back with the main reason is that the talent is here. The talent is here. It's like you think everybody in Ohio is just an idiot. That's crazy. There's a billion... no, there are what, 22 million people in Ohio. Something like that, right?
And I think most importantly, you have academic centers and research centers like Case Western, where you and your brother studied, and Ohio State and Carnegie Mellon and on and on and on, that are producing the type of talent that a Path Robotics is looking for. There is amazing talent here. I think you have one of the best stats and I can't remember it, but the amount of software engineers that are graduating in the Midwest is more than that on the coast. Something crazy like that.
There's amazing talent here. There's amazing people here. People that are grounded in reality. People that have amazing values. I would never build a company anywhere else. I've been to the coast. I go to the coast all the time. I have a bunch of investors out there. It is a different culture. It is a culture of, you know, frankly, kind of assassin-like engineers that move from company to company to company. We've all seen on LinkedIn, ex-Twitter, ex-Meta, ex-Google. Okay, great. You want people that are going to be bought in, that are going to grind, that are going to want to win. And what I've seen from Columbus has been an amazing, amazing talent pool. There is no IQ test that says Columbus or Ohio is below the coast. Everybody is brilliant that we have worked with, and we've seen that everybody's incredibly teachable and they are willing to grind and they want to win. So to me this is an amazing place to actually build a company and build something that's going to last a long time.
And then further, I'm one of the only AI robotics companies in the Midwest. There's a million that are happening on the coast, and all of their customers are going to be in the Midwest. We decided to stay really close to our customer base where we're going to actually be deploying equipment, so we can be there quickly, so we can be in the weeds with them, so we can learn from them. And so again, to me, Columbus and Ohio has been an amazing place and I wouldn't change it.
I thought about asking another question, but that was too good of a way to really end the main part of the podcast. So, hey, let's give a round of applause to Nils, Nima, Andy. Thank you so much.
Now, we'll probably do some swapping of seats up here in a little bit, but we would love to open it up to Q&A. Does anyone have a question? If you do, come on up here. Ask it in the mic so that way we can capture it here on audio as well. All right, jump on up. Jump on up.
Full disclosure, I've been a supplier to Path and have loved following their story the last few years. But my question for you is, what other shining stars in addition to Path have you been working with in Columbus, in central Ohio, perhaps in Ohio in general? What other shining stars, maybe not shining as brightly as Path yet? I mean, the last couple months has just been stellar for you guys. That's been fantastic. Stellar for you, too, if you're the supplier. Yeah. Well, but what's coming up in Ohio? Either specific companies that you might be able to name drop or industries that are on the rise in the manufacturing and the robotics industries that all of us, whatever industry we're in, should be paying attention to.
I almost feel like, you know, from a what manufacturing startups are doing well, I almost wonder if Andy has a better finger on the pulse than I do. We're constantly looking at companies that are automating, whether it's manufacturing or back-office healthcare tasks, and you see a lot of them coming out of Columbus, and I think the reason why is because you get a critical mass of people working on a relatively similar problem and they start to feed off of each other. Same thing that happened in Silicon Valley. Which company specifically? I mean, I do see a READY Robotics sweatshirt in the front row. They were an investment from a past life that is no longer, but you know, there's been a number of investments. I'm trying to think specifically which I can name right now in Columbus that's specific to manufacturing. Help me out, Andy. Are there any?
Not that I know of. There is another AI robotics company, I mean, there's Skild that's out of Pittsburgh, so in the Midwest group, only a couple hours away, but in Columbus specifically I don't have any others.
You know, Pittsburgh sometimes gets offended when you loop them into the Midwest, right? They're like, "No, we're our own thing." Okay. Sorry. Sorry, Pittsburgh. Okay. Thank you.
Any other questions? Surely we've got another out there. I do have Q&A as well. Tim, do you have questions? Oh, come on. You want me to ask? Yeah, jump on up. Jump on up. You requested the Q&A section. I want to make sure we get you in here.
I want to hear from Nima, like, in the five, six years that you've been involved with Path, what have you seen, not just in Path's technology advance, but in the last 12 to 24 months of AI in general? You've seen the buzzwords, but what's the real advancement that we're applying to Path?
So like I said, I mentioned this before, the main thing that we are again learning is going with larger data sets, larger models, and generalizing to more and more tasks instead of being very specific and very specialized to a given task at hand. So this is something that we've been working on. This is something that is on the forefront of our minds, and we are thinking about it constantly because, you know, we're not going to stop at only welding and assembly. We're going to think about the processes that come before and after. And we don't want to reinvent the wheel every single time we onboard a new product, a new task, a new customer. We want to make it super easy, frictionless, and seamless when we are thinking about those things. So definitely that's something that we've been thinking about and been doing.
Leslie, welcome back. Thanks. So actually, as you guys have talked about continuing to, I guess, swim upstream with the manufacturing process, you've gone from welding to the assembly. There's always this talk about the fact that nobody is losing their job, they're just continuing to upskill. And so I really am fascinated by the fact that the AI keeps taking everything further with what these robots can do. How do you keep seeing humans and the robots continuing to interact and people continuing to upskill as that happens?
I guess... go ahead. I just want to say, you know, with any technology, I think it is fair to say, no, some jobs actually going to be eliminated. Let's go back and think about the horse-drawn carriage. Before there was a car, there was a horse that you have to feed, and there was a person that had to take care of the horse, and that person doesn't do that anymore when cars were invented. And you look at, I don't know, a barber shop 200 years ago, when you would go to your barber shop and get a haircut, and at the same time you would, I don't know, pull out a tooth or something, and that person was your... as well. And we think about it right now and it's crazy. You want to go to your barber and say, "Hey, can you, I don't know, look at my teeth as well."
And I think it's going to be the same thing with AI. This is a new technology. There's a famous thing from Steve Jobs that says whenever there is a new technology coming in, people keep applying the same old problems to that new technology, and then over time they realize or they learn how to actually use that technology. So I would say I think there's going to be some displacements, but as a result of that there's going to be some other new jobs being created. Humans are not really good at thinking about exponential growth, and as these new things come in, we're going to learn and we're going to adapt and we're going to create new things and new possibilities for people.
Nima, you talked about a GPT moment earlier. I feel like since that GPT moment, the narrative has completely shifted around jobs being automated and specifically what jobs are being automated, where I feel like, you know, in the 2000s, 2010s, it was all about more manual labor type jobs being automated away. You know, if you're making 20, 30 bucks an hour and you can automate that job away, you can save a company 20, 30 bucks an hour per person. If you can automate a lawyer at 300, 400 an hour, it's a completely different calculus and completely different economics. And I feel like that GPT moment when we saw that you could just upload a PDF of a contract
and say, "How is this unfair to me?" And it can give you a pretty good lawyeresque response. It's like, "Oh, that's the investment." It's not figuring out how to remove minimum wage workers or even the $20 an hour folks. It's like we can automate lawyers. It's a completely different calculus now. And I think we're just at the beginning of seeing who actually is about to get automated out of the economy.
Yeah, I would say from us being in manufacturing, actually those type of jobs are very, very hard to automate as opposed to lawyers or the medical field. I feel like the only thing that the medical field has got going on for them is the fact that there are so many regulations, right? If that wasn't there, there's so many things that could happen. Again, think about the barber and the dentist, right? And think about in the future how good the healthcare or the medical field could be because of that.
It's like the less in the real world your job is, the more that your job is on a screen, the more screwed you're about to be.
That's actually a really good point. And Andy, you go first, because I know, in true founder fashion, he has to jump off and take a 7 pm Eastern call. So you go first.
As a guy whose barber is still his dentist, let me start there. I think I'll go in a different direction. The gap of jobs needed is still super high and it's constantly growing. So, at least in our industry, we aren't feeling that at all. Every company that we are talking to, nobody has ever been laid off. Nobody's ever been replaced with the robot. We have always been an addition. And so, we believe that's going to be the truth for a long time. And we also believe, just back to Nima's point, there's going to be a ton of opportunity. How do you work with these systems to be even more efficient, to grow even faster as the economy grows, as the population grows? I don't see this as a tomorrow it's robots, goodbye humans. I see this as a constant growing, and we're still focusing on how do we help fill the gap to accelerate manufacturers. But yes, with that, I do have to go.
No, Andy, I'm glad you went first because you said 80% of what I was going to say.
So, hey, this is actually the end of this portion of the podcast. I have two more guests I want to bring up here. But hey, another round of applause for Andy, Nima, and Nils. Thank you so much.
So, I'm going to invite Josh from A3 to the stage. I'm also going to switch microphones as well, just so I can be back in the center.
But no, I've heard the point made before that in the US there's maybe an inordinate amount of people going into roles like legal, for example, like we were just talking about, and not enough people going into manufacturing. And I do believe the stat is still something like there are 2 million open manufacturing jobs right now, or jobs that will be open with everyone retiring right now. So, to Andy's point, no one is getting replaced with all of these. There's just some, I see it as repurposing of jobs, and we've seen that for centuries as a civilization.
But anyway, now that Josh is up here, another group we want to give a shout out to is A3, the Association for Advancing Automation. And we've had a lot of member organizations from A3 showing up throughout these events on this tour. And one of the reasons is a lot of these dates have been very close to Detroit, Michigan, which is where the Automate Show is taking place on May 12th through 15th. Mark your calendars. If you are wondering to yourself, is Chris going to throw a banger of a manufacturing happy hour party there? The answer is yes, on Tuesday. But hey Josh, welcome to the stage. What would you like to share about A3 and the Automate Show that we've got coming up?
Awesome. And you trust me with a mic. That's dangerous.
Of course.
All right. Oh yeah. So actually I'm brand new with A3, so I'm probably the worst person to talk to.
No, I can help you.
He's going to guide me through this. But no, actually I just started in November. I come from the education space, so my passion is actually K-12, former teacher. And that's one of the big initiatives we're looking at. But A3 is, as he said, the Association for Advancing Automation. We are automation. We are manufacturing. We're just a trade association of members. We're here to support that industry, to champion it, to drive it forward. Like I said, everything from education to safety to regulation. If some of you are members, that's awesome. Come hit me up. If you aren't, I'm not in charge of membership, so I'm not here for a sales pitch. I'm just passionate about this. And like I said, we want to see it be something that grows and is fostered.
Like Chris said, we host Automate. If you've never been to it, I haven't either, but I'm super pumped to go this year. It is the largest automation conference. Just huge. It's up in Detroit this year, so it's close to home. Again, my passion and my background with the education side is something that we're growing. So, we're going to have a big education presence, including our education pavilion, educators day. We're going to have, I think the last number I heard is almost 400 and some students and growing right now. They're going to be coming. So, why does that matter to some of you? It's the next workforce, right? That's what we're passionate about, trying to get young students excited about automation, excited about manufacturing. We're working with programs like VEX and FIRST and getting kids in robotics and trying to get them to come connect with you. I'll pause there because I've already said a lot.
No, and I've been to that show 2022, '23, '24, so the past three years. And I was literally just on a phone call today with someone that's just getting involved in the manufacturing industry right now. And he's like, can you send me a list of all the conferences you're going to go to? I'm like, you know what? If you don't have a huge budget right now, just pick one. Go to Automate. First of all, it's free. A lot of the biggest voices in industry are going to be there, people driving it forward as well. So, hey, it's just two months away from the date that we're recording this. So, I know we've had A3 members out here this whole tour. I have A3 members in the audience right now, so I know we'll see some of you there. But Josh, thanks so much for jumping up, giving us the quick spiel, and go to Automate.
The last person we're going to bring up here tonight is Adrisu from Makerpace Central here in Columbus.
Josh, thanks for quick last pitch.
Hit me up if you need another drink when your tickets run out.
So, yes. Okay. Josh has extra drink tickets. That's very important. So, all right, Adrisu, if you want to jump up here, and I'll give some context. One, I don't know if you read the fine print when you signed up for this event, but one of the things we do on manufacturing happy hour tours is we like to support the folks that are getting the next generation, getting underrepresented populations, getting veterans into the manufacturing industry. And you and I just met, gosh, probably just like a month ago. You're based here in Columbus, and the things Makerpace Central is doing is key to getting the next generation involved in our space. So, a portion of proceeds from this tour are getting donated to Makerpace Central. But if you could tell us about the organization that you started, we'd love to learn a little bit more as we wrap up.
Yeah, I think I may follow Andy's lead here. But yeah, so Makerpace Central, what we do, we are a nonprofit here in Columbus. We focus on serving underserved, underrepresented youth and getting them interested in STEM careers. So, how do we do that? We do that through hands-on workshops focused on a cohort, project-based learning approach. We work with schools, libraries, community centers. We work in informal learning environments for the most part.
What really started this all off is the fact that I'm an engineer by my training, right? So, actually Andy was talking about Case Western. I actually went to Cleveland State, right down the street, like 10, 15 minutes. So got my bachelor's, my master's there in mechanical engineering. Flew out to California to work on rocket engines and built those, tested those in the deserts of California. Very hot and dry out there. Perfect environment for all that. But what I really had the privilege of doing was going out to the STEM youth engagement events, right? Talking to these kids about rockets and 3D printing and how do we do it all, right? But then they would hear rockets and then they would ask me about black holes, which is fine. Not exactly the same thing, but what it showed me was that these kids were really hungry to do this kind of stuff, right?
So, moved back to Ohio because I had a baby and I can't do that in California. It's super expensive. I don't know if you guys have been out there. So, moved back here and really had it in my heart to get this organization started where we work with kids and show them that there are really interesting career pathways out there, and STEM is for everyone. Not just for people who go to really wealthy school districts, but anyone who is interested and curious about this has the opportunity to learn this and to advance their lives and change their lives in a really strong way.
Excellent. Well, hey, make sure you meet Adrisu before you get out of here tonight. Learn more about Makerpace Central. Last thing I want to say is, well, my beer's empty, so I can't really cheers you right now, but thank you all for sticking around tonight, hanging out. But we've got plenty more parties still to go. So stay innovative, stay thirsty. Thanks for being a part of Manufacturing Happy Hour here in Columbus, Ohio.
Article published