SigmaCorns: A New Robot Generation

FIRST Tech Challenge (FTC) is a global robotics program for high school students that teaches skills that prepare them to be the tech leaders of the future. Over 2,500,000 students have participated in over 100 countries. Meet the SigmaCorns, from NC’s School of Science and Mathematics, a top ten ranked team in 2022.

There’s a lot of AI products out there now, but most of them are just thin UIs over ChatGPT. So how can we build AI tools that are truly revolutionary?

Jacob and Ayush share their experiences and thoughts on this problem as they describe their journey in the world of AI, from being students at the North Carolina School of Science and Mathematics to becoming researchers at a Venture backed AI startup.

Slides

Find out more about BoS

Get details about our next conference, subscribe to our newsletter, and watch more of the great BoS Talks you hear so much about.

Transcript

Ayush Paul 

Okay, we’re all good. So hey everybody. My name is Ayush Paul.

Jacob Van Meter 

I’m Jacob Van Meter.

Ayush Paul 

We’re so excited to be here at the Business of Software conference in front of some amazing minds in the SaaS space and amazing founders as well. We’re really glad to see that BoS has decided to relocate to our home state of North Carolina, and it’s a center of entrepreneurial growth and innovation. Yep.

Jacob Van Meter 

Yeah. So I’d like to say I’m very grateful for being here too. It’s pretty amazing to me. I think it’s just really cool to me, how far you can get if you’re just interested and have a commitment to curiosity and exploration. So yeah.

Ayush Paul 

So let’s start our journey. Where most journeys start the beginning. I would say our journey began when we went to the North Carolina School of Science and Mathematics. For those of you not aware, NCSSM is a public residential high school that takes the best and brightest from all across North Carolina and assembles them into two campuses, providing them a world class education for absolutely free. This place has been a huge a huge inspiration for a lot of our growth in life, and it really instilled in us the mindset that we can help change the world using the power of technology. One thing that introduced it introduced to us is the famous Paul Graham quote, “Live in the future and build what’s missing.”

Jacob Van Meter 

So the first instance of this, where Ayush and I actually originally met. A met was in our school’s first robotics team, which is a competition robotics program, and in that, we were able to meet a community of like minded people interested in technology and collaboration.

Ayush Paul 

Yeah, you might have noticed them all around this conference today. They’re right here in the front row with us today, right? They’re robots.

If you’ve been lucky, you might be, might have been able to get a demo with the robot that’s right outside this auditorium. And if you’re really lucky, you might even get to drive it. We’ll see.

Jacob Van Meter 

Yeah, and so we’re very thankful for our bikes experience. We learned a lot of technical skills from programming and CAD and design as well, very important, almost more important, probably softer skills like leadership, collaboration and problem solving with people.

Ayush Paul 

However, our first huge turning point in our journey happened when I was browsing on developer news forums, developer social medias, and I was hearing occurrences of this GPT three, right? I remember back to when I was experimenting with GPT two, and I found it to be, like, an interesting product, but like, trivial in nature, it couldn’t really do much. But these screenshots I was seeing of GPT three, these were absolutely mind blowing, like it was able to write text nearly exactly like a human. I knew I had to get right on this, so I jumped straight over to open a eyes website, and we I put in my email on the wait list, hoping to get a response. It took months, but eventually I received a letter in my inbox saying, you have, you have early access to the open AI beta.

And as soon as I still got that email, I hopped on my laptop, I opened the website, and I started playing around. I knew there was huge potential for this technology as soon as I started interacting with it, and I was just blown away. I’ll admit, at the first I used it for some trivial things. I used it for a bit of help on my homework, if you know what I mean. But yeah, there were huge potentials, and one person who saw those potentials right alongside of us was a mentor on our robotics team. I’m sure many of you here are familiar with him, Carl Ryden.

He inspired us to use this new API to build a product for ourselves and apply it to the field of competition robotics.

Jacob Van Meter 

Yeah, so when we first had access to this technology, Carl, Ayush and I were sitting around trying to figure out how to do it, and one problem that we all recognized was that a big problem in competition robotics is the lack of mentors for a lot of teams, it’s just hard to find people who are able to fulfill that position. And so we were thinking about how we could use AI and GPT to reduce some of that burden, and the way we did that is by building an AI based, always online robotics mentor that we call Foster.

Ayush Paul 

The FTC optimized sentiment and text engagement robot.

Jacob Van Meter 

Yeah and we worked on that a lot, and had, you know, it was pretty amazing. We could just put in, like, all the engineering documents and competition rules and all that, and it could give really great responses. But as we worked more on it, Aisha and I started to recognize the like incredible the more broad applications of it, particularly within the field of Ed Tech. And so that’s where we focus next. And so we next worked on computer science education, mainly because computer science is increasingly becoming a curriculum and requirement around the nation, and I know at least here in North Carolina, that’s a requirement that’s going to be in the coming years here. And so, and it’s also an issue where there’s teachers are stretched a little bit thin, just because there’s not as many of them as in other subjects. And so we were, we were thinking about, how can we use AI to build a product that allows every student learning computer science to get the individualized attention they need. And that’s how we then we used AI and GPT to build that product.

Ayush Paul 

Yeah, we founded learned, going back to that Paul Graham quote, live in the future and build what’s missing. We decided to put ourselves in a future where CS was a required course in every single stage, let alone North Carolina, schools would be spread incredibly thin trying to find the teachers and the resources they need to help teach this new batch of students. So what was missing was a tool for our existing teachers to spread their reach and spread their focus and specify in computer science, that’s that’s where Jacob was heading with LRNT.

LRNT

Ayush Paul 

We realized that we could use AI to give teachers the tools they need for individualized approach with still with maintaining the human connection that’s essential for the human connection that’s essential for like teaching. One philosophy you really took to mind was build Iron Man suits, not Terminators, rather than building a Terminator, which would be, in this case, an entire like autonomous CS platform, like Alex, where all the assignments are done. Through this one platform, we built a suite of tools so existing teachers could extend their reach, rather than replacing the teachers who are already trained in pedagogy.

Jacob Van Meter 

Yeah. And so through this, we had a lot of great success. We had a lot of great success. We had interest from a number of educators, include including David J Milan from Harvard, CS50 some faculty from Berkeley’s data science department. And we also brought it back to our roots, and we’re actually running a pilot program later this semester of our product.

Ayush Paul 

Yeah, we’re running a pilot program back at our roots in North Carolina, School of Science and Mathematics, where we’re going to be helping teach data science to students at both of the NC SSM campuses.

Luck Raiser

Jacob Van Meter 

Yeah. So another core philosophy that we’ve kept in mind along this way is how to get a lucky and so it’s called the luck raiser. And so essentially, we’re trying to comes down to two things. So one, increasing your luck surface area, and two, trying to get rid of anti like things. Now that’s pretty abstract, so let me, like, break it down for you, for example, for us, what that means is, rather than us just sitting together and building our platform or going out and engaging with people you know, going to conferences like this and stuff, to increase the chance of getting lucky, because you’re probably not going to get lucky if you’re just sitting building your product at home or something.

Ayush Paul 

Yeah, so we networked with many people. We researched competitors. We basically maximized the interactions we had with others in chance that eventually we would find something that puts us in good luck. And soon enough, our net was cast, and we finally picked something up. I got a notification on my phone. I checked I see this discord message, let’s see, yeah.

Ai education start a startup that works like on a track, very similar to ours, was looking for some part time help in working on their AI Socratic tutor – Bloom. We decided this would be an amazing opportunity to, like, build our skills and possibly use our knowledge in working with this to work on learnt and other projects we’ve been working on. We decided to do a shot in the dark and reach out to them. As soon as we contacted them, we knew we were instant click. They were a small team of three, but they had deep expertise in the field of education and working with AI, and they their goal was to change use AI to dramatically change the way that us humans learn.

Jacob Van Meter 

Yeah. So we started working with them and we helped develop this AI Socratic tutor, and it was immensely successful. We, like boosted their users tenfold. We added a lot of new features, and we have a lot of active users to it, and it’s really cool. But as we worked with them more, they pivoted a little bit from just education, and we had another instance of living in the future and building in the present. And for us, this was making our building AI agents that are like truly personalized, like AI agents that can get to know you.

Ayush Paul 

Yeah, the problem that we’re in essence solving is the principal agent problem. It’s a problem that derives when an agent acting on your behalf has an interest that’s misaligned from your own interest. And thinking in terms of this conference, that sounds an awful lot like the mach two problem that we were facing earlier today. In essence, what our startup was working on was aligning the goals of the AI’s AI agents that anybody can create using personalized information with your goals. And as soon as everything aligns, we can start accelerating Ed Mach 2. We want you to take the lessons that we use to guide us in this journey, in bootstrapping our career. I guess, in your day to day life, every single one of you here lives in the future in some way, shape or form, but so you have to know, like, what’s missing around you. What can your fellow founders do to help fill in these gaps and build the future together? We want to open a discussion, rather than the traditional Q and A format where you if, if you are interested in asking us questions, feel free to go ahead, but we want to ask you, what do you think is missing, and how can your founders help you achieve these goals? Yeah, thank you.

Q&A

Mark Littlewood 

Yeah, you gonna ask some questions?

Ayush Paul 

Yeah.

Mark Littlewood 

Why not? I’d just like to point out that if we add your ages together, you probably bring down the average age of the speakers quite just saying, Who do you want ask?

Ayush Paul 

Who is somebody who is really living in the future here?

Mark Littlewood 

Oh, Olive, Joseph, you’ve got mics. I’m going to pick on people unless they volunteer. Who’s a volunteer? Well done. Richard olive got a mic to Richard there. He’s always got something to say for himself. Up there.

Audience Member 

That’s a tough question to answer, isn’t it? I’ll tell you one thing someone said to me, which I really liked. I went to a conference, small little event, where someone spoke about the way he ran his business, and he said, in his company, he has two CEOs, a CEO of today and a CEO of the future, and I find that really inspirational. It changed a lot of the way I run my business. My chief operating officer is responsible for keeping the train on the rails, making sure the business runs, which liberates me to really think about where we’re going in the future and what that might mean. Now, the wonderful thing about the future is I can never be proven wrong, because, well, yeah, it’s all to evolve. I did know it was going to turn out that way, really. But I think that if you do run a business, especially in our sort of space where things are moving at such tremendous velocity that you know trying to work out where you where you need your organization to be is more important than ever. We can’t just keep on relying on our momentum delivering success. So I don’t know whether that’s an answer to the question, but it’s just the way I think about my role as a CEO and a founder is absolutely to think about. You know, this is this new these new opportunities that inevitably will come our way, and these new risks and how we want to reinvent ourselves?

Ayush Paul 

Yeah, thank you. That was, like, really insightful one analogy that I could drive from that. So imagine you’re like, on a constantly moving train, right? Every single startup here is a constantly moving train, but eventually the tracks are going to run out. We have to plan ahead and place those tracks as we go along to make sure there’s room to grow. Anybody else?

Mark Littlewood 

Before I pick on someone? Anna.

Ayush Paul 

Okay. To reiterate, the question was like, What is your company doing? How are you living in the future, and what like around you do you see is missing in a day to day life?

Mark Littlewood 

Emily, sorry. Olive. On here in the dark. How are you living in the future? She’s an investor, so.

Audience Member 

I suppose always challenging the team to think bigger, and not being afraid to bring those ideas to the table. There’s no crazy ideas. They’re just ideas, and so we can talk about them, so that we can find that future to build those tracks.

Jacob Van Meter 

You have any thoughts? Yeah, no, awesome. Yeah. I think that’s a big part of this is just, you know, thinking big, because especially of a lot of this new technology, there’s just so many things like that we just haven’t even thought of yet.

Ayush Paul 

Yeah, and we’d also be open to a traditional Q and A, if anybody has any questions for us, feel free to raise your hands and can get you an answer.

Mark Littlewood 

Let’s go with one here, first one to get to it

Audience Member 

While they’re moving. I’ll just let you know that I really appreciate the approach that you guys are taking to using AI as an enabler of existing expertise, as opposed to trying to make the moonshot, replace a person with a machine. I think we have a lot of potential for huge growth in that space. One of the things my company is doing is basically trying to automate your whole process of creating digital content, like slide decks, if you happen to be maybe an engineering leader and you’re not so great at design. The machine fills in the gaps there. And I really like that approach.

Jacob Van Meter 

Yeah, no, awesome. Thanks for that. It’s a big focus for us is just identifying what things AI is good at and what things humans are good at, and seeing how we can best combine those. x

Ayush Paul 

Yeah, to go off that the company we work at, plastic labs, was actually part of beta works camp. And the theme for this year’s inaugural beta camp. Beta works camp is augment the camp focused on technologies that were meant to augment human potential, rather than replace them.

Audience Member 

Great talk. I really appreciate yall. I have a question about putting guards. Can y’all hear me? Yep, I have a question about putting guard rails on the AI. Because I remember when I was a high school student, if I could ask GPT for the answer, I definitely would have jeal put any guard rails on the AI to prevent students from directly asking GPT for the answer and only doing it as sort of a coach model, as opposed to and question answer model

Jacob Van Meter 

Yeah. So I mean to say, like, for example, with LRNT the product we made, that we How it works is the AI always, like, always knows the because you’re, like, working with a specific question, it always knows the answer. But it’s like prompted to not just tell you the answer, right? It’s the idea of, I guess, to get into the architecture of how it works a little bit it. What it tries to do is identify, like, What misunderstanding the student is having. And then, how do we fix that misunderstanding, rather than, how do we just fix your code to make you pass the problem? How do we identify what misunderstanding led you to make this mistake, and how can we rectify that?

Ayush Paul 

And to fall back to the stuff we do for plastic labs. Plastic labs approach to bloom our Socratic AI model is using theory of mind and Socratic and aristotleian methods to address the problem, and that entails like helping the student when they need help, but that usually entails them, like the AI, prompting the student to go where they need to go, rather than giving them the answer directly. Yeah.

Mark Littlewood 

Very quick one, mark

Audience Member 

I was going to ask so, so when you’ve been a CEO for a couple of years, you get used to, like, there’s always a revolution. So I started the .com bust, and we’ve had the financial crisis, and we’ve had VR and now a house the next AI is the next big thing. And I also think AI is really interesting, because the the overlap with the COVID pandemic, which has essentially thrown all workplace practices up in the air, I think is very interesting. I think when you’re in tech, the tech is actually the easy bit. The difficult bit in technology is the people. So I’m curious to know, from the angle of actually how it interacts with people. What are your thoughts on how we make this sort of a useful tech, rather than just a sort of another busted VR solution? What do you think its real benefits are to people?

Jacob Van Meter 

So I think, like, the really important thing is getting it to work. Because there’s a lot of like, AI tools that are in this space where they’re interesting novelties, but they’re not really like finished products. And I think in all the AI tools I’ve seen and what we’re working on, the important thing is just really making sure you have put it into, like, a polished finished product that has, like, demonstrable gains, rather than just being like cool, because there’s a lot of AI tools that are cool, but there’s a difference between being cool and being a genuinely helpful tool.

Ayush Paul 

Yeah, to iterate on that. So obviously there’s like, a bunch of fun chat bots out there that emulate characters from movies and TV shows those, I feel like those fall along the trends of, like, something that appeals to a huge target audience, but falls flat because it doesn’t have many practical uses. But we can use AI for practical uses in automation day to day. It’s gonna save. It is saving. I’m not gonna say it’s gonna save because, yeah, every day it’s being used around us to save like, hours and hours of human time, like, for example, this presentation, a lot of it was pre drafted from a bunch of bullet points put into chat. GPT back out.

Mark Littlewood 

Fraud, nobody told me. Yeah. Well, I’m going to decide who’s got the best AI at the end. Correct? Laura, oh, OK, go on last one.

Audience Member 

So I really appreciate what you guys are doing and the way you’re doing it, you ask about living the future. And AI is a big part of the way we live in the future. Now, what my perception after being an engineer for decades now, is, is that the trend is relentlessly towards the commoditization of what used to be an engineer stock and trade, which is getting data from books and and vending it out to people who don’t have the you know, there’s too much friction for people to get access to that now. And the automation of routine calculations, the automation of research, of existing data, is all flat now. It’s completely flat. So the result of that is that if you want to continue to add value as an engineer or as a scientist of any kind, you have to go to the unanswered and the unaddressed and the open questions, and it forces all of us to be extremely productive in terms of the frontiers of human knowledge, because you’re automating what is already known, and you’re automating the transmission of that data, or that information, that way of thinking, to new generations. This is very exciting moment, and I don’t think it’s a flash in the pan. I don’t think it’s a, you know, it’s a little an anecdote here, but I do think this is a major transformation. And by the way, you guys make phenomenal robots. I’m an FTC coach too. You know we tremble when we go against you. So.

Ayush Paul 

Thank you. Thank you. Thank you.

SigmaCorns

The SigmaCorns are all students at the NC School of Science and Math, a public high school in Durham NC that aims to foster an intellectually stimulating, diverse, inclusive, and collaborative community of students.

More from SigmaCorns.

Jacob Van Meter

Jacob is an AI Research Engineer and Entrepreneur who’s always been focused on building in the future. He’s the Co-Founder of LRNT, the startup building Athena, a revolutionary AI data science learning assistant currently undergoing pilot programs. They’re also an AI Research Engineer at Plastic Labs, a startup solving the principal agent problem and just happens to be a student studying CS at UNC Chapel Hill. 

Ayush Paul

Ayush is a developer, designer, budding entrepreneur, and a high school student at the North Carolina School of Science and Mathematics. As a co-founder of LRNT, they utilize AI and Computer Science pedagogy to build Athena, an innovative AI Data Science learning assistant currently making waves in pilot programs. In addition, Ayush is an avid roboticist, currently the software captain of FTC #22377 the SigmaCorns. Furthermore, they’re the founder of Queen City Hacks, Charlotte’s largest high school hackathon, and an AI Research Engineer at Plastic Labs, a NYC startup solving the principal-agent problem. 

Next Events

BoS Europe 2025 🇬🇧

🗓️ 31 March – 1 April 2025
📍 Cambridge, UK

Spend time with other smart people in a supportive community of SaaS & software entrepreneurs who want to build great products and companies.

Save your Spot

BoS USA 2025 🇺🇸

🗓️ To be announced soon
📍 Raleigh, NC

Learn how great software companies are built at an extraordinary conference run since 2007 to help you build long term, profitable, sustainable businesses.

More Details