Loading summary
Transcript104 lines
- [00:01]
Sai
So good, so good, so good.
- [00:03]
Nordstrom Rack Announcer
Score holiday gifts. Everyone wants for way less at your Nordstrom Rack store. Save on Ugg, Nike, Rag and Bone, Vince Frame, Kurt Geiger, London and more.
- [00:13]
Brian McCullough
Cause there's always something new.
- [00:14]
Nordstrom Rack Announcer
I'm giving all the gifts this year.
- [00:16]
Brian McCullough
With that extra 5% off when I.
- [00:18]
Nordstrom Rack Announcer
Use my Nordstrom credit card. Santa who join the Nordy Club at Nordstrom Rack to unlock our best deals. It's easy. Big gifts, big perks. That's why you rack Meet the computer you can talk to with Copilot on Windows Working, creating and collaborating is as easy as talking. Got writer's block? Share your screen with Copilot Vision to help spark inspiration and use Copilot voice to have a conversation and brainstorm ideas. Or maybe you need some tech help with Copilot Vision. Copilot sees what you see. Let Copilot talk you through step by step guidance so you can master new apps, games and skills faster.
- [00:55]
Sai
Try now@windows.com copilot today, hypercubic is allowing Fortune 500s to understand, preserve and modernize their old legacy infrastructure. And so what I mean by legacy infrastructure is there's a lot of enterprises that still run COBOL and mainframes, and this was technology that was built back in the 1960s that runs almost all of the big industries today that you can think of such as airplanes, airlines, banking, financial services, logistics, the government like irs, Social Security, and so on. And so there's a lot of legacy technology out there that's still running. And what we do is simply help these enterprises understand them and modernize that into a modern tech style.
- [01:42]
Brian McCullough
Welcome to another bonus episode of the Tech Brew Ride home. I'm Brian McCullough. As always, this is a portfolio profile episode which I It's been a couple months since we've done one of these, but I'm super excited today to talk to the founders of Hypercubic, which you can find out more about. At Hypercubic AI, we have Sai and Ayush. Please introduce yourselves.
- [02:11]
Sai
Hey guys, I'm Si. I was previously a machine learning engineer at Apple for a couple years where I worked on the iPhone multitouch and before that I did research in grad school and undergrad on LLMs way before they became mainstream and today working on Hypercubic with Ayush on using AI for the mainframe ecosystem.
- [02:32]
Brian McCullough
Ayush.
- [02:34]
Ayush
Hi, I'm Ayush. I'm co founder CTO of Hypercubic. I was a software engineer at Apple AIML before this and my research has been at Grad school around robotics, nlp, AI, and excited to be here today.
- [02:50]
Brian McCullough
For whoever wants to take this, give me the 1 minute elevator pitch of what hypercubic is doing and then let's dive into what hypercubic is doing.
- [03:02]
Sai
Yeah, so today Hypercubic is allowing Fortune 500s to understand, preserve and modernize their old legacy infrastructure. And so what I mean by legacy infrastructure is there's a lot of enterprises that still run COBOL and mainframes. And this was technology that was built back in the 1960s that runs almost all of the big industries today that you can think of, such as airplanes, airlines, banking, financial services, logistics, the government, even like irs, Social Security and so on. And so there's a lot of legacy technology out there that's still running. And what we do is simply help these enterprises understand them and modernize that into a modern tech stack.
- [03:44]
Brian McCullough
Do you have any stats? I mean, you're mentioning like, you know, major pillars of the economy and things like that, but do you have any stats off the top of your head in terms of how, how pervasive these legacy systems are? Like, are we talking about 30, 50% of like the systems running out there? Like how big is this problem that you're solving?
- [04:07]
Ayush
We have a few stats and some of them are like something like 70% of all Fortune 500 use mainframes for something. Another is like, I think Cobol Sai mentioned is like a programming language that many of these systems run on. There's something like between 200 to 800 billion lines of COBOL still running out there. So these are pretty big numbers. People think like COBOL has died, like this is like a dying market, but there's just so much of it out there that's not true at all.
- [04:37]
Sai
Well, and I think the key stat, one more thing to mention is the average age of the developers working on these Systems is about 55 and rapidly going into like 60s, 70s. There's plenty of 70 year olds working on these systems today. They simply cannot retire.
- [04:53]
Brian McCullough
Well, I was going to. I'm glad you interrupted me because I was going to make a joke about dying. You said dying. But the problem you're solving is that it's not just legacy systems, it's legacy knowledge. It's knowledge that can age out of an organization. And so the code is still running, but the ability to tweak the code to improve the code, to fix the code, that sort of knowledge might be, and I'm not going to say dying, but walking out the door, getting a gold watch, that sort of thing. And so that is the problem that you're trying to solve.
- [05:33]
Sai
Exactly. So it's the system understanding of these legacy infrastructure today. What's happening is a lot of people are focusing, the competitors or the incumbents are focusing on the code aspect, line by line, explainers, code explainers and so on. But the other missing part of the puzzle is basically the institutional knowledge in the minds of these subject matter experts. Like I mentioned, these 70 year olds that have worked on these Systems for about 30, 40 years of their careers, and when they leave the workforce, all of that simply vanishes or leaves with them. And so capturing all of that is the key mission that we're going after.
- [06:13]
Brian McCullough
So yeah, go ahead. No, go ahead.
- [06:16]
Ayush
Adding on to that when you're talking about this knowledge, a lot of the common question we get is, isn't all of this knowledge much of this knowledge already there in large language models, or aren't they eventually going to, eventually going to be in these models?
- [06:30]
Brian McCullough
Right, right.
- [06:31]
Ayush
And I think one key thing to understand here is yes, the system operational or manuals are going to be there in the large language models, but every company has its own cultural way of doing certain things and that is almost never going to be in the language models. And that is one of the key things that we're focusing on right now.
- [06:51]
Brian McCullough
Right. The language models can be trained on cobol. You can have IT code in COBOL for you. But it is more of that institutional knowledge mixed with the legacy coding that you're trying to fix here.
- [07:08]
Ayush
Correct?
- [07:09]
Brian McCullough
Yes. Can AI agents, can they fix this problem in the sense of what do I have to do if I'm, if I come to you and say, hey, we've got all these legacy systems, how can you help us? What is the installation? What do I have to do to get up and running, aside from showing you my code base? What do I have to do to get hypercubic up and running to help me?
- [07:43]
Ayush
I'll start and then I can continue. We have this crawl, walk and run approach that I think we've talked about internally, where crawl is like an easy way to get started with us, which is just, I think you mentioned, or even before the call, like hyperdocs and hypertwin. So step one is just the understanding part. Right. We try to come in, we will understand your code, we'll sort of interview your experts using AI to get the knowledge in. And that's the crawl part. The AI agents part come later. There's more to our vision eventually. Like I think we said initially, we Want to be able to move entire systems onto modern tech stacks. But I think going straight there is like, has been done many, many times and has failed. So we're not sort of going there directly right now. We are sort of doing two things, like I said, the capturing knowledge from code and capturing knowledge from experts.
- [08:36]
Sai
And to go a little bit deeper into that, just to elaborate on sort of what we're speaking about when we say going deeper into the code, we have a documentation platform that ingests all of this legacy code. You know, it's millions of lines of code, tens of thousands of files to spread across their mainframe systems. And we simply ingest all of that and convert it into a very readable, highly structured documentation that their engineers, business analysts and leadership can simply go through to understand these now black box systems. So that's one part of the puzzle.
- [09:08]
Brian McCullough
That's the hyperdocs, right? That's the interesting. Okay, and you're using, like, what a hybrid of like, deterministic and generative AI. Like, how are you. How are you making sure things like accuracy and audibility and trust are part of that ingestion process?
- [09:30]
Ayush
One of the things that we're doing is like, sort of having. So for many of the things that we're generating, we're sort of trying to tie them back to proofs. And by proofs, I mean when we're just ingesting code documentation for every statement or every paragraph or every, like a diagram that we generate, right? We sort of link back to like, okay, these are the code. Like, these are the files, these are the code blocks where we're sort of extracting this business logic from. And this is where it comes from. So it's like there is an element of, like, human verification to it, but we make it very easy for that verification to happen. And we have all kinds of editing tools and versioning to make that as verifiable as possible. And like you said in the beginning, there's like, yeah, we have a bunch of generative models and like, some deterministic stuff together, sort of making it more reliable than like a purely generative system would be.
- [10:30]
Brian McCullough
The second part being the hypertwin part, that's. That's the. I think you said, the knowledge layer that. So, like, you're building a digital twin of the experts inside the organization. How does that work? How do you, like, are you learning from observing their work and stuff like that?
- [10:55]
Sai
Yeah, so this was a correct. So this is a combination of a few different modalities of information that we collect and the whole goal is to basically replicate their mental model of the systems, the processes, the expertise that they're working with. And so the way we collected, without going too deep into the proprietary stuff, is first is ingesting the existing documentation they have written. So this could be SharePoint, GitHub, Jira, anything that's been out there. The second one is an AI driven interviewer. So the AI driven interviewer would ask very targeted questions to the subject matter expert. For example, there's an issue last month the interviewer would ask, hey, we've seen this issue with this specific lpar. How did you resolve or debug this specific error that we're seeing? Could you walk me through the problem solving process? That's one way. The third modality is simply workflow capture. That means recording the screen of some critical tasks or workflows they're working on and then converting the information from that task into structured workflows. And that essentially means, hey, Richard, the SME maybe has gone through these different files, he's accessed all of these different things, here's the functionality he's implemented and this is the outcome of that. We take all of that structured knowledge of how we went through that task and by combining these three modalities of information, we can create almost a digital twin of the subject matter expert. And the way they work, problem solve and architect solutions.
- [12:25]
Brian McCullough
Is that though, is that something that again, when I'm getting set up with you, you do it for a period of time and then it's ready to go, or is it something that's always there and always learning so that again, maybe it's a two week process or whatever to get set up or do the interviews and things like that. Or is it something that you're monitoring so that when new problems arise, you can still. It's like, hey, you'll get a ping, we need to talk to hypercubic again and give them more info?
- [12:59]
Ayush
I think it's kind of both. So one, the interview process itself, I was surprised to learn that many would want to because the knowledge that's in these experts has been accumulated over like 30, 40 years, right? It'll take them a year of just doing interviews to just put everything in. So there is an aspect of we might not even need the really current or present things are not even just getting what's in their head, it's just like a year long or even longer process. So that's one part of it, but to the other part, which is there's the workflow capture that Sai talked about and that is much more proactive or even reactive in the sense if things happen or there's a new thing, the expert can just do the thing and hypertwin can watch you do the thing and that just goes into it automatically. And the next time it happens there's.
- [13:54]
Brian McCullough
Like it's captured and it like it's continuously learning. Yeah. And ingesting. Yep.
- [14:00]
Sai
So the twin is almost self evolving in the sense that as new information is being fed in continues to get updated over time.
- [14:09]
Brian McCullough
So for enterprises the value proposition is obviously, you know, preserving knowledge, but also is there like risk, risk reduction and like cutting efficiency in terms of production ready like producing things like is there efficiency here also beyond even just the preserving the knowledge?
- [14:39]
Sai
So the key value proposition is it's preserving the knowledge. But then what is the outcome of preserving the knowledge which is really just risk mitigation of these legacy systems. Because if you think about it, when a large bank like JPMC or Bank of America, when they're running these systems with millions of customers and tens of billions of dollars of transactions, they cannot afford to lose or any downtime on these systems. And so it's really about mitigating any and as much risk as they can that could potentially occur on these systems. Whether if it's subject matter experts retiring some critical issues within these systems. Trying to mitigate all of those issues is sort of the business outcome that enterprises are getting out of the tools.
- [15:22]
Ayush
We'Re building just to add on to that. Even like the initial conversations that we've been having for pilots and all they've all come out of like okay, here's a risk. Here is like potential people that we might lose. How do we mitigate it? So it's always like all the conversations we have had is always in the background, larger background of mitigating risk that is like the main thing that they're doing with cartoons.
- [15:46]
Brian McCullough
And is there a layer of proactivity in the sense that hypercubic, once it's up and running, once it's learned and ingested that it can proactively solve problems that might arise.
- [16:03]
Degree Cool Rush Announcer
We all love a legendary comeback and degree. Original Cool Rush is back and better than ever. Cool Rush isn't just a scent, it's it's a mvmt, a fan favorite that delivers bold, fresh vibes and all day sweat protection. Whether you have a man that spends hours in the gym, heads into the office early, or is just trying to stay fresh on a long day, Cool Rush has their back. Head to your local Walmart or target and grab degree Cool Rush, the fan favorite scent from the world's number one antiperspirant brand.
- [16:35]
Netflix Announcer
This episode is brought to you by Netflix from the creator of Homeland. Claire Danes and Matthew Rhys star in the new Netflix series the Beast in Me as ruthless rivals whose shared darkness will set them on a collision course with fatal consequences. The Beast in Me is a riveting psychological cat and mouse story about guilt, justice and doubt. You will not want to miss this. The Beast in Me is now playing only on Netflix.
- [17:04]
Call of Duty Announcer
Hello, friends. Guess who? That's right, it is I, the replacer. Once again, I've been called on so you can play the new Call of Duty Black Ops 7 with three expansive modes, 18 multiplayer maps, and the tastiest zombie gameplay you've ever freaking seen.
- [17:25]
Brian McCullough
Call of Duty Black Ops 7 available now.
- [17:28]
Call of Duty Announcer
Rated M for mature.
- [17:32]
Ayush
So to understand correctly, proactively solving problems. So you mean like if that is.
- [17:39]
Brian McCullough
Like if there's an outage, can I turn to hypercubic? And maybe you're not there yet. And again, I don't want to go into product as it's developing, but like, can it say, okay, this has happened, this outage has happened. I think this is why and this is the solution that I can make for you.
- [18:00]
Ayush
So I think it's more like they're still going at its current state. There's still going to be a human kind of involves like the person who's going to go in to resolve this incident. Right. Instead of them just having to go in without any context, they would sort of go in and just our interface is very minimal. If we have a demo, we'll show you. All you do is hit a shortcut and just quickly ask a question. It's able to look at your screen, look at stuff on the screen and just give you some idea of what's happening very, very quickly.
- [18:29]
Brian McCullough
It's still guidance. It's not agent in the sense that it's going to fix it for you. But that could be down the road.
- [18:34]
Ayush
Which is fixing, modernizing, actually building stuff. Because we want to be able to get into enterprises with an initial. That's our advantage. This is how we get in. This is what we want to do initially. Solve an actual problem that helps them out. But we do want to do bigger things eventually. So I think more of actually rearchitecting these systems. That is something eventually. What we want to do is be able to take an old system and completely re architect it based on all the knowledge we have gathered so far and how can we make it with as minimal risk as possible. That's the longer term vision that we are thinking of.
- [19:16]
Brian McCullough
Yeah, it's extremely early in the sense that I think I only spoke to you, what, in April of this year or something. Actually, let's go into that for a second. When I spoke to y', all, you still hadn't really launched yet. You were your former Apple folks. So can you tell me where this idea came from, what the inspiration was and why you felt like you were the ones that could solve this problem?
- [19:51]
Sai
Yeah, absolutely. So as you know, Ayesha and I met at Apple, we were working at a hackathon, randomly sitting next to each other and then just happened to hit it off. And over time we started building a bunch of different projects together and we had a similar vision in terms of the things we wanted to work on and so on. And one of the components of the vision was knowledge. And knowledge retention was something very big that we focused early on. And so we realized there's a lot of institutional knowledge in these enterprises and that's very.
- [20:22]
Brian McCullough
Let me interrupt you real quick. Have you personally, either one of you encountered that where there's a gray hair, there's a veteran that has this knowledge. You're working at Apple or wherever, there's a problem that you couldn't solve and you're like, man, that person solved it for us because they know. Have you ever had any personal events like that that maybe inspired this?
- [20:48]
Ayush
Yes, we've had so many. Even though, like for our situation at Apple, even though it wasn't like a gray haired veteran, like we've come across that situation, like, oh, if that person X had left and if he had his brain somehow or if he had his mind somehow, I'm saying he. Because I know exactly who I'm thinking about when I'm saying all of this. If that person had not left, we could have just solved this in minutes. But it took us days. And I think some version of the problem of knowledge loss or tribal knowledge retention, it's most critical in the mainframe space. But I think almost all companies have to deal with this.
- [21:24]
Brian McCullough
Well, I was going to say even not just again, we're using gray hairs. I don't want to be pejorative to anyone that's older than even me, I'm older than you all. But it's also like someone could get hired away or go off and do a startup somewhere or something. And so that's the institutional knowledge that you can leave. So even the larger thing is it's not about Retaining talent. Talent can move on, but you don't lose the talent if you have the legacy knowledge of that talent. All right, I interrupted. So you were doing a hackathon and you hit it off. I'm sorry, continue.
- [22:06]
Sai
Yes, no worries. I mean I just sort of hinted, foreshadowed some of this as well. We, you know, we realized that institutional knowledge was a very big problem, especially at some of the workplaces, without going too deep into what that is was. And we've had a hypothesis that maybe there's other large enterprises out there that are also facing the same issue. And so we just simply went out there, spoke to tens of, you know, tens of people out there in different sectors such as banking, defense, retailers, government and so on. And they realized this was a major problem across all the other industries. And throughout those conversations we've discovered there is a specific space called the mainframe ecosystem. And specifically COBOL engineers who are facing this as sort of a hair on fire problem because of, you know, the average age is about 55 and 60. A lot of them are retiring. These are very critical systems that are close to the heart of the business. And we realized we might have struck some gold here. And so we started pulling the thread on this and over time evolved into the company we have built so far.
- [23:13]
Brian McCullough
Six months later and where is the company right now in terms of the product? Like we'll get into more if you're interested, how to get in touch in the end. But like if I am an airline or an automotive company or a bank, somebody that's using these legacy systems, how ready is your product for me to use and in what way can I use it?
- [23:42]
Sai
Currently we're working with a few pilots right now, some of the largest enterprises today. And so we have fully functioning products of the two things that we've just talked about, the Docs and the hypertwin. And so we are working on design partnerships, pilots with these folks. And so if someone is interested in coming along to test them, we would be happy to onboard them, use the products and see how we can work together on some of the challenges that they're facing in their orgs and institutions.
- [24:11]
Brian McCullough
And we're thinking of things like financial services, aerospace, utilities, government sectors like that, right? Correct.
- [24:21]
Ayush
And one thing I was just going to add because we're sort of like in earlier stages and like doing design partnerships, right. So the one thing, the one kind of benefit that people who just reach out to us right now is like we're able to sort of build the product towards their Use case. Like, it sort of develops with them and with their cases, so it, it becomes very specialized for like the kind of. The early design partners will have like a lot of say in our products. So that's like a. Both. It's, it's a real.
- [24:56]
Brian McCullough
Let me ask you a personal question in the sense of you were both at Apple. That's where you met. Good job, great company. Personally, what is it like to. When you're at a big tech company, good salary, stock options and stuff to say, you know what, we're going to go off and do this on our own. Is it something that you had considered before? What is the thing that for both of you that says, you know what? We really, we're gonna do this?
- [25:31]
Sai
I think one of the things we had in mind was without going too deep into specifics of the workplaces we've been at, one of the things we had in mind was, you know, we're super young. There's just a lot of opportunity right now. And the most important was when there's a special window of opportunity right now with AI and LLMs and sort of the applications that are possible that weren't possible one or two years ago. And so it seemed like the perfect time right now to sort of take a bet on ourselves, take a leap of faith. And that's something that was a shared vision for both of us. If I just have something very specific as well.
- [26:08]
Ayush
Yeah, I think for more personal. If I want to say more personal, I think this holds true for both of us, like from the perspective of personal constitution. I think both of it were just like, we would honestly go mad if we had stayed back there or back in. It's not a tug on Apple or anything. It just talks about how fiercely independent or fiercely. We wanted to go and chase out something. So it was gonna happen. Both of us were even long before we met, there was a main bucket list item. It's even too strong to say bucket list item. We want to start our own thing that was always there even if the AI wave had never happened. So, I mean, AI, we've just made it more urgent than ever, but I think we were going to do this no matter what.
- [26:54]
Brian McCullough
Well, right. So you feel like this is our moment, let's meet the moment. But also, as mentioned previous AI background sort of helps. So it's sort of like. Let me use the analogy of like, if, like you're in a city and all of a sudden there's a. A music scene that hits and you're like, well, look it's all taking off. Let's start a band or something. Like, did you sort of like be feel like your background sort of made you ready for when this moment hits?
- [27:30]
Ayush
Yeah, I think absolutely. Because like Sai talked about, we met at a hackathon and we built like a, like something related to tribal knowledge capture. Even back then, like, that was our like original idea and we're just playing around with it and even the whole. Even our research that came before that. So a lot of this stuff, it was really easy for us to pick up as things were moving really quickly. It's not like we know everything, but I think we do have many younger folk who are just sort of doing research in AI or who built an AI before coming into the AI wave, we had an upper hand of, oh, we already understand a lot of it. We just, we just keep being. All we need to do is be at the forefront and keep being at the forefront and we've got this.
- [28:19]
Sai
And the interesting thing I'll add is, at least personally, a lot of the dots looking backwards started connecting immediately. For instance, in grad school I was supposed to specialize in computer vision. Luckily that department wasn't that great. In my college at Virginia Tech, I chose the second best option, which was natural language processing. GPT2 just came out. There wasn't much of a buzz out there. But then I did my research, published my papers and over time this LLM Web just started taking off. I was already at that forefront of this frontier that was just opening up and now looking back, it's all just connecting the dots.
- [28:56]
Brian McCullough
Basically, as you get older, you see the dots where you're like, well, that obviously was going to happen. Where at the time you didn't know it was going to happen, but you were in the right place and knew the right people and done the right education. So you were mentioning earlier projects and things like that you are going through Y Combinator right now. Demo day is coming up at the time of this recording, November 10th. But I think you had applied to Y Combinator a few times before. Like this isn't your first rodeo. Tell me the story of applying before and not getting in.
- [29:40]
Ayush
I think SAIS applied a bunch. Like SAIS applied like seven times and I think we together we've applied like four times. So we've been rejected a bunch of times and the last time we got accepted. But before that we've applied with all kinds of ideas. Like there was a time like I think it was last winter where we applied with like a dating app because we were just like throwing ideas on the board and it's like, I don't know, Sai, if you want to add more color to that.
- [30:07]
Sai
Yeah, I think I've applied a bunch of times during grad school and undergrad almost every semester. I was looking for a way to simply leave college and just go out to San Francisco and build something. And eventually the seventh time was a lucky charm, basically. But in the last six months, what Ayush and I did before we started this was basically we started throwing different ideas along. We made a list of 100 different ideas that we can basically work on and started just checking each one off in terms of what is the feasibility of it. Could we actually go into that space and over time narrowed down into this space? I think a couple other second contenders were something in the health space, if I remember correctly, and so on. But eventually narrow it down to this mainframe space.
- [30:53]
Brian McCullough
Well, a dating app or a dating whatever. That's how YouTube got started. A lot of startups that started out as dating apps and then people pivoted to things. Let me ask you this though, again, connecting the dots later. Do you feel like this was the idea that probably maybe you shouldn't have done it seven times? This was the one that was really the one that could hit. What is it like when you apply and then you apply so many times you get turned down and then you get accepted? Like what is. How does the application process differ versus just throwing something in? Versus. Okay, we want to talk to you more. We want to talk to you more.
- [31:42]
Nordstrom Rack Announcer
Okay, you're in the family that vacations together, stays together. At least that was the plan. Except now the dastardly desk clerk is.
- [31:52]
Netflix Announcer
Saying he can't confirm your connecting rooms. Wait, what?
- [31:55]
Brian McCullough
That's right, ma'.
- [31:56]
Call of Duty Announcer
Am.
- [31:56]
Brian McCullough
You have rooms 201 and 709.
- [31:58]
Netflix Announcer
No, we cannot be five floors away from our kids.
- [32:01]
Brian McCullough
Eh, the doors have double locks. They'll be fine.
- [32:03]
Netflix Announcer
When you want connecting rooms confirmed before.
- [32:06]
Nordstrom Rack Announcer
You arrive, it matters where you stay.
- [32:09]
Netflix Announcer
Welcome to Hilton. I see your connecting rooms are already confirmed. Hilton for the stay.
- [32:16]
Degree Cool Rush Announcer
This episode is brought to you by Rumchata, A delicious creamy blend of horchata with rum. It's best enjoyed over ice or in your coffee. Delivering vacation vibes anyway. Or anywhere you drink it. Find out more@rumchata.com Drink responsibly. Caribbean rum with real dairy cream, natural and artificial flavors. Alcohol 13.75% by volume 27.5 proof. Copyright 2025 Agave Loco Brands, Powkee, Wisconsin. All rights reserved.
- [32:45]
Netflix Announcer
I am so excited for this spa day. Candles lit, music on, hot tub warm and ready. And then my chronic hives come back again in the middle of my spa day. What a wet blanket. Looks like another spell of itchy red skin. If you have chronic spontaneous urticaria or csu, there is a different treatment option. Hives during my next spa day. Not if I can help it. Learn more@treatmyhives.com Brian I think that's a.
- [33:16]
Sai
Great question because that was something very clear. I noticed in the application where we got accepted compared to all of my previous applications where it was simply a matter of I was working on something else on the side. YC deadline is coming up. Hey, let me go ahead and apply with an idea that I have. Whereas the application that we actually got in, we were very intentional in terms of making progress. Regardless of YC worked out or didn't work out, we were super focused on this idea. We knew there was a big market and a big opportunity for us to go into this space and simply submitted an application and then we simply forgot about it after. And then we started doing our customer discovery, building the product and just building the business. And so whatever YC says in terms of their advice on building a startup and Paul PG's advice about not caring too much about YC is actually the truth, which you should be focused on building your business and YC is just a byproduct. If you get in, you get in. If not, continue building your business.
- [34:14]
Ayush
Adding some more color to that just to give an example of like every application that we submitted before, I remember just doing the YC application was so hard. Every question was just like we don't know how to answer for every question. We have to spend hours and hours to come up with an answer. And I think for our last application everything was just so clear. I think Sai just during one free evening he just filled it in and submitted it. It was just like a very offhand thing because all the things were just so clear. So I think just adding on to that. If you just build something with intention and are really in the space full time and just focusing, the answers sort of just write themselves.
- [34:56]
Brian McCullough
That is so amazing in terms of clarity of if you apply to YC and this can apply to anything, like you apply to college, you apply to try to get an agent for a book or something. If it's like, well the application or what you're trying to do is well this will only work if you accept me or even apply for a job, it's almost like a hiring agent can understand that, that like, you know, I'm. I'm coming to you and I'm crossing my fingers versus what you're saying is on this eighth time or whatever of applying, you're like, well, we're going to do this anyway. We're applying this time because, you know, it would to get into Y Combinator, but we're going to do this one way or another. And it's almost like that sort of intentionality comes through in an application. I don't know if you agree with that, but I think that's an interesting lesson.
- [35:56]
Sai
Absolutely. I think that's precisely it. Which is just simply not caring about the outcome and just focusing on the. You want to do. You want to build the business, basically.
- [36:11]
Brian McCullough
Yeah, I'm going to make this happen one way or another.
- [36:14]
Ayush
Exactly.
- [36:15]
Brian McCullough
When you do get in, and this is the last question about yc, but just give me a little bit of color on what are the resources, what is the onboarding, what is the excitement and things that you didn't expect that Y Combinator does for a cohort company.
- [36:38]
Ayush
One thing that was like, really, really useful that we kind of, even though we kind of expected, we were still surprised by it because there's like an. So we had heard, like, when you get into yc, it's really intense. It's really intense. And every time I've asked people, like, what do you mean by intense? And they've. No one's ever really been able to describe like, what that. Where that intensity comes from. But it's sort of the. A combination of like, your partners helping you set like really ridiculously ambitious goals, plus you looking at your peers doing really well. And like, there's like the social pressure. There's like a combination of things that's like, makes you want to work really, really hard and just push what you thought was possible. So that was both surprising and not surprising. That was one of the things for.
- [37:22]
Sai
Sure for me, I think, to add on to that, I think it was second week of YC and we went into our group office hours and sort of our partners made us commit to this absurdly high goal that we've never even thought about. And so we walked out thinking this was a revision, but it actually should be up here essentially. And so I think YC raised up the bar in terms of sort of shooting for the stars in terms of your vision. So that's been incredibly helpful. And then obviously all the partners are. They've seen and worked with hundreds of startups over the years. And so the amount of information and insights that they have on company building, you know, how to build a product, talking to customers, how to scale, it's all incredible information that we get access to on a weekly basis.
- [38:08]
Brian McCullough
What, what do they do in terms of introducing you to investors? I think, I think you meet with investors before demo day a little bit. But also demo day is, is announcing yourself to investors. But like what are the introductions like in terms of do they, do they sort of match make where it's like you would be right for this person. I want to. How do they do that?
- [38:32]
Ayush
I think the heavy lifting is done by sort of the YC ecosystem and the YC brand. So what they've done is like made it a norm for make the investors sort of approach us in the sense they go to like the YC startup directory page, look at all the startups and they kind of come in. So like the majority of the work is just done by the brand and like the way the whole, I think during the whole like three months the partners heavily recommend you to not talk to investors, just like sit down and just build stuff and work and do all of that. So rather than it just being like a matchmaking or like an introduction, it's just because it's just more investors sort of just reach out to you because you're a part of IC and you're doing something cool.
- [39:22]
Brian McCullough
Yeah, sorry Sai.
- [39:24]
Sai
No, I think the only thing I was going to add is it almost helps you as a founder because during the duration of the time, fundraising isn't even on your mind and your only focus is to talk to customers, build your product. NYC takes care of the sort of the legwork of reaching out to investors, getting attention, booking meetings and all of that. So just the brand itself helps you with the entirety of the seed fundraising process. And it culminates with a demo day at the end of the batch where you pitch to some hundreds of investors at their dog patch hq.
- [40:01]
Brian McCullough
Well, let me ask you about that final question about yc. Are you feeling pressure? Are they giving you training on demo day? How are you thinking about demo day and do you feel prepared for it?
- [40:16]
Sai
Yeah, Demo day is in about two weeks or so. We're super pumped up to go into the meetings that we have scheduled the next two weeks. Most of the hot companies tend to raise before demo day and by demo day they're sort of fully, the allocation is complete. But we've been getting a lot of investor inbound, a lot of interest just because of the niche space that we're working on. So, yeah, just super excited about what's coming in the next few weeks.
- [40:45]
Brian McCullough
Okay, so let's wrap this up by saying, I'll put this out this weekend. I think if you want to get involved with this, first of all, if you're a company that could use this, how should I get in touch with you all?
- [41:03]
Sai
So you guys can go to Hypercubic AI and simply book a demo and we will reach out to you within 24 hours to schedule something on your calendar. And we'd love to show you the products that we have and how we can fit the challenges that you guys are working on in your old.
- [41:20]
Brian McCullough
What about if I want to get involved? Are you hiring anything like that?
- [41:26]
Ayush
Yeah, we're just getting started. So, like, we've talked about fundraising. Three, four weeks from now, we'll probably have been done with our fundraising and we'd be really, really looking for amazing founding engineers. So if you're someone, like, someone who's like, who wants to build a company eventually, but wants to get really, like, wants to compound themselves and like, learn like crazy over the next year, two years, three years, we'd love to have you. Love to meet amazing people. And for that, I think just reach out to me or SAI on LinkedIn or email at us at teampercubic AI, we don't have a formal careers page yet because we're quite early, but we are starting to look at hires and that's like the biggest reason for us and most companies to fundraise.
- [42:12]
Brian McCullough
Quick edit here. There was one more thing that SAI wanted to include that I forgot to ask. So sai, go ahead.
- [42:22]
Sai
If you have any leaders, connections to leaders in the mainframe ecosystem or large enterprises in the Fortune 500, we would love to speak to them. If they're maintaining or have ownership over these mainframes, or if they simply have connections to those that are working on these systems, we'd love to get in touch with those. So please feel free to reach out if you have any connections in that space.
- [42:44]
Brian McCullough
You don't even have to be those people. You can be the people that know the people is the point. Yes, exactly. I would also say if you're a listener to this pod and you know me well enough, email me@brianidehomefund.com and I will put you in touch with the guys and amazing opportunity. Like, literally, if you think this is a good idea, like I did eight months ago, to invest in it, you can get in on the absolute ground floor. Listen, again, I want to repeat hypercubic. AI Sai. Ayush. Good luck on demo day. But thanks for coming on to tell us all about this.
- [43:30]
Sai
Thanks for having us.