In this episode of The Next Next, host Jason Jacobs speaks with Uday Chandra, co-founder of Singular, a startup developing an adaptive business operating engine. Introduced by former guest Mady Mantha, Uday shares his story of transitioning from a failed fintech venture to a bootstrap approach with Singular. Singular leverages AI to continuously crawl the web and create 'software Lego blocks' that replace multiple SaaS subscriptions with a unified platform. The company has achieved $45k in MRR from five customers in just three months. Uday discusses the challenges of integrating AI, the importance of open source tools, and the mission to build ambitious companies while maintaining flexibility and control. Despite the small team and lack of external funding, Uday expresses high ambitions of reaching $100 million in ARR and becoming a unicorn. The conversation delves into the potential for AI tools to disrupt the SaaS market and explores strategies for customer acquisition, pricing models, and maintaining a lean operation.
From Startup Struggles to AI-Driven Success: Uday Chandra & Singular's Adaptive Business Engine In this episode of The Next Next, host Jason Jacobs interviews Uday Chandra, co-founder of Singular, an innovative startup focused on building an adaptive business operating engine using AI. Introduced by Mady Mantha from Happypillar, Chandra shares his journey from a failed fintech startup to successfully bootstrapping Singular with his co-founder. Singular aims to consolidate various SaaS tools into a single unified system, leveraging open-source technologies and AI-driven agents for coding and business rule integration. Despite being only three months old, Singular has achieved $45,000 in monthly recurring revenue (MRR) from five customers, with a vision of scaling further without outside funding. The episode delves into the challenges of integrating AI, customer acquisition strategies, and future ambitions for the company.
00:00 Introduction to Uday Chandra and Singular
00:31 Uday's Journey: From Traditional Startup to AI-Driven Success
01:13 Singular's Innovative Approach and Early Success
02:09 The Vision Behind Singular: Consolidating SaaS Tools
02:41 Meet the Host: Jason Jacobs and The Next Next
03:28 Uday's Background and the Birth of Singular
04:02 Challenges and Learnings from UKey
08:38 Identifying the Problem: SaaS Fatigue
10:10 Building Singular: Customer Development and Market Research
15:01 Targeting the Right Market Segments
16:42 Developing Singular's Features and Functionality
19:53 Leveraging AI for Software Development
24:24 Pricing Model and Customization
27:38 The Future of SaaS and AI Integration
36:53 Opinionated Approach to Software Development
38:49 Initial Customer Acquisition Strategies
39:54 Pitching to Mid Enterprises
40:24 Overcoming Customer Skepticism
41:41 Navigating Tool Integration Challenges
45:25 Leveraging AI for Business Optimization
50:19 Scaling with a Small Team
01:00:31 Open Source and AI Innovations
01:08:07 Building a Collaborative Community
01:11:46 Final Thoughts and Future Outlook
Jason Jacobs: Today on The Next Next, our guest is Uday Chandra. Uday is co founder of Singular, which is building an adaptive business operating engine. Uday was introduced to me by Mady Mantha from Happypillar, who came on the show recently, and she said that Uday was interesting because he had a small team doing big stuff, leaning hard into AI.
I didn't know much more before Uday came on the show, but what a fascinating discussion. It turns out that Uday had just got done trying to build a business the more traditional way, and they struggled with product market fit, and they shut it down. And so Uday and a co founder set out this time to not try to go out and raise money, but to bootstrap and to Stay small by design to lean hard into AI and agents specifically to do a bunch of the coding and to see how big they could take it with an agent army, so to speak.
Well, they are three months in. They have a two person founding team. They're up to 45k in MRR from five customers. They have an agent that's crawling the web continuously to identify which tooling areas need coverage in these. And this is, small medium businesses or, mid size enterprises that have a bunch of SaaS subscriptions.
And what they're doing essentially is they're consolidating so that if they work with Singular, Now you can have one subscription and you can still get those SaaS tool functionality but through these Lego blocks of open source products with these business rules wrapped around them. And that's Singular.
So, at any rate, 45k in MRR from five customers and this is across task management, communications, customer management, content management, etc. And then they've also got an internal agent who's building a bunch of these software Lego blocks Again, with agents writing much of the code using open source products to make viable alternatives to the individual SaaS provider incumbents.
Now three out of the five customers have already booted out their SaaS subscriptions and saved a bunch of money. So really interesting and Uday has big ambitions. He's thinking about getting MRR, granted they're just getting started, but. Even more interesting is he's thinking of doing it with a really small team all the way through and with no outside funding Without further ado Uday Chandra, but before we get started
I'm Jason Jacobs, and this is The Next Next, it's not really a show. It's more of a learning journey to explore how founders can build ambitious companies while being present for family and not compromising flexibility and control, and also how emerging AI tools can assist with that. Each week, we bring on guests who are at the tip of the spear on redefining how ambitious companies get built.
And selfishly, the goal is for this to help me better understand how to do that myself. While bringing all of you along for the ride. Not sure where this is going to go, but it's going to be fun.
Okay, Uday Chandra, welcome to the show.
Uday Chandra: Hey, Jason. Thanks for having me.
Jason Jacobs: Thanks for coming on. Yeah, so we got introduced through Mady Mantha an earlier guest And it sounds like you guys are friends and and your story's interesting because you It sounds and correct me if I'm wrong, but that You tried to do a startup more the traditional way and had difficulty in terms of raising funding and now you're on to a new thing where you're doing another startup but keeping it much leaner and profitable and no outside capital and no plans for outside capital.
Is that all?
Uday Chandra: Absolutely. Yeah, my previous one, UKey, we shut it down a few months ago because more than having difficulty raising funds, it was that we didn't find the product market fit that we thought we would. We built a fintech platform based on open banking and we were relying on the upcoming FedNow payment rails, which is going to be a Compliment to credit card payments.
In North America and Europe, we are so used to credit cards and the reward system that it's basically there. There was a lot of consumer friction. It has to consumers. Habits have to change, which is always going to be extremely hard. So yeah, it wasn't sustainable for us to so we closed it.
Jason Jacobs: And is it the same team that's after this new one?
Uday Chandra: No, not at all. So a little bit about my background. I call, I like to call myself a software builder, maker, been in this business for a decade and a half. I did my bachelor's and master's in computer science. So I just love building software. And I had a nice stint at Oracle Endeca, where we built search and things around search, from data ingestion to, Traditional machine learning and natural language processing and those kinds of things.
So I am going back to my roots with this new company, Singular, where we are leveraging all the latest advancements that have been happening over the past few years in artificial intelligence and leveraging that to build software with a spin. So obviously, SAS has been around for so long and, we're all used to subscribing to various.
Sass providers for various things that we want to do in order to conduct our business, right? So we Are challenging that status quo and saying why do you need so many tools to perform or operate your business? because of the latest Advancements that have been happening in AI can we leverage AI to build a bespoke?
So to speak a business operating system that can do it all or most of it. I'm not saying we'll replace Email providers, but for the rest of the stuff, like your task management, your chat, your specific business rules and all of that. Can we combine all of that into one single cohesive operating system?
And so that's the journey we are on. And so far, I'm happy that in the last three months since we started, we are at close to 45 K. M. R. We have a few customers who are really happy with what we're building. and yeah I believe that we are on the right track this time around bootstrapping ourselves along the way.
Jason Jacobs: And when you shut down the last one, were you immediately on to the next? Were you planning to take some time off? Because it sounds like you didn't waste any time in terms of getting right back at it. How did that come about and how did you lock in on this idea?
Uday Chandra: So we saw the writing on the wall with the U Key around November 20, 24 last year. And all along, even while we were building U Key in our FinTech platform, we were heavily focused on also incorporating machine learning for fraud fraud, for detecting fraud, and for optimizing.
Transactions, especially since we were looking at micro transactions, which are under a few bucks and those sorts of things. So we were my co founder Sumesh and I, we were always, very close to using AI, but not directly with Yuki. So when we saw the writing on the wall, we immediately started brainstorming on, okay, is there a path forward for Yuki?
If not, let's be honest with ourselves. It's going to be hard because we spent it. Two, two and a half years and it was hard, but we are also wanting to be more pragmatic if, we can't raise money and more importantly, if we cannot find the product market fit, then there is no point in wasting more time and trying to, continue to go with Yuki.
So once we decided, we took a week and a half off and immediately started brainstorming on how can we. Use all the learnings while we build Yuuki along with all the latest advancements that have been happening in AI. Combine those two and come up with a problem. And we weren't this time around searching for a problem to be solved.
We just, while we were operating Yuuki, found out that we had to spend a lot of money, getting all these various SAS tools and using multiple interfaces, subscriptions, and sometimes forgetting even if we are not using it. So the sass fatigue was real. It was a personal experience for us, and we said, Hey, hold on a minute.
There are too many sass tools out there. Every feature at this point is masquerading as a sass platform. And in fact, you don't have to, especially given the scale of most companies. I'm not talking about Google scale or meta scale. Of course, it's a different ballgame for some companies. You can't build everything yourself.
And Build versus hire that kind of a thing decision is now going to be challenged because I can help you deliver all of this. Like I said, business rules incorporated into one single platform. It's more easy to do that now. I'm not saying it's a solved problem, but we thought that was fascinating. And since we personally faced issues with Having to juggle multiple SAS tools, we said, Hey, let's take a stab at it.
And that's how we started Singular. And yeah, we didn't take much long to get on with shutting down U Key and starting Singular.
Jason Jacobs: Okay, and so you had the pain personally and you thought, Oh, this seems like an area that makes sense to attack. What came next? And how did you balance building versus talking to the market and and doing customer development, market research.
Uday Chandra: Yeah, again, we have to thank all the learnings from U Key over the last two and a half years. So we obviously spoke to many founders during that period. So we went back and spoke to all of them. Twenty, twenty five close founders of network that we had, we tapped
Jason Jacobs: So was Yuki was focused on startups.
Uday Chandra: Now, Yuki was focused on serving publishers to sell their one what do you call it? Articles in a piecemeal fashion that complemented subscriptions. So anytime you see a paywall, typically you try to find it somewhere else or you just abandon the conversion rate, as is not that great. So we said.
If you can also offer an additional thing where you, instead of subscribing to this publication, you can just unlock that specific article because if I'm interested in something that my friend shares on New York Times, I would just want to read that one article and move on, not necessarily subscribe. So that was the original idea of UKey.
And because of the transaction fees being so high with credit cards, we thought, yes, banking payment rails, which are very cheap, would be the right way to unlock that opportunity of micropayments. So we were serving publishers, content creators, and also stores, so we extended that to not only micropayments, but traditional payments using banking rails instead of credit cards.
So that was our market. Anyway no what was I saying? Yeah
Jason Jacobs: so then once you knew you had the personal pain, you said that you had talked to a number of founders. So were these founders of startups or founders of small businesses or what types of founders do you mean?
Uday Chandra: a mixture of both, I would say. Yes, some were just getting started. Some raised There's a fair bit of money in their, I would say, C or Series A stages. Some running small businesses or mid size enterprises too. And we asked them all the same question. What's going on with your usage of SAS? And some said, we don't even know how many tools we are using right now.
And there are SAS tools that offer you To manage the other SAS tools and the subscriptions, keeping track of all the subscriptions, how it's being used. And then, of course, for regulated spaces, you also need compliance and security. So there are SAS tools that offer that. So it's a huge market.
I have absolutely no problem with it. I'm not saying it was, it's a bad thing, but I think we are. Absolutely in a pivotal time where you can reimagine all of those things instead of saying, I need one SAS tool to do my task management, one SAS tool to communicate with my employees, one SAS tool to do my financial projections for accounting and so on and so forth.
There's so many things that you need to operate a business once you are past the initial stages, right? We asked those sorts of questions to a lot of these founders, and they said, yes, like I said, some even when they were at a, series a stage or these mid size enterprises, they didn't even have a clue.
Some of them as to how many SAS tools they were using and in a highly regulated environment. They were frustrated with, not being able to use the tools that they want immediately because they will have to go through a long process to get the legal approval. To incorporate a new SAS tools and so they were in this weird space where in their productivity state is going down because they need XYZ software, but they can't procure it immediately and we said, you know what, let's tackle this.
This seems like a good enough problem and it was previously impossible to do it. If you said to me five years ago, Hey, here's a company that can even. The business requirements describing your business will be able to put everything together because traditionally building software while on papers seems simple, like anyone can do it.
It's not that easy, right? There are so many nuances associated with building a production worthy software system. So it wasn't possible. But with the advent of everything that has been happening, I'm sure you know this. You've been tracking it. You've been talking about it, which is a I being able to generate code.
We said, yes, this is the right opportunity to embark on such a challenging task. And so far, we have been pretty happy with the results. And there's a long way to go for sure.
Jason Jacobs: It sounds like you were talking to a wide range of different businesses. How did you and how are you thinking about prioritization in terms of Customer segments if you're even thinking about it that way are there criteria that make a company a good target? Are you starting with a certain size of business, or type of business, or industry, or businesses that have certain pains, or businesses that are in a certain stage in terms of what types of tools they're using, or how they're using them?
Like, how do you think about focus?
Uday Chandra: Great question. Yeah. We obviously want to narrow down and focus our energies on serving a specific sector. But right now, in terms of size, we are targeting on companies that are at a seed stage where we believe that we can grow with them in terms of when you are at a seed stage or pre seed stage, you're not going to use 20 different SAS tools.
You're just getting started and you're trying to figure out what SAS tools you need to use to operate your business. So that's the right market for us. At this point, we also had some success with the health tech health care sector. Where we are, we have a few customers in the healthcare sector, mid enterprises, they're not too large, they're not too small.
And yeah, so right now that those are the two areas where we are finding success. Three months old again, we, I don't think we have figured out exactly which segment to focus on yet. But we have a good pipeline, so we're just trying to serve. These folks at the moment. But great question. Yeah, we are thinking about it.
Who? Where should we spend our energies on at this point in terms of size of the companies that we're serving? I would say precede and seed state companies.
Jason Jacobs: And how did you think about the initial set of features, functionalities platforms to support, et cetera? It just sounds so daunting to go from all these different tools within an organization to essentially one tool, and especially a new organization that's three months old with a tiny team that hasn't raised any outside funding.
There's a lot of trust, and it sounds or seems risk involved as well and how have you navigated that in terms of what you include and then how you position it to these companies?
Uday Chandra: Yeah, great question. Trust. That's the key word there. We demo our product. We have so The first one month, one and a half month, we have built a tool, which is basically a crawler that has done a pretty good job in terms of figuring out, given a segment, an area, what are the top five SaaS tools that dominate the market right now?
What are the five open source tools that offer similar functionality, because it's open source, you still need someone to host it and maintain it, those sorts of things. So we built this, and we are continuing to build this database where we have access to the top five SAS tools in a given segment.
And like I said, the open source alternatives to it. So we're continuing to build this database using a crawler. That's one of our agents. So Singular, we, when we started out, we said, okay. No one is going to give you or describe you their business. It's hard, right? Prompting is hard. And so how do you make it easy for someone to onboard?
And how do we demonstrate that we have this business operating system that will perform pretty well? Now, we're not talking about petascale, right? Most companies don't need extreme scale. That's the thing we tell them up front that, you know. We will maintain the business operating system for you. We will have a single unified data store.
So there is no moving data around for you to do your business. So one unified store, one unified interface, and we will provide you the functionality that we have already built. We call them software Lego blocks. So once you show them The functionality that we have already built as Lego pieces that can be put together.
Like I said, most companies do need a task management system, do need a communication system. And then on top of that, some customer management, some customer content management perhaps, some have learning management systems. So along the way we have built close to ten different modules, what I call software Lego blocks that, have been well tested.
We use So we're not claiming to, do some deep research in machine learning or AI, but we are pretty good at applying those concepts to build software. One thing that we do is what we call a deterministic engine. Basically saying if this software Lego block, this piece of functionality is, I would say around 30 percent coded by AI.
What's the guarantee that it actually does its job? So we have built an internal tool where it makes sure that anything that's generated by AI is still deterministic, right? Which means we have great unit test coverage. We have formal verification systems, basically proving that whatever has been built by the software actually performs well, does its job well.
So it's deterministic in nature, essentially. Using those techniques, we have built these Lego blocks. And I think over the next 18 months or so, we'll continue to build many more Lego blocks that are pretty common out there in the world. Now, when it comes to building the business rules that are very specific to an organization, that is where the challenge lies in terms of at least the technology that we have is not enough to fully automate it.
So we handhold when it comes to Building something very specific to a business like an example would be the recent customer that we have been building their I don't want to reveal too much information about them yet, but it's called a clinical intelligence system. So that's very specific to how they conduct their business.
So we gave them. A chat bot, so to speak, where they can describe some of the business logic that they want to incorporate into this business operating system. And then behind the scenes, my co founder and I will have to actually look at what the software what is the software that needs to be built?
How did our coding engine, how is it doing its job? What are the kinds of formal verification systems that we need to put in place to make sure that the business rules are fully incorporated, those sorts of things. So that's like much more challenging than, leveraging the software Lego blocks that we have already built.
So there are two aspects to it. One is we continue to crawl the web, identify these various common themes and build the software Lego blocks that can work. In a cohesive manner using a single data store, right? And the other one would be very specific to a business. That is where we are playing a key role at this point, but we are pretty confident that we'll continue to automate that aspect of it as well.
Where with the recent advancements, we heavily rely on open source and we love open source and everything that's happening in the open source world in terms of Bringing AI models up to speed with these massive proprietary AI models. That's one thing. And then all the tools around being able to incorporate these base models, fine tune them.
There's lots of techniques like distillation, right? Fine tuning all the latest advancements in reinforcement learning. That's happening right now with instead of using human in the loop, the techniques that DeepSeq has come up with recently over the last few months. All there. The noise was around how DeepSeek has basically upset everyone else with their large language models like OpenAI, right?
So instead of thinking about what's happening there, we are really incorporating the techniques that have already been built by these awesome open source community, leveraging them and, building our own DEXE model, which is to build customized business operating systems. specific to a business.
So that's our ultimate goal. I don't know how far we'll go with the current way of conducting our operations, because like I said, we still have to make sure that we read the business business criteria, business rules given to us by the company. And there are, there is sometimes pushback from them saying, I don't want to describe my business.
I just want to pick X, Y, and Z. Give me it. Okay. I want this functionality A, I want this functionality B, like I said, task management, project management, financial projections, those sorts of things. And then you build the software. That is why we are building our own crawler, our own software Lego blocks.
But at some point, there is some customization that has to be done. And that is where AI is not great. I would say 20 30 percent of it can be automated through AI. And the rest is something Sumesh and I would do also using AI. So we speed up the entire process, regardless of whether it's fully automated or barely automated.
Jason Jacobs: So do you have a SAS pricing model for your customers?
Uday Chandra: We do right now it's very specific on what they ask for us. But we are playing around with some options in terms of given. The number of features that a company chooses, we have put up a calculator on our website where you select the number of features that you would want the file storage space and there are a few other things that you can select and then it gives out a price and yeah, we're still experimenting with the pricing model, but we found, I would say we were very lucky because Just in three months.
We are able to get to a pretty reasonable monthly recurring revenue and We'll see how it goes when we open it up to the public right now. We are focusing on Identifying a specific business talking to them going through the typical b2b sales pipeline and we are at Close to 50k in MRR, but
Jason Jacobs: And how many customers is that, just ballpark?
Uday Chandra: Yes,
Jason Jacobs: I think you said 45k in, in MRR. Is there a services component too? Because it sounds like the customization varies a lot from client to client. How do you how do you account for that time and resource on your internal team?
Is it, is that priced as well? Or is that built in
Uday Chandra: yes.
Jason Jacobs: And that's one time, or that's just like consulting?
Uday Chandra: That's just like consulting. Yes.
Jason Jacobs: Huh.
Uday Chandra: Depends on the complexity of the changes that are required to really build that specific business rules. The rest of the stuff, it's a monthly recurring model, right? Yeah.
We don't want to do a whole lot
Jason Jacobs: Huh, and it sounds like you're using a bunch of open source products. How much of the functionality that the end customer is utilizing is stuff that you and your team are building and that's part of a singular offering versus stuff that the singular offering just sits around and provides a consistent layer across?
So how, like what percentage of the. Value prop of the overall offering, are these open source products doing versus your offering that kind of sounds like ties it all together?
Uday Chandra: like libraries and frameworks not. Entire products or even features, right? So most of it is like I've been referring to the software Lego blocks that Singular has built so
Jason Jacobs: open source is powering the Lego blocks, not powering the, it's not utilized directly by the end
Uday Chandra: Yeah, because it's hard once you start thinking about, okay, I'll just piece everything together that's already built. Then the whole goal of having one single unified data store and one unified interface. It's hard to maintain because of all the integrations, it's going to be a mess. So we said we will design the unified data store layer.
We will design the interface. We'll obviously have APIs. So when we are building the back end, yes, of course, we'll rely on open source frameworks and libraries. But yeah, most of it is has been built by us.
Jason Jacobs: There's lots of rumbling out there that because of AI, the SAS model as we know it is under significant threat. How do you feel about that and why?
Uday Chandra: Again, because I'm biased with Singular I don't think the world needs too many SAS tools. They do need one cohesive interface and an engine that can help them successfully run their business. So when you think about the current SAS model, which performed really well over the last two decades, it's about, okay, you, do you want to do a B testing?
Here's a SAS tool for that, right? Do you want to monitor the customer's activity? Here's a tool for that marketing. Here's a tool for that revenue management. Here's a tool for that. So there's so many different tools that at some point, there has been a lot of research done in this, what they call SAS chaos, too many SAS tools and those sorts of things.
At SAS.
Jason Jacobs: how I feel about AI tools, by the way, but that's another whole discussion.
Which, by the way, have a I guess they don't have a SAS model, right? It's more about usage, right? Tokens.
Uday Chandra: yes, yeah,
Jason Jacobs: Plus SASS. It's like a little bit of SASS and then a lot
Uday Chandra: I
Jason Jacobs: or, actually, I don't have it, I'm not using them at a scale enough that I know is it a little bit of SASS and a lot of token, or is it a lot of SASS and a little token.
For me, it's all SASS today, but it's just because I'm not doing anything industrial strength yet.
Uday Chandra: No, I agree. Everything is SAS, right? If you go online and you're just going to a web app, sign up and do something that SAS, of course.
Jason Jacobs: It's ironic that the same people that are saying that Sass is dead are backing all these AI tools that they say are going to kill Sass that have Sass models.
Uday Chandra: I know.
Jason Jacobs: on there?
Uday Chandra: I know. I. Yeah. That's. That's what I think too. I don't want to be using, I don't want to be forced to use too many SAS tools, whether it's built by AI or AI powered or AI generated or it's the traditional SAS tools that we know and love and have a love hate relationship with it. I think there is a nice middle ground there where sometimes you are absolutely You know, based on your needs, you will need certain kinds of SAS tools to use, like email providers.
They're not going away. The underlying infrastructure providers, they're not going away. But all the other wrappers that are, like I said, A, B testing, marketing, customer management, content management, media generation, all of these various things that you do and use, you don't need thousands of tools.
Of course, at this point in time, Especially right now in 2025, we are not at a stage where you can use one tool that does all of these things really well. But, just like Singular, I'm sure there are other companies that will pop up that's going to tackle this space, saying, you know what? Don't subscribe to tens and thousands of SaaS tools.
Here's one business operating engine that can do all of it for you. Most of it for you, at least, if not all.
Jason Jacobs: One of the earlier episodes I did was with Sahil Lavingia and one of the things that he said was that he thinks that Sass is going to go away and that in its place will be the tools will get so cheap and it'll get so easy. to build efficient and inexpensive, et cetera, and automated, right?
That that companies will just build their own personalized GloveFit software in house to serve their needs. What I'm hearing from you is it sounds like Singular is taking a hybrid position where you say yes, it's moving in that direction, but actually there's going to be someone like us that pulls together the open source stuff into Lego blocks and then wraps them.
Customization around it on a customer by customer basis. And today that's through consulting, but over time, hopefully AI will. We'll automate that too, but you'll still need someone like us to consolidate all the SASSs versus wanting to fully DIY it. Is that right?
Uday Chandra: Exactly right. You got it. That's my take on it. Yes, that's our stance. Larger enterprises, which do have an army of software engineers and ML practitioners, they can definitely build a lot of stuff, right? Like the build versus buy debate, we might go away fully when AI gets really good.
And I said, when I say AI, I mean to these Coding agents, AI software engineers, AI software engineering agents, those sorts of things, when they get really good at managing large chunks of code bases, yes, perhaps these large enterprises can do it on their own, but for the rest of the folks who do not have necessarily the resources and the time and the experience to do that on their own, We are going to serve them.
We're going to say, yeah, you still need one business operating system. We will build that for you. You don't have to go and, use 10 or 50 different SAS tools. Yeah, that's the happy medium for us.
Jason Jacobs: Now why has this not happened earlier? Because open source has been around for a long time and it sounds in your mind, AI is what's now making this possible. Is that right? Am I correct about that that's your view? And what is it about AI that's making it possible now where it wasn't possible before?
Uday Chandra: It's all about scale. With a small team so far at Singular, we are able to manage five, six customers being able to First of all, there is something that's continuously running in the background, which is the crawler that I mentioned that's gathering all this information about various functionalities and features out there.
The second
Jason Jacobs: And that's to inform the Lego blocks approach in terms of what order and prioritization and which ones to build, et cetera.
Uday Chandra: Yeah, that's.
Jason Jacobs: Let's say it's the crawlers running continuously.
Uday Chandra: into our other agent. We call that maestro, which is basically to build these software Lego blocks. So those two,
Jason Jacobs: the agent is actually building the blocks. You guys aren't building them yourselves.
Uday Chandra: the agent is building the blocks. It's pretty terrible at this point. So we go there, review it, refine it. We take a very highly opinionated approach to this, by the way. So we're not trying to compete with other AI software agents out there, which say, throw it any language, anything, it will try to do it. But that's not. What we are doing. We're taking a very opinionated approach in terms of the programming languages that we use in terms of the open source libraries that we use in terms of the U. I. That we're going to build. So by doing that, we are ensuring that the Lego blocks themselves are ready to go.
Mostly bug free. I wouldn't say there wouldn't be any bugs in software, whether it's written by AI or humans, but so that's our approach. That is what we're
Jason Jacobs: And what what percentage of the, once you get done factoring in like cleanup and any final coding to like last mile coding to get it Production ready, what percentage of the work is actually getting done by the agent, would you say?
Uday Chandra: I would say somewhere between 25 percent if it's pretty new in terms of functionality to 60 percent when, the AIs are really good at building the auth module, for instance, authentication, it's, it's pretty, Pretty solid in terms of doing that. Regardless, on
Jason Jacobs: Where might it not be as solid and why?
Uday Chandra: it might not be good enough to, first of all, handle edge cases. AI is very good. AI coding agents are pretty good at giving you the happy path, which is given an input, you get an output, everything looks great, but then users behavior is unpredictable, and there's nothing wrong in it. That's the way we all operate.
And it breaks. It doesn't think through all the edge cases. And that's the fundamental difference between being a really good software engineer versus being an average engineer. Is to, keep thinking about the edge cases. How would the user use the system? What could be wrong with the way that you think that the user will always provide a certain form of input.
Definitely will never be the case. So how do you manage all these additional edge cases? That is where formal automation and verification systems come into play. And that's what we are also using so that we know that if a special method, if a specific method or function is being built by AI, we want to verify that the functionality is correct.
And we use AI to build that automated verification system too. So that's why I'm going to say on average we are around 35 percent in terms of AI fully automating the software Lego block building. Sumesh and I, we review all of the code. Like I said, we have some templating.
Jason Jacobs: it, Uday, is it just the two of you on the team?
Uday Chandra: So far, yeah. Yeah.
Jason Jacobs: Are engineers by training?
Uday Chandra: Yes. Yeah. Both of us
Jason Jacobs: Dude, are you opinionated in terms of how you build with AI, the tools that you know what what is your your, what's in your toolbox for building and how you utilize those tools?
Is that consistent?
Uday Chandra: Yeah, we are very opinionated in how we use it, what kind of languages we use, how we want to design systems in general at a very high level.
Jason Jacobs: What are your tools of choice?
Uday Chandra: In terms of programming languages, that would be Java, Rust, and JavaScript for the front end, specifically back end Java and Rust. We have opinionated approaches on the open source. Libraries and tools available in these specific languages, right? For instance, we don't want to add additional complexity or abstraction layers.
Open source is really great. I love it. But they also offer too many abstractions, one on top of the other. That makes it hard to read, manage in the long run, right? So we reject that idea. We reject the notion of Building abstractions because it's just more complexity. Now, if we are to truly believe in the promise of AI, then you should also reject it.
Why do you want to build these abstractions when AI can, in fact, understand and build code? So simplify everything. Do not add too many abstractions. So that's the fundamental opinionated approach that we are taking.
Jason Jacobs: You know what's funny is that this conversation started and you were like really soft spoken and quiet and I feel like the longer we're spending together, it's almost like the beast inside of you is starting to slowly emerge.
Uday Chandra: I do have some strong opinions when it comes to mixed software.
Jason Jacobs: Oh, it's great. It's great. It's it's you have to for what you're doing because maybe not now when you're three months in with five customers or whatever, which, don't get me wrong. That's amazing. But that's not going to scare SAP, right?
But it's fast, fast forward a few seasons of momentum, right? And and you guys start to be a pretty big threat pretty quickly.
Uday Chandra: Yeah. I hope so. We're working towards it. Yeah.
Jason Jacobs: how did you get these initial customers? What did that go to market look like? Was it cold calling? Was it yeah, like what was the approach? What is the approach? It's just the two of you, and there's a lot of code to write, and AI is, doing between 25 and 60 percent of it, but that still leaves a lot of code and customization to do.
Uday Chandra: absolutely. A lot of late nighters. Obviously, we still believe that with a small team, we will be able to achieve our goals that we set for ourselves over the next two to three years. At this point, we're doing a lot of hard work. But in terms of acquiring customers, I Again, luckily, because we have spent two and a half years on this failed attempt, we do have some network that we can tap into.
And so it has all been, so far, all the five customers that we have acquired is WOM intros, or people who we already know who are working at these mid enterprises.
Jason Jacobs: And what's the typical entry point? What function do you, have you found the most success with?
Uday Chandra: VP of Engineering, or COOs, CEOs even. Yeah.
Jason Jacobs: And what's the pitch?
Uday Chandra: The pitch is basically ask them what are the pain points that they are having in terms of managing 20 or 50 SAS tools. So far, all the five customers, they are at that we have onboarded, the average is 20 SAS tools that they are using. And the one mid enterprise we are working with, they don't even know how many they are using. That's the pitch.
Jason Jacobs: And what's their answer? Where's the pain?
Uday Chandra: To be honest, a few said it's too good to be true. I don't believe you. So we had a, we had to show them the demo. We had to have multiple follow up calls. We have to prove ourselves our credentials and how we're building it. We had to show them that, because again, luckily with Yuki, because it's in the payment space, we dealt with a lot of banking regulations, compliance, security, like a crazy amount of it.
So we already know some of the things that. Enterprises look for when they are evaluating a vendor. So we check a lot of boxes. Luckily for us, we have already gone through that process for the last two and a half years. So we know security and compliance and how we need to build systems that already adhere to these compliance and security things.
So that helped them. And to get, for instance, SOC 2 licensing or PCI DSS. Compliance and those sorts of things. What are the bazillion things that you need to do to be in compliance with those kinds of regulations? We have answers to that. So we continuously talk to them, show them, prove that we have all these things.
And once we get a few more of these mid enterprises, I think the ball will continue to roll. That's the
Jason Jacobs: I would imagine that even if it does Do what it says it's going to do, it would have ripple effects because it would affect like all of their business processes across the whole organization and then there'd be a training component with new tools and new logins and and it, it just sounds like a.
Uday Chandra: Yeah.
Jason Jacobs: nightmare. So how, like how does that, is that true? And if so, how do you accommodate for that? And if it's not true, how is it not true if you're swapping out all these SAS tools?
Uday Chandra: Absolutely. I think it would be a nightmare for the large enterprise because I have worked at large enterprises. And I know how hard it is to get new tools in and also how hard is it to remove certain tools that people are used to. Even if they're clunky, it's just a, you just get used to it.
So there will be pushback from users when you say, reject all these tools, just use us. But that's why our focus has always been with the companies that are at a pre seed stage or small mid enterprises who are okay to experiment, who are ready and willing to take on new challenges when they realize that will really benefit them in the long run in terms of streamlining their businesses, opera, optimizing their businesses.
So that's the challenge for Sumesh and I to really convince them, the C suite, that, hey, I know you have used XYZ tools before, and if you are happy with it, then we are gone. We can't do anything. But so far, we are able to convince these folks to say, one unified system is better than 10 different things, not just, and we are not competing around just the cost benefit of it, right?
If you're paying 3500 bucks for 10 tools versus paying us 1500 bucks for one tool, that's definitely a good thing. But we don't want to compete on just that because others can, decrease the price. And what will we do with that? So we have to offer more than just reducing the price.
Jason Jacobs: And what else do you offer?
Uday Chandra: The unified store. So less data integrations. More time to make decisions versus, just move moving data around to different tools. For instance, you have a customer management system and then you have a revenue projection system or a subscription management system, a payment system. All of these have to be integrated to, get make sense of it.
So there's another tool for that. There are reporting tools, business intelligence tools, data warehouses. There are lots of things and singular speeches that Not only do you get a unified interface in the UI layer, it's behind the scenes at the back and there is one database, one data store. So it's really easy if done right and executed right to be able to use a tool and spend less time wrangling the data and moving data around.
Jason Jacobs: The customers that you work with so far. Have they ripped out what you're meant to replace?
Uday Chandra: Yeah, a few, three, three of the five customers have replaced. task management systems. Some used Basecamp, some used Linear. The one bigger enterprise using JIRA they are still looking into it, but they have already started using a data issue management system that's built by Singular, which is similar to JIRA in some aspects, but very specific to their business in the healthcare sector, right?
Yeah they are very happy with Being able to use a single tool for that. And the other thing that I want to mention is that not only is Singular using AI internally, but we also offer it as a feature inside the business operating system. So for instance, for a healthcare company where they are doing these clinical trials and stuff, to be able to assess that data to bring forth certain Issues that they find in the data, those sorts of things.
That's a very lengthy, time consuming process that they have right now. And we have incorporated these base models, integrated them inside this business operating engine, and they are using that. They are chatting with their own data using AI. That's the other big thing that we are offering. And we showed them.
We can not only just integrate with existing models out there, but because this is a very highly regulated industry, we are also given them the option of hosting their own models on their own cloud provider like Azure or GCP and further fine tune them. So just if you think plug and play will get you far.
No, not at this time. You still need fine tuning. You still need that iterative cycle. To leverage a I fully and so singular has also built tools just like there are some open source tools and proprietary tools out there for, tracing for prompt engineering, for prompt management, for evaluating the prompts, for evaluating models, those sorts of things, even that tool is part of our business operating engine.
So you see, that's where we're getting it. It's a massive project. It's a I would say very ambitious, but we are able to execute on our vision so far, and it's not easy by any stretch of the imagination, but we're getting there. So that's the other big thing that we offer. Not only do we use AI for ourselves, but we provide that within the business operating system.
So they have been really happy with that because most of them do not have the engineering team, the resources needed to do those sorts of things themselves. And unless it is about, dealing with content, you cannot simply rely on ChatGPT or just the great open source, not open source, but these proprietary models like, be it Gemini or OpenAI or Anthropic and others where you go there, you chat, you have APIs to integrate with.
That can only take you so far, but if you're dealing in a, dealing with your own data, especially in a very highly regulated environment, you need a lot more pieces to it. And that's something that we offer as part of the business operating system. And that's part of our Lego blocks.
Jason Jacobs: So in, in terms of pilots, is the strategy to have it run alongside the existing tools for some period of time? Or do you aim to get them to swap out one that's maybe lower stakes, lower risk? Or, easier to pull out or put back in? Like how you Get over the hump of getting them to feel comfortable to start somewhere.
Uday Chandra: Yeah, absolutely. Like you said, low hanging fruit start with something low risk. And then once you start getting used to the system and see how it can know, optimize your day to day operations and workflows, then you can slowly nudge them to continue to use it. But we still have to solve this problem of migrating the data.
If it's a large enterprise that already has been using these tools, SAS tools, how do we migrate all that data into our unified data store, right? Again, our Lego blocks and our data integration connections will help, but our whole promise is to get away from using data integrations, connections, and all of that.
So we're trying to figure out what's the happy medium there. Again, for large enterprises, we're not going to tackle them at all at this point. That would be the challenge when we, hopefully at some point when we get these larger, bigger B2B enterprises where this will definitely be the number one thing.
How can we use your tool when we have 100 different SAS tools with different data stores? How do you pull that all into our into your business operating system? That's a question that anyone would have so we're not We haven't solved that yet. I don't know the answers for it. I do Know that AI will definitely help speed up all of those things in terms of solving those kinds of challenges, but at this point our focus is on Small mid enterprises and series a seed companies, which do not have that kind of a challenge.
Jason Jacobs: And you're the second person in two days that mentioned to me that they've built agents internally to write code. Is that a common thing? Are there a lot of people doing that? Did you build your own agents or using? Off the shelf tools and how much work was it to actually invest the time up front to get the agents doing the code?
Would it have been more what's the payback period look like where it becomes worth it to invest in that versus just writing the code yourself?
Uday Chandra: Great question. I think it all depends on your own experience as a software engineer. In summation, I, we are pretty lucky to have had this great decade plus each in terms of Building large scale B2B and B2C enterprise software, full stack, up and down the stack using traditional machine learning models at the same time, data pipelines, data ingestion, data cleaning, building these inverted indexes, natural language processing.
So having dealt with those kinds of things over 15 years or so, I believe. We have that, it's art and science at the same time. So we do know how to leverage build these AI agents that can build, like I said, 20 to 60 percent depending on what kind of things that you're building, but it's definitely an investment and it's a continuous investment.
You can't be like, yeah, write this code and not, no, you can't be like, I will not look at it. I know I can trust you. We are not yet, we are not there yet. But I think in the next two years, we will continue to make significant progress where it's mostly automated and at that point, since you have this head start of having these agents continuously running, not just the code, but the tests to validate the code, the formal verification systems that you build in place.
All of those. And even while you're sleeping, it's crawling the web, finding these functionalities, it will make a lot of mistakes. Like I said, it's not even close to where we want it to be. And our vision is to continue to have this very small team, but still be able to deliver big on in terms of building these Lego blocks, putting them together and serving the needs of our customers.
Jason Jacobs: And is that with a bigger and bigger agent army? Is that how you see it?
Uday Chandra: Yeah, lots of them. Lots of them there's this concept that's been around for some time, mixture of experts, right? That's what we want to do. We want to have a, an army, a cluster of AI agents working together with us while we still call the shots. At least for the next two, three years. Where we are Reviewing the code.
We are following all the best practices. We have also built a solid foundation to begin with, right? We have used our expertise. Like I said, I'm very opinionated in terms of what we want to use, how we want to use, how we want to build things. So that's all already built into these agents that work for us.
Jason Jacobs: What about go to market? Are you going to have a sales force? Is it just going to be word of mouth? Are you going to build agents to do the outreach on your behalf on that side too? Have you thought about any of that?
Uday Chandra: We haven't honestly thought about that yet, because we're still in the initial pieces. A word of mouth and warm intros
Jason Jacobs: And you're like you're constrained on how much you can deliver, so you don't want too much too fast in terms of customer demand, right?
Uday Chandra: Absolutely. I would say we are lucky enough that we have our hands full at this point. But that's a great question. Can we leverage AI agents to do some of the marketing, some of the cold outreach for us? Although
Jason Jacobs: If you do both sides, all of a sudden you start looking a lot like Sam Altman's prediction of of the sole founder billion dollar company, except in your case it'll be two founders.
Uday Chandra: Yeah. Perhaps you can't get away with just one person doing all of that. Yeah, I see small teams being able to deliver big using and leveraging AI.
Jason Jacobs: Huh. Are you feeling any pain that you think you'll need some additional humans to address, or are you going to try to take it with the two of you as long as you can?
Uday Chandra: The goal is to keep going with just with the two of us. Once we hit this bottleneck of we can't scale anymore with just the two of us, then we'll definitely look at hiring more people. It's not like we don't want to hire people. We're just bootstrapping ourselves. So that's all we have. And we want to see that MRR climb every month and at some point we'll probably reach our threshold of serving all our customers, doing a bazillion things.
As you very well know, being a founder, you have to juggle a lot of things, prioritize, reprioritize those sorts of things. So we'll see how far we can go with just the two of us. We continuously update ourselves on what's going on. Out there in the AI world that can help us to speed up our own internal processes to serve our customers.
And the biggest problem is to just leave out all the noise out there because every day there is a big claim about something happening, groundbreaking that's going to, make everyone lose their jobs and those sorts of things. And we know that to find out what it is that technology can offer.
Jason Jacobs: How ambitious are you feeling about this? In your wildest dreams, if you look out three years, five years, ten years what does success look like in its fullest form?
Uday Chandra: For us, it's a hundred million in ARR. When we hit that, I think that would be a great start. We want to be a unicorn. Not just for the sake of being a unicorn or chasing something, but at the same time, enjoying the process, delivering our promise of a unified business operating system to hundreds of customers.
And in the process, we make lots of money. Hell yes, I'm for that. It's worth the pain.
Jason Jacobs: And I know you don't You need to raise money, but if you keep on the path you're on, you would likely have access to capital on some pretty attractive terms that would put smart people around the table helping you and enable you and give you the resources to hire a lot faster, although it sounds like hiring is not even in your plan, but but there is a path where you Get the money, and then you invest in tooling, invest in infrastructure, and invest in help, invest in, go to market, invest in brand, invest in whatever.
Maybe it's not people, but there's plenty of things to invest in that can help you get there faster. Do you think about that at all? Is it even on the table for consideration and and if not, why not? And that's not to lead and say that it should. I'm just curious how you think about it.
Yeah.
Uday Chandra: No, obviously capital would help scale faster. There is no doubt about it. But we want to be careful in terms of when do we really need capital and raise it. When we really don't need capital, that way you get better terms. So I'm thinking as a founder myself, I want to make sure that I raise capital when I really need it.
And I don't give away equity just like that. But I'm sure that we would need capital in the near future where we would want to raise some money. And it all depends on why we need that money for. What are we trying to achieve with it? Yes it will really help when we raise money to hire some really smart folks out there within our network that we know of.
We do want some of them. We do want to hire them and, they're pretty expensive. Long as we can continue to bootstrap and continue to generate revenue, we want to go on that path.
Jason Jacobs: Okay, so given how much leverage these tools are giving you, and these tools are on their own path of accelerated improvement that's the same leverage that it's giving to all of your competitors that exist or don't exist yet, and the more success that you have, the more you get a target on your head of other people that are going to copycat and just try to replicate your approach.
So how do you think about defensibility and competition?
Uday Chandra: I think for us at this point, it's our Lego blocks. It's not easy. It takes time to build these Lego blocks, especially when you're just not weaving existing open source tools. Wherein, again, you are just at the point of having various data stores. UI on top of it, and then you have these massive data pipelines to move data around.
You can do that probably easily, but to build this crawler to gather this database where you understand various functionalities that's one big agent that we are very happy with in terms of what we are collecting as intelligence that feeds into our maestro engine that builds these Lego blocks again.
We have a, yeah, it's just been three months, but as we build and as we see, and as we review the code that's being generated, as we. See how hard it is to build these Lego blocks. We're not too concerned at this point. Yes, it's going to get easier and easier. Yet there is a lot of software engineering that needs to happen.
It's not as easy as it sounds to be.
Jason Jacobs: How do you think about IP, if at all?
Uday Chandra: I think our IP is again in addition to our expertise in how we build these AI agents, the agents that we have already built and the Lego blocks that all work in a
Jason Jacobs: huh. How do you think about it in a formal way, in terms of legal protection?
Uday Chandra: Good question. We have, we leverage open source AI models, but on top of that, we have done a lot of tooling. And we are very proud of what we have built. That's perhaps the best IP that we have. It's helps us like with the formal verification of the systems. That's not easy to build, but in underlying base models that is fully open source, everything around it.
That's our IP.
Jason Jacobs: I don't know much about open source, and I'm really interested in continuing to learn more about it. That's my caveat, since I'm probably asking a beginner question. But given the rate that AI is improving is open source Keeping pace and and what is the motivation within the open source community?
Because with AI, it's clear it's like a, it's like a, there's like a glass chewing capitalist motivation. But what is it that keeps the open source fire burning so bright? If it is burning so bright, I'm not close enough to it to know.
Uday Chandra: I think it certainly is doing a phenomenal job in terms of catching up with every open proprietary model that OpenAI is releasing. It's doing a phenomenal job. The motivations behind I'm not entirely sure. Obviously, there are some well funded open source initiatives. Mira is doing that, for instance, with Lama class of models. I guess biggest thing is that they don't really have a mode and they just want to say, okay, we'll just give it for free since we don't have the cloud infrastructure like AWS as GCP has, we have nothing to lose. Let's just open source it so that the open source community can further improve on it.
That's meta stake on it. And then there are these large data sets that have been refined to it. Start building these base models by some folks, again, funded, some not so well funded. But in terms of the hunger and the desire to compete with the best model, proprietary models out there, there is a lot of fire in there.
There is also funding for sure. Otherwise, it's not easy to build these base models at least. So open
Jason Jacobs: it's for, this is
Uday Chandra: Sorry, what's that?
Jason Jacobs: for profit funding profit seeking funding. It's not like philanthropic capital or things like that.
Uday Chandra: Yeah, I think so. I think so. There are two different different types of funding happening in the open source world. One is just we have got nothing to lose. Let's just keep releasing and leverage the community to make it even better. And then there are companies that are like, okay, let's just open source the base model and then let them come to us.
In terms of deploying these models, fine tuning, distilling, everything that's required to actually operationalize these models. Of course, there's money to be made there, right? Open source is all about the free components. That's the core. But then to really use it, you still need people, you still need infrastructure, you still can make money out of open source.
So I think that's what's driving the continuous innovation, not just in AI, but even before that, it has always been the case that, you release a search engine for free. But to really use it in your e commerce engine, you would need engineering resources, you would need infrastructure, you would need maintenance support, all of that.
Jason Jacobs: And I know this is a core part of your thesis, but if you take a step back and just look at the market more generally, as actually, before I go there what do you think happens to the big SaaS players of today, tomorrow, and beyond?
Uday Chandra: I think those who kind of pivot or embrace AI fully and reinvent themselves, and some of them are, In terms of not doing just one thing, but more than one thing. And again, for instance, I would think of Stripe. Not just payments that they are offering. A bazillion other things. So in a way, they are unifying lots of things for the businesses.
Those kinds of companies will continue to flourish. But those who don't, or think they are, they have enough moat. Yes, distribution is a big thing, of course, at this point, but, who knows what happens in three years if you continue to just rely on, I have the distribution, I don't have to innovate.
I don't, those kinds of companies will have a hard time.
Jason Jacobs: As AI continues to evolve how dependent Is someone like you on open source? Does that dependence go down as the actual AI tools keep getting better? Or is the open source going to be a big fundamental piece of the equation for as far as the eye can see?
Uday Chandra: Reliance on open source is not going away anytime soon. It's because the amount of tech that you get without spending, Your time or your money is massive, like really massive for specifically in terms of singular. Unless we start building the base models ourselves and invest heavily on that research area, we're not going to not rely on open source.
Yeah, I think for us, at least as far as I can see, will continue to rely on open source unless the innovation for some reason dies, and I don't see that happening at all.
Jason Jacobs: Is it, is it, am I understanding right that AI is essentially the catalyst that will help open source have its moment? In your view?
Uday Chandra: I think it has already had, and it will continue to have like software in general, cloud computing in general, everything happened because of open source. There's no doubt about it. So yeah, even AI with AI, it's the same thing. I wouldn't say AI was the only thing or is the only thing that's making it have its moment.
It has had its moment for the last two decades.
Jason Jacobs: I think, that makes sense, and that's just my own lack of understanding of open source that I need to get. Smarter on, but I think the thing I was, but like my follow up would have been if you had said yes to that would have been, but you said that AI isn't great at doing the customization yet, and so if it's the customization that's the glue that makes it possible for open source, to be good enough, right?
Then that's still human and not AI powered, but but you answered that that it isn't, that open source is already there, and it's just, it's like amplifying it, but more incrementally than putting it on the map.
Uday Chandra: Exactly. Yep.
Jason Jacobs: Yeah, cool. If I want to get tomorrow an open source, what should I read?
Who should I talk to? What, what organizations should I check out? Any, and either answer that now or that'd be a helpful follow up for me if you don't mind.
Uday Chandra: Who would follow? I think Andrew Ng. I love the way he presents the concepts. No marketing, no buzzwords. He just says facts. And I love the way he articulates everything that's happening in the machine learning and AI world. So I really love what he does and How he contributes to the community by teaching everyone machine learning AI and the various concepts.
That's a good start.
Jason Jacobs: What did you have any inspirations or role models when it came to this Lego blocks concept?
Uday Chandra: Inspiration was from software engineering. It's the same concept. You always have these various. I think I should go back to the Unix philosophy of doing one thing and doing it right. So all these various things The command line tools that Unix has built is how everyone thinks about building software.
Do one thing, one thing which is your framework, which is your library, and you compile those pieces together to build software. We're taking it to the next level of abstraction, calling it a software like block.
Jason Jacobs: Do you have peers that are building companies with similar philosophies in other areas that are not competitive to you?
Uday Chandra: I'm not sure, I'm sure there are, but we haven't done the research
Jason Jacobs: So you guys are mostly just heads down. It's not like you've got a tribe of people that you're banging, comparing notes with and building shoulder to shoulder with.
Uday Chandra: I'm sure it
Jason Jacobs: just insular, it sounds
Uday Chandra: at this point, yeah we did come across a few companies recently that said similar things to what we are saying. We have the Lego blocks, you just tell us what what you want and we'll supply it. But it's more like a procurement of various open source tools. That concept is slowly being embraced by everyone.
Jason Jacobs: Would having a more active peer group be helpful to you?
Uday Chandra: Absolutely. Yeah. Yeah.
Jason Jacobs: of the things I'm thinking about because I'm bringing on so many people like you coming on the show that are like neck deep in this stuff at the bleeding edge, hitting it from different angles, uncovering a lot of learnings and Lego blocks, for example, that so it's not just the learnings.
It's actually the tooling that could be useful to other companies that aren't competitive. And so do I have a role? Does the next have a role in pulling together the people and the Lego blocks in a way that could benefit it? The ecosystem, and if so, what does that look like? Because we don't want to violate anyone's confidence.
We don't want anyone to share anything they don't want to share, but it's like in order to get, you need to give, right? And so how do we, for people that opt in like. How do we build that tribe and how do we organize it? How do we manage it? Scott Weller, I had on the show the other day, you should listen to that one when it publishes because he, in fact, you guys should probably talk.
I'm going to tell him about this episode because, I think there's a lot of overlap there. They're automating. The process of commercial lenders evaluating potential debt recipients and a lot of the mundane tasks, and they're building agents to do that, but then they're also building internal agents to write a bunch of code, right?
But but the, yeah but the point is that he was saying the same thing. He's I want to share, but I'm busy, and who do I, and I want to learn, but I don't know who to talk to, who's doing what and I don't know, do you have any thoughts on what would be most helpful, and how we might structure something that would actually, you know, that you'd actually be excited to engage with?
Uday Chandra: I think that's a great idea. I would love to be part of a community where we can exchange ideas and, talk to each other, learn from each other and the next playing a key role in that,
Jason Jacobs: But where would that live? Is that a WhatsApp thing? Is that a Slack thing? Is that a Discord thing? Like in Scott's mind, he was like, you should build agents to, to make those connections. But I don't know. What do you think?
Uday Chandra: I think you should start off as a community. I don't know. Yeah,
Jason Jacobs: Feels tired, doesn't it? Slack and Discord feel just, like I don't want to go in there.
Uday Chandra: Yeah, I'm not. Personally, I'm not a fan of Discord, but Slack works. You can, jump on asynchronously and communicate in person would help. Some group discussions like this where more than one, two folks, round table things, those kinds of things. It really would help where, the next, next can make those connections happen and organize those round tables would be great.
Yeah, I would love that to be able to talk to. People who have been doing these kinds of stuff. Yeah. Yeah.
Jason Jacobs: If we actually have the beginnings of something to react to I'll come to you for some. Feedback then because we don't, we don't want to do it and we is just me. Like you, I have a tiny team, but I have a team that's half the size of your team.
But but I think that, we as the next, next, I think. Don't want to do, we don't want to do it just to do it. We don't, we'd rather do less things great, and only do great things. And and there's a lot of opportunity. But if we're not going to do it right, we'd rather not do it at all.
And if it's not something people are excited about and digging into and getting great value from, eh, not necessarily initially, but if we don't see the path to get to that point, then why do it?
Uday Chandra: Yeah. No, I think you're on to something. And I really wish you good luck and success. I'll be following your newsletter. I'll be watching your podcasts. And yeah, I'm rooting for you. Yeah.
Jason Jacobs: a little slow, and then as it went on, I was like sitting further and further up in my seat, and now I'm like ready to run through walls. I'm so excited about what you're doing. Any, anyone you want to hear from? Any any homework you want to give people?
Stuff to think about? Party words? I don't know, like you, how do you want to close this out?
Uday Chandra: I would say everyone should start using are thinking about how to leverage LLMs in their day to day lives. It's it play, it really levels the playing field. Democratizing has been thrown around a lot over the last decade, but I think AI and LLMs in particular are really going to level the playing field.
What you do with it and how you take it is on you. It's still not replacing you, but it will help you. Get a headstart. So everyone would, should figure out how to use LLMs in whatever they do. Do not ignore it.
Jason Jacobs: Uday, that's a great point to end on. Thanks for coming on the show and yeah, amazing what you've been able to do so far, and I can't wait to watch your progress.
Uday Chandra: Thanks, Jason. Likewise, it was great chatting with you.
Jason Jacobs: Thank you for tuning into The Next Next, if you enjoyed it, you can subscribe from your favorite podcast player. In addition to the podcast, which typically publishes weekly, there's also a weekly newsletter on sub stack at the next dot sub stack. com. That's essentially for weekly accountability of the ground.
I'm covering areas I'm tackling next and where I could use some help as well. And it's a great area to foster discussion and dialogue around the topics that we cover on the show. Thanks for tuning in. See you next week!