Reimagining software development: Guy Podjarny's AI Vision

0:01Senior people will have opinions and you want them to have opinions, you want to give them independence, but you do need to continue steering the ship. And so there's just a lot that you need to invest in getting them bought in, in simplifying core principles in a way that is not that dissimilar to what you need to do when you get out there and you communicate the product. Welcome to the SaaS Revolution Show, a podcast by SaaS.com. Here we interview SaaS founders from around the world who've been there and done that, as they share the ins and outs of how they built their businesses, their operations, their path for securing investment, and more.

0:38Our mission with the podcast is to help you, the founder, learn how to scale your SaaS, maintain your well-being, and navigate the complexities of this ever-changing industry. I'm your host, Alex Deema, and together we'll explore the good, the bad, and the ugly in the journey to SaaS success. ready to connect directly with AI and SaaS decision makers who need solutions like yours this October SaaS Stock Europe is introducing hosted meetings the performance-based sponsorship that puts you face to face with qualified SaaS founders and c-suite executives buy packages of pre-scheduled 15-minute conversations with buyers who've opted in to meeting you Every conversation matters.

1:21Every connection counts. And the best part is you only pay for meetings that actually happen. It's event sponsorship built for results. Connect with the decision makers who matter. Secure your hosted meetings at sasstock-europe.com slash hosted meetings. Limited spots, real connections, pay for performance. Okay, welcome back to the SaaS Revolution Show. I'm your host, Alex Thumer, CEO, founder of SaaStock and also general partner of Back Future Ventures. Delighted to be joined today by Guy Pajani, who's the CEO of TESOL. How are you doing, Guy? I'm doing well. Thanks for having me. Yeah, good to have you on the podcast.

2:01And this is ahead of you coming to Dublin in October to speak at SassDoc Europe, where AI meets ARR. Exciting times this year with everything that's happening in AI. And also, you know, you've seen this and this is also, I guess, kind of influence in terms of what you're building now with Tesla. And we'll get into that. But Guy, we'd like to start getting to know our guests a little bit more on a personal level. so I would love to know who is Guy Pajani? Sure, so I'm a geek, a nerd, kind of a computer lover from early days when computers were not as common. Software developer by trade, worked in sort of an AppSec company in Israel as sort of the first commercial job of it, moved to Canada during which, with an acquisition of that company, which got acquired by IBM and sort of during that time become product.

3:01So I'm sort of a direct developer turned product manager. And I was wondering how would I become a CTO? And I asked someone and they told me that the best way to become a CTO is to found a company and call yourself CTO. So I did that. I left IBM and I founded a company called Blaze that made websites faster in the DevOps space. That was a quick successful run, sold it to Akamai where I became CTO of about half of Akamai. and with Akamai I moved to the UK in 2014 and so I spent about a decade in Canada and 2014 moved to the UK and after about a year I got the age to to do it again so I did three and a half years good years at Akamai and learned a ton you know during that time but felt like I want to get back to the roots and try to go bigger and so I founded Sneak and Sneak was probably like first line of the obituary at the moment, you know, the claim to fame, it's gone well.

3:56It sort of wandered the desert for about sort of two, two and a half years until it kind of hit the mark to help developers secure the software that they're building and build that up. At the peak, it was at eight and a half billion dollar valuation. It sort of went down as the market corrected, but still now a sort of 300 plus million ARR business with over a thousand people kicking along well. And then kind of the last and latest chapter is about a year and a half ago, I left Snyk, where I'm still involved and still supportive, but I'm no longer operating there, to found TESOL, where we're looking to reimagine software development for the AI era.

4:35So obviously, you built this unicorn business in Snyk with, as you said, quite a significant valuation. Why leave when you left to found TESOL? It wasn't an easy decision. I think it started from just sort of searching, general soul searching about what it is that I want to do at Snyk. I think the good news about Snyk was that it was set up to be a long-lived company that has escape velocity. And that does imply it doesn't have an end point in time. People talk about the IPO. The IPO is just a financial liquidity moment. The company actually continues. uh i brought in this amazing ceo about five years in uh and i kind of co-ran the company with him for about three more years after and then i wasn't entirely sure what is my role in the company and i was soul searching and did some part-time and then uh i worked on sneaks ai strategy trying to say okay sneaks what should sneak do and adapt to this world and you know sneak secures software development and so to figure out sneak's future you have to kind of form a thesis about what is the future of uh of uh software development in the era of ai and as that picture kind of formed in my mind i got more and more enamored with it um and i guess kind of the last step was accepting that you know i'm an addict and i uh i want to do it again that what i like doing that what i am good at doing and gives me satisfaction is to you know solve big problems with technology and and so that's what I want to do and it was a nice having a comfortable setting for about a year there but at the end of the day you know satisfaction comes out of struggle you know and you need to be in a place where you're trying to do something that is hard for you and meaningful if you succeed and when you get that done you get that satisfaction and I sort of set out to do that And what is TESL?

6:33Can you share a little bit about what that is and maybe some information like data points about the company? Sure. So TESL is looking to reimagine software development for the AI era. We believe that software development will evolve or transform really from being code-centric to being spec-centric. So we think today you get some requirements, you implement something, and then you throw away the requirements and the implementation is the only thing that goes on. And over time, it forgets some things that were the intent and it includes some implementation decisions that are not actually important, but they get sort of cemented in.

7:12And that's kind of the trajectory of every software project. And we think LLMs offer the opportunity to change that up and to actually make kind of this sort of document of intent, this spec specification that captures what you want to build with some definitions of guardrails, of what it means to have built it correctly. That becomes the canonical document that defines your software. And as you expand to say, well, how do you build in this fashion, right? What is the spec? How do you create it? How do you debug a problem when it happens? How do you observe a system? and how do you collaborate, how do you version software or use it, you really evolve that into a new development paradigm.

7:50And we're looking to kind of create that with the initial entry point really being the people that feel this pain the most today, which are developers using a Gentic dev. So they're already feeling it. You give some intent, some prompt to the agent. The agent builds something, and then you come back the next day, and you tell it to build another feature, and it would burn the house down to get that feature done. and it would delete this feature he said yesterday. It's like, no, no, no, that was important, but it doesn't know. It looks at the code and it doesn't know what in the code needs preserving and what doesn't.

8:24And so that is kind of our initial entry point. But we think the real opportunity is really to build a new software development platform, a new ecosystem. We can talk about that more that revolves around this sort of new way of developing software. I guess maybe some vital signs. So the company was officially incorporated a little less than a year and a half ago, around March of 2024. We are about 32 people now. We raised$125 million at a 750 post in a seed round and then an A round that happened last November, in November 24. and uh and yeah and we're sort of building uh building the product we've been in uh in the growing private beta uh for a while now gearing up towards a launch uh very soon it's a pretty healthy seed in series a we're probably getting a little bit more used to it in in some respects uh in this era that that we're living in uh at the moment i think uh was it lovable series a was around 200 million, right?

9:38And yeah, it's a bit of a gold rush that is happening. A lot of excitement, certainly, well, both in the VC community and in the software sort of founder community. Like, what are your thoughts around this? I mean, you saw it when you were at Snyk, thinking about your AI strategy. and now as you're coming into Tesla and building this, you know, AI native software development of the impact of AI, this AI era on, you know, onto software, into the software industry? Yeah, I think there are several reasons why these kind of valuations explode and sort of the opportunity arises. And, you know, a lot of it is just that there's a limitless amount of value to be created, especially in software development.

10:32Software really creates, brings ideas to life. And as long as we have ideas, there will be more work to do. And generally, LLMs and specifically spec-centric software really unlock more and more people to create software. And so I think the opportunity is vast. There's no TAM question, right? The TAM is infinite. and the pace of change is so high that the importance of having the right people build is more critical than ever. It's already true in the VC world and the experienced investor will tell you that it is about the team and the market. Fundamentally, the idea is less important. And so I think there's broad acknowledgement that it is doubly true or 10x true in the world of AI because things move so quickly that your ability to not just make good decisions, but also make them fast and attract a good team fast and avoid some mistakes in the process and adapt yourself is just higher.

11:34And so I think in that combo, investors are just very, very keen to invest. I mean, I think there is definitely a gold rush and there's definitely probably an underappreciation of the risk side of this equation. I mean, investors look at a company and say, if this works out, it is incredible. And I think it is more the upside is bigger than ever, but also the risk is bigger than ever. And that doesn't always manifest into the calculations. On the founder side, I think you care less about that equation. You kind of care about getting the right investors and the right terms. but I do think that there are upsides and downsides to raising substantial amount of cash you have to make sure that you are disciplined about how you spend them you don't spend them before you're ready that you don't fake product market fit by just buying it and and in general that you are able to kind of stay sane and remember that you haven't actually succeeded yet eventually success comes from product, it comes from delivery, it doesn't come from fundraising.

12:44So I find I got a lot of congratulations for our sort of series A. I know it's nice, thank you. But in practice, what I want to be congratulated for is for building a product that people love, that provides a lot of business value for building a successful business. Those are the things to truly celebrate. So I think there's good reason for investors to be bullish about AI. I don't think it is vaporware. Um, I think, uh, I think founders should take advantage of that as long as they're, as they trust their ability to stay sober. Uh, I've personally split up the funds at the beginning of like, Hey, these are my operating funds to sort of reach the next milestones.

13:25I think it's hard to really manage it in that fashion, but you still need to be mindful of use the money to grow faster, but, but be smart about where you're spending it. Don't just think of it as a huge amount of money that you need to spend in 18 months to reach the next milestone. I want to get into fast growth in a moment, but you mentioned about obviously getting the right investors. Who were the right investors for you and why? I think, so I did two rounds. I did a seed in the Series A. So in the seed, I really, I've had the benefit of knowing a lot of great people in the industry and having a little bit of my pick of the leader to who do I want to engage.

14:06And so I re-engaged with Ed Seam at Boldstart and Tom Hume at GV, who are both, you know, incredible people and incredible investors. And I've had the pleasure of working with them at Snyk and with Ed, even with my previous company, the first one at Blaze, and co-invest with them and they've become friends. And so they were people that I wanted for the journey regardless. I thought of bootstrapping. It's a bit of an unfair term. I can fund it to a substantial degree, you know, myself with self-funding. But when you have these incredible people that you can bring to the journey, you want them there and they will actually be able to help.

14:40And so that was the first one. And I did want very good terms. And so I raised what ended up being a$25 million seed at a 160 post. So it was already high with great people. and then I brought in some really, really smart angel investors, the founders of HashiCorp and Intercom and CochurchDB and Netlify and just sort of like really, really capable individuals that could help me get advice that, again, I had the benefit of knowing them. So that cohort was, I think, most obvious of just people I want in the journey. And then when I felt like I know what I'm building and I want the freedom to not worry about fundraising and to be able to subsidize big moves without needing to be distracted by fundraising.

15:31Then I opened the door to three investors at the beginning in sort of September, October of last year that again, all of whom are people that I knew and said it needs to be kind of a preemptive Series A. And here are sort of the terms. And I ended up choosing Index as the lead and Accel as the follower. I think they're the two top funds in Europe that are sort of tier one funds. With Accel, I've had a lot of experience at Snyk, and I've had a lot of great relationships over there. And with Index, which I was happy to have them lead, I really think highly of Carlos, who was a CTO at GoCardless before and very technical and gets the domain.

16:20And I really like Index. I like the culture there. I feel like they have a very strong operating arm today that helps on the marketing side, helps on the hiring side. And I've generally been only positively impressed since the round. Fundamentally, I brought some other great people, like First Minute Capital joined in a bit. And I think fundamentally you want to get people that would make you better on the journey and that would, I guess, as someone with the benefit of sort of a stronger negotiation stance, it was important for me to be held accountable but not give up control. And so in the terms, and as we discussed those, I was able to maintain that doing this journey without accountability has the risk of becoming a bit of a vanity company.

17:11Then maybe also sort of one last thing, which is there is an advantage of having an externally acknowledged valuation and share price from a credible investor or actually multiple investors. When you talk about hiring, especially for the people that you hire in the AI world out of the AI labs that oftentimes are bigger companies that are very familiar with RSUs. And so I think making up how much you think your share is worth is not as valuable. So looping back to the fast growth sort of point, like AI enabled or AI native software companies, the goalposts have changed in terms of the metrics to grow faster than ever to get to that kind of next milestone.

18:01right and what we're seeing and i mean i'm i feel like i'm seeing it on linkedin all the time at the moment and maybe not just because of of lovable and whether they're saying okay fastest company to get to 100 million ar and then i'm seeing another post i can't remember uh another company it's like fastest company to get to 10 million or fastest to get to 50 um but the these ai native software companies are growing very very fast right um what do you think of the reasons uh there and then is this like adding in you know a bit of pressure uh like to you as you said you're in uh like beta so like soon to you know uh fully launch the product right and this expectation of like shit all these guys are like getting to 100 million in whatever uh you know seven eight months and you know are we going to be a failure if we if we don't get to you know certain milestones and what those are?

18:57So I feel, first of all, I'm excited by these rays of growth. I think there are a bunch of drivers to it. Part of it is just the entire industry is in a state of fast adoption. Everybody acknowledges something big is happening. Everybody acknowledges nobody knows how it's going to land. And so really, you need to get stuck in and evolve. And this acceptance doesn't just happen on the developer side. It actually happens at the board level. and so budgets are being approved and people really kind of lean in to try out different tools. There's also, for good and bad, we'll get back to this, a very high willingness to try out different tools.

19:35Before trying to get a developer to change their IDE felt like, are you crazy? There's no way I'm going to do that. And now with all these AI IDs, user loyalty is actually very low. People will bounce around. They might use multiple IDs at the same time and even explore things like warp for instance which is like a terminal semi-ide alongside a cursor or windsurf or whatever that is you know has a terminal within the id so there's a lot of willingness to experiment a lot of budgets could go alongside it um so so i think it's it's an exciting time to get people to try new things and when something catches i think it is very everybody's again very spending a lot of time learning and looking for the right things and so when it solves the problem it's easy to get it in.

20:19I think there are maybe two or three concerns around what's happening that I'm mindful of. One is a lot of the actual values in the underlying LLMs and a lot of those dollars being spent are really in practice being spent on finding product surface that unlocks the LLMs' ability to help. When I think about cursor, when I think even about Lovable and definitely about many other tools, they are product surfaces, but the majority of the magic comes from the LLMs that are well-employed. And similarly, maybe the related concern is that these products generally operate at a sort of zero or sometimes negative margin.

21:00And so they make available dysfunctional, but most of their funds aren't going back to the sort of the big labs that are providing them the the systems in the world of coding. Anthropic is a big portion of it, but OpenAI has a big piece as well. And so they're not necessarily viable businesses yet. They are fast growing, but it's interesting to know, you know, what is the profit, like not even like a profit for the company, but like operating profit chart of that zero to a hundred. I think those charts would look a lot less good. And that is coupled with very high churn rates, Very easy to transition very low loyalty amidst users.

21:45And nobody's a bad actor here. It's just nobody knows what's going to happen. The LLMs change and they change the rules of the game. So when I think about TESL, I first care about creating value and solving real user problem. I think when you do that, the ground is set to get fast adoption. Uh, I, I care about getting revenue, but I care about getting revenue as a proxy for delivering value and people are happy to use it. But I also think a lot about sustainability, about how do you, how do you build something that has a place in the world today, but also that if that succeeded, it continues to build power.

22:20Um, and, and those are much more elaborate in the AI world, uh, uh, and, uh, and harder to be certain about. But if you don't think about them, the chances that you will happen upon sustainable power are very low. Reflecting, because you've got the benefit of, well, actually, experience beyond Snyk, right, with the other companies you founded and sort of work with, including Akamai. But with your time at Snyk, what leadership lessons have stood out for you, especially around assembling a team that has the ability to become a high-growth team? I think, um, uh, probably the primary lesson is that you can tell people what to do.

23:04You can tell good people what to do. Um, you need to convince them and to do that, you do need to invest in clarity of your messages. You need to invest in repeating that message when it's relevant. Um, and, and, and so I think, uh, um, I hired, uh, have the pleasure, right. or feel humbled by the caliber of the individuals that I managed to draw to the Tesla journey. They're very, very senior, very capable individuals. They're excited by the vision and they come with opinions. Senior people will have opinions and you want them to have opinions. You want to give them independence, but you do need to continue steering the ship.

23:50There's just a lot that you need to invest in getting them bought in, in simplifying core principles in a way that is not that dissimilar to what you need to do when you get out there and you communicate the product. So I think that's a big deal. The second thing is that independence isn't just a declaration, it needs to actually get built into the company and the software that you create. And so you have to think about how do you organize the team in terms of ownership and mandates, you have to think about how do you architect the software to allow for that. And the ideal is really a set of independent teams that are able to work on their domain across clear interfaces and are even able to work on independent timelines.

24:33You will have launch times and things like that where people can build, but otherwise that there's sort of minimal dependencies. And so really investing in trying to build TESL in this fashion, I think we're in a pretty good place. And yeah, I guess the last bit I would say is people are like the best and worst part of the journey. Most of my stress is caused by people-related things and most of my joy comes from that. And when people don't work out, when a person is not the right person for the job, any procrastination can cause a lot of damage. It puts you in a bad state, it puts them in a bad state, but it also puts the company.

25:16And so I think one, like an obvious lesson that I needed to learn the hard way a few times at Sneak is if there's a doubt and if something doesn't work out, don't delay, like act on it. I've really yet to be in a place in which I regretted parting ways. And when you do that, you try to do it in a way that is most friendly to the individual. But you can't avoid it because you're hurting everybody else how do you see um with this new ai native landscape affecting the role of the cto and i think developers as well um yeah because i mean again it's like these are kind of click-baity posts potentially but seeing that the role of the cto might be you know eradicated with the new ai native uh companies but you welcome your thoughts around that yeah i think um so i think everybody levels up when it comes to software development and so a developer uh effectively becomes a manager of a bunch of ai labor uh and and by doing so has to think about like the broader sort of suite of of functionality and has to think about how do they capture and communicate what they want to build and how how they want it built, and then supervise the work, ideally giving the guardrails or the creations in our world the specs and the defined plans to execute on it so they don't need to supervise.

26:44And they need the ability to drill down, but success is when you don't need to drill down often because you just sort of see the success at a high level. And I think that's really where a lot of our focus is right now is how do individual developers grow and build. I think this similarly moves up everybody else in the stack. And so if every developer is a manager, at some point maybe they get good enough to kind of really be a director. Does a development manager really become a semi-exec and they need to think about business needs and they need to think about customer interactions and so on? So I think just everybody levels up.

27:22I don't think the CTO role changes, but I do think CTOs need to work with individuals or sort of very small teams that get a lot done. And so they have to be more empowering in that sense. I think they have to embrace a CTO who is very tech oriented and implementation will have to expand their view to understand more business needs, more company needs. And I think every company will just do more. There will be more that is being done. And so ability to be agile, to rule out ideas, to kill things that you've tried, understanding how do you evaluate those. So all of that loop from software creation to deployments to customer engagement, coming back and improvement is something that I think will be supercharged.

28:08And then the last thing I would say that I think is interesting is the profile of who is a software creator will change. And a certain class of software will definitely be created by people that are not developers, that don't have technical skills. And I think you still need to manage the estate of software that you have. It represents an attack surface. It represents value opportunities, cost structures. And so there's still a need to manage all of those. But you have to think more about how, as the technology leader of the organization, you kind of liaise and interact with maybe people in sales and maybe people in finance and maybe people in HR who are building applications now and creating them or in product who are shipping applications to customers to try.

28:57And so there's, again, a certain amount of breadth of technology management that is maybe similar in some ways to what a CFO has to do, dealing with everything in the organization. But finally, looking ahead to Sassock Europe, so you're coming over to speak October the 14th and 15th in Dublin. What central message are you looking or hoping to convey about AI-nated development, the AI era, you know, with the audience at Sassock? Yeah, I guess my hope was to share the learning so far from the journey and combining them with some sneak knowledge. The AI building in the AI world has sort of two hard problems.

29:39One is you're building for an unknown future. And so you have to guess your way. And two is the ground is shifting from under you and things change all the time. That makes for a very tough combo. At TESOL, we're launching what is really effectively the third product that we've built because of this constant change. And I think there are some best practices that hold true in this world and you have to preserve. And there are some beliefs that you have to shake off or adapt. And I'm hoping to share a perspective on it. It's not a solution yet. We're still early in the journey, but hopefully some lessons learned and some perspectives for the future.

30:15Awesome. I'm looking forward to seeing you in person in Dublin and having you share this with the SaaS.com community live there. So I appreciate that. But Guy, if people want to reach out to you online, if they've been listening to the podcast, have any questions, where's the best place to find you? So I think easiest is just go to tessle.io. You know, if you're curious about the product, join the waitlist. It is not going to be a waitlist for much longer. So try that out. And then personally, I'm probably best on LinkedIn and a little bit on Twitter X to engage with probably more LinkedIn. Awesome.

30:50Well, thanks so much, Guy Pojanuk, for being on the podcast. Looking forward to seeing you in Dublin and yeah, looking forward to see the future of TESOL and the impact on the software industry. Thanks so much for being guests on the podcast today. Thanks for having me. Thanks for listening to the SaaS Revolution Show. If you enjoyed this episode, please leave a review and follow the show. It helps more SaaS and AI founders to discover the podcast and keeps us bringing you the leaders who are shaping the future of the industry. For more insights and to join the SaskTalk community, head to sasstalk.com.