Best of 2025: From Startup to Exit Podcast

Join us for a dynamic highlight reel featuring short segments from the 2025 season of From Startup to Exit, TiE Seattle’s flagship podcast showcasing world-class entrepreneurs, innovators, and industry leaders sharing candid insights from their journeys. Each mini-segment brings you the essence of these remarkable conversations — from foundational lessons in product and leadership to reflections on how technology shapes markets and culture. 

We kick things off with Paul Maritz, discussing his pivotal role in Microsoft’s product strategy during the Windows and server era and how a deep commitment to innovation shaped a career spanning multiple iconic tech companies. Next, Jeff Raikes reflects on building transformative software platforms like Microsoft Office and his later work in global philanthropy focused on equity and systems-level impact. Robbie Bach takes you inside the historic launch of Xbox, offering a rare behind-the-scenes look at driving innovation within a legacy tech giant. Meanwhile Steve Wood shares foundational memories from Microsoft’s early days as the sixth employee, emphasizing the importance of grit and experimentation in startup culture. 

We also spotlight Brad Chase on the launch of Windows 95 and the strategic decisions that made it a defining moment in personal computing; Laura Jennings on shaping MSN into a major internet portal; and modern voices like Sabrina Wu, who breaks down the future of intelligent applications and the role of AI in venture investing. Yash Wagh delivers insights from his startup journey with Gone.com and building sustainable commerce models. Ece Kamar closes with a forward-looking conversation on responsible AI, sharing how human-centered frameworks will guide the next phase of technological growth. Whether you’re an entrepreneur, technologist, or curious learner, this highlights compilation is packed with actionable wisdom from leaders who’ve navigated the path from startup to exit.

Brought to you by TiE Seattle
Hosts: Shirish Nadkarni and Gowri Shankar
Producers: Minee Verma and Eesha Jain
YouTube Channel: https://www.youtube.com/@fromstartuptoexitpodcast

Transcript

SPEAKER_16 0:00

You're here because you want to understand where AI is headed and how companies like Microsoft turn innovation into lasting dominance. This is from Startup to Exit, the podcast where we sit down with serial entrepreneurs so you don't have to, uncovering what really drives their success. In this episode, we've distilled the most powerful insights from our 2025 conversation, featuring top executives from Microsoft and leading a Seattle-based AI startup. If you want to know what it takes to build a startup that skills, endures, and wins, start now.

SPEAKER_02 0:38

First, Paul Murrit, CVP for Microsoft's desktop and server software.

SPEAKER_08 0:43

Microsoft also missed a few things. The browser, right? Early on, there was a mess. Mobile was a mess. Was there something in the leadership that Microsoft had that allowed them to overcome such large misses?

SPEAKER_04 0:59

Yeah, it is remarkable. And I think it's attributable to two things. One is just the strength of those assets that Windows and Office were. And even today, the majority of the profits are still coming from those two sources. So it gave Microsoft an enormous cash flow stream with which to invest. And they tried several investments, as you point out, like buying Nokia, et cetera, that didn't pay off. And I think the other so they had this very strong two cash cows in Windows and Office that that gave them a strong cash stream. And then the third time they got they succeeded, which and I give that full credit there to Craig Mundy. Craig Mundy was the guy who went and did the deal with OpenAid and gave them billions of dollars in return for royalty-free access to everything that OpenAI does until such time as they reach AGI, which is what I think is giving Sal Sam Altman, you know, partake today. But without that, Microsoft would not be relevant today, and we'd be having a very different conversation about Microsoft today. We'd be saying, well, you know, they missed the boat once again. It is Craig who went out on the limb and persuaded the board and and uh uh and Steve and and and Bill and then Satya to basically put billions of dollars into allowing uh essentially the scale to happen at OpenAI. OpenAI's fundamental breakthrough, they may argue with this, was not the algorithms, but was just simply the scale at which they applied it, you know, allowing them to go to models that had billions and billions of parameters, which is you know brute forced by essentially building out this enormous hardware back end that's enabled them to succeed. But it's a strange relationship. I mean, uh I I it's debatable how stable the relationship with OpenAI will be because you've got two different entities now that at the end of the day have different ownerships and different economic incentives. So I don't know, I'm not not privy to what's going on in either company, but it's it's gonna be interesting to see how that gets squared in the long run.

SPEAKER_08 3:06

But at the at the you know, uh if you go back uh say 1995 and the browser really came into being and and the adoption was at a at may not be as the rate as ChatGPT, but it was still at a at an incredible pace at the adoption, right? That you know, Amazon it literally went to IPO four years later, once the browser became ubiquitous, if you think if you sort of use that as a parameter. Um at that time you were building Internet Explorer. Did you think that you should just buy out Netscape? Because that seems like you could have done it. That was Barksdale was an old ATT guy.

SPEAKER_04 3:47

Yeah, we very early on we had an opportunity to deal with do a deal with uh Thendriesen and uh I forget what they called it before they started calling it Netscape, and uh but we we kind of somewhat arrogantly said this is 50k bytes of code again. Why can't we just write this in a weekend? And I and we did actually license some code from one of their competitors. I you'd have to ask Ben Slivka, he probably would know the history here. So we did license some code, uh which was the first incarnation of what became Internet Explorer. Uh and uh you know I I argued very strenuously to the Department of Justice in the in the the uh antitrust trials that this was a huge threat to us, that uh, you know, if the browser took off new applications, it would become a new applications delivery platform, et cetera, et cetera. They didn't buy that argument. In the final analysis, it actually did happen.

SPEAKER_02 4:41

Jeff Riggs, president of the business systems division at Microsoft.

SPEAKER_13 4:46

Well, you know, when I when I joined Microsoft, my my first day was November 13th, 1981. And and there were some really fascinating there was uh there was an all-company meeting. I think it was the second time they'd brought everybody together. And there was a guy named Kay Nishi that was Microsoft Japan. And Kay says, you know, the telephone is an extension of your ear, the television is an extension of your eye, but the computer is an extension of your brain. So that was that was kind of my introduction to Microsoft Vision. And then Charles Simone, who I was very fortunate to work with, Charles does this presentation which became known as the Simone Revenue Bomb. And the idea of the Simone Revenue Bomb is all about leveraging code. You leverage code from one application to another, from multi-plan to a word processor to a database, and you leverage code from one machine to another using a P code or pseudocode interpreter. And that was the application strategy to be able to leverage code. And if you did had that leverage, you'd have to hire so many people in sales that there'd be this revenue bomb. And so that was the Simone revenue bomb. And that was our strategy. And then as Sharish, as you said, January 20th, that's the date, 1983, Lotus 123 launched. About the same time that the Compaq IBM compatible launched. And so suddenly you had a clone of the IBM PC that could run Lotus 123 as well as the IBM PC. The Lotus 123 really uh shaped my thinking about software strategy. How did the IBM, how was the IBM PC different from the Apple II? Well, it had a faster processor, a faster microprocessor, it had more memory, and it had higher resolution graphics. What were the three key features of Lotus 123? It had bigger spreadsheets, faster spreadsheets, and built-in business graphics. Lotus 123 differentiated the IBM PC on one of the most important types of software, spreadsheets. And so what I learned from that, and I used that phrase in 1983, is that if you want to win big, you have to make the right bet on the winning platform. And I could see that the IBM PC and then compatibles were going to be the winning platform. And I could see that Lotus had made the right bet. You know, they thought they won because it was integrated software, but really why they won was because it was a better spreadsheet. And by the way, they made a mistake thinking that they won because it was an integrated software product. They did this product called Symphony, which was even more integrated and never was as successful. So they didn't understand the seeds of their own success. And so on February 1st, 1983, we started the PC Word project. Microsoft Word on the PC was not designed to run across all of those computers. Uh it was designed to take advantage of the IBM PC architecture. So we could do bold and italics on the screen. We could use a computer mouse, so on and so forth. And Microsoft Word on the PC shipped on October 25th, 1883. So that was a very accelerated development project. And that was the pivot in our our the first pivot. And then if you want me to go on, I can tell you about the the the second pivot which occurred in that next year.

SPEAKER_07 8:33

Right, yeah. Please uh talk, I assume we're talking about uh Excel first on the PC platform and then eventually on the Mac.

SPEAKER_13 8:40

That's right. In October of 1983, we held a retreat of engineers and Jabe Blumenthal, the program manager. Jabe was brilliant. I mean, he had evaluated every spreadsheet out there uh because there were a lot of spreadsheets and financial planning products, and and you know, Doug Clender was an awesome you know developer. And they together designed what was codenamed Odyssey. And um it was gonna be a uh a competitor to Lotus 123 on PCMS DOS. And in the spring of 1984, this is a few months after Multiplant uh had shipped on the Macintosh. Apple Macintosh shipped in January of 84, and Microsoft was the applications company, the first one to have software products on the Macintosh.

SPEAKER_02 9:33

Robbie Buck, chief Xbox Officer at Microsoft.

SPEAKER_08 9:38

You had this unique relationship with publishers, right? Because you could you you are a publisher, and then you had to bring other publishers to get on board with your hardware. There's whether you that would be a long-term thing or not. How did you at that time convince everybody this bet's the best bet they could make? Because the publisher had to make a bet. Because they were they were working on other devices at this point, or they were publishing for other for either Sony or Nintendo or something like that.

SPEAKER_11 10:07

Well, we had an advantage in this respect, which is that Nintendo increasingly didn't want and didn't care too much about third-party publishing. I mean, they had so much of their own content and so many of their own franchises that they didn't want the competition. And so they weren't recruiting heavily for other people to join their platform, and they actually were making it difficult for people to publish on their platforms. And so it was just not very third-party friendly. You know, they sort of felt like, hey, we're the only game in town. If you want to pu Electronic Arts Activision, if you want to publish, come to the PlayStation. But this is what it's you're gonna have to pay us the royalties. And sure, we'll help you a little bit. But in particular, if you were a second-tier or third-tier publisher, you got zero love from Sony. So our strategy was to be the publisher's best friend. And so basically, when we went to meet with them, first thing, first, first things first, I went to meet with every publisher twice a year. And I don't think the head of Sony meant to meet with any of them except when they were at a trade show, maybe. But we did visits at their facility every year. And we shared with them everything we had. We answered every question they had. So there's a personal commitment aspect to this, not just from me, but from the senior person at Microsoft. Um, sometimes we got Bill and Steve to meet with them. They knew Microsoft cared about what they were doing way more than Sony did. The second thing is we were really quite aggressive about making it clear to them that if you want to have a balance to Sony's power in the marketplace, you need somebody with Microsoft's pocketbook there. And you need somebody with Microsoft's capability to balance out what Sony's doing. Our success will make you more money. Not just because you'll sell units on our platform, but because it will keep Sony honest in the marketplace. I think that was a key reason why Electronic Arts signed up is because they were struggling in their relationship with Sony and arm wrestling over power and royalties and all the rest. And then the third thing we did, which was uh a George Peckham innovation, George led our third party group. We basically said we're not going to offer volume discounts to publish. So you might say, well, electronic arts and activision probably didn't like that. And the truth is they didn't. But it did attract the next level of partners to the platform because they were getting killed by these volume discounts that Sony or uh PlayStation, Activision and Electronic Arts were getting discounts on royalties. So they're paying less royalties to Sony, which gave them some pricing power. And people like Ubisoft and THQ and others did get those discounts. So we basically said to people, there's no volume discounts. If you come to us with an idea, we will jointly market it with you. So we'll pay for some marketing for games we like. But we don't care whether those games come from Activision or Electronic Arts or Ubisoft or THQ or somebody else. And you know, ultimately that leveled the playing field a little bit in the publishers. And if you were EA in Activision, you didn't mind it so much because you had Call of Duty, which you knew we were going to support. You had Madden, which you knew we were gonna support. So you got your discount in marketing, not in royalty dollars. But if you were Ubisoft and you had the Tom Clancy series and you're getting no love from Sony, you can come and be the first guy to do a great Xbox Live game with a Tom Clancy series, and suddenly you had a big market and you got supported by us. So the combination of all of those things worked really well uh with our publishers. George and his team were a little known but secret to our success. I go in a meeting, you know, sometimes you go into a meeting with the publisher and you'd be talking about their upcoming game titles, and it was clear that George's team knew as much or more about the upcoming game titles as the senior leadership that the publisher did. They were really deep with the development teams, really deep with the external development teams. And having that level, this comes from Microsoft's historic work with developers. Right? And there's some DNA in Microsoft about developer support and management that carried over into the Xbox world that was really powerful. And again, it doesn't get written about a lot. They don't get a lot of love or credit, and they deserve a ton of credit for what we did with those guys.

SPEAKER_02 14:35

Steve Wood, Microsoft's sixth employee.

SPEAKER_08 14:40

At that time, as you articulated, there was no independent software company. There were hardware companies that bundled software. And so the the reason that they thought that could work at that time was what? Well, in Albuquerque, especially, because there was no clear path.

SPEAKER_03 14:59

The reason that they thought it would work was actually the reason that it turned out it did work, which was it, it's the the whole vision, the ability to look at the first microprocessors that were coming out, and to then say, you know, have the vision to be able to say that's gonna create a market that standardizes around processors and instruction sets. And then the software becomes unique, uh, becomes the unique thing that you can sell these things with as much as the hardware. So you know, hardware's obviously always important, but it does in a lot of cases these days has sort of become more of the commodity. Um and the the unique part of it that makes you decide whether to buy one PC versus another is the the software that comes with it. So um it was all being driven by what ended up being called Moore's Law. So you know, the the these standard microprocessors were coming out and they're evolving at a very fast pace. And back then, uh Bill Paul first, and then Bill had the vision to be able to look at that and say, five or ten years, this is gonna be enormous.

SPEAKER_08 16:25

So now that sort of makes me kind of think of the choice you made. You're married, as you mentioned, to Marla at Stanford. So you could have made a fairly safe bet. There was Hewlett Packard locally next door to you in Palo Alto. Apple was, I think, uh early stages, but there were other players, right? IBM.

SPEAKER_03 16:48

Yeah, Apple wasn't quite there yet, but yeah.

SPEAKER_08 16:50

Yeah.

SPEAKER_03 16:51

There were other people. But local.

SPEAKER_08 16:54

Yeah, yeah. And you could have gone to Intel. I mean, you could have gone to many places at that point, right? You picked an unknown company in a city you have never visited to move after Stanford with your wife. So that seems like a that was way early comparatively. Maybe that's the power of being young, right? Because what's what's the big deal that could happen, kind of a thing. But what stood out in that that post that you saw in Stanford that made you pick them over what by then were established next door companies like HB?

SPEAKER_03 17:31

Well, I think the I mean the the first one was the point you made that you know you're young, you don't know any better, you don't have much to lose, right? So but you know, the second question you asked, the answer is I think that it was the opportunity to be a key contributor, um, you know, maybe even leader on uh a new product or a new technology being developed, as opposed to just being, okay, I'm now a junior member of a team of 50 people working on something. Even though that the something that HP was working on was pretty cool. The opportunity to be um an influencer in where that technology goes and how it gets developed, I that was really attractive. And then, you know, in terms of personality, we were both pretty adventurous. You know, it was interesting to to see other places and experience other places. So sure, why not?

SPEAKER_02 18:35

Brad Chase, senior vice president of Windows 95.

SPEAKER_07 18:39

So your marketing strategy for Windows 95, the the uh three E's, uh, which is I think really nice uh you know uh way to encapsulate it, which is uh educate, excite, and engage. So tell us more about uh your your marketing strategy uh for Windows 95. Because that's a really important you know release. First time I think Microsoft had a Windows product that you could say was better than the Macintosh.

SPEAKER_14 19:09

Yeah, it was an amazing product. So, you know, one one of the things I get a lot of kudos and my team uh for the marketing of Windows 95, but it all starts with the fact that the development team built a great product with our input too. But they wrote the code, obviously, and and and those guys, you know, David Cole and his team did just an amazing job. Uh Windows 95 was the product that ushered computers into the mainstream, and there um and I think people may have sensed that possibility because there was incredible interest in Windows 95 everywhere you went. And so when I became the marketing leader, all that was happening was we were trying to push away everybody who was trying to find out more about the product and learn about it, and you know, everybody wanted to hear the latest rumor about it, whatever. And one night I went home and finally, after the kids were to bed, I was thinking about it, and I I decided that we needed to flip the whole marketing strategy on its head. And instead of trying to keep people from learning about Windows 95, that we should encourage people to learn about Windows 95, even though the product wouldn't be out for a while. And so I came up with the educate and engage and excite strategy, the e strategy, as you mentioned, to try to get people to learn about it, to get excited about it, and for third parties in particular to engage about it with it, figuring that that would speed up the adoption of Windows 95 when it launched because people would already be very familiar with it. And the ultimate goal was to make Windows 95 a consumer phenomenon. So everybody would know about it, would have heard about it. And I like to say, well, in fact, I told the team that about a couple weeks before the launch, when Doonesbury ran a comic strip for a week making fun of Windows 95, that in fact we had achieved our goal to become a consumer phenomenon. So that was that was the goal to become a consumer phenomenon. And the strategy was all about stopping from preventing people from learning about Windows 95, but instead to try to teach them, get them excited, get them engaged so that when it launched, it would the transition would be simpler. Got it.

SPEAKER_07 21:20

So what are some of the key elements of the launch? I'm sure uh, you know, uh, you know, getting Wolf Mossberg to bless Windows 95 was uh pretty important. Did he write about it before uh it launched or was it uh at launch?

SPEAKER_14 21:37

Well, his main review was at launch, and I remember it very well because Walt and I were were very close. And I I was the first Microsoft person he met with when he became a technology reporter, actually. And that at that time I met with him about MS TOS 5. But when Windows 95 shipped, he did a review, and that review was very positive. But he kind of said, Well, I won't give my full recommendation. To upgrade until we find out if the product's got any kinks or problems. And so after a few months, I call up Walt and said, Walt, you need to do another review of Windows 95. And he said, Brad, I already did a review of Windows 95. And I'm like, But Walt, you promised your readers you would get back to them. And he's like, Well, you're right. You got me on that one. And so he did. And then he recommended Windows 95. You know, that was a that was a key moment. Because Walt and I had many debates, you know, where we disagreed about stuff and agreed about stuff. And but it was all in a very uh positive and constructive form and tone. I really, really respected Walt a lot. Still do. But that was only one part of what we did. I mean, the PR effort was gigantic, obviously, with Windows 95. But we had all sorts of programs from demo disks to trainings of corporate people to working with OEMs to consumer workshops to the whole beta test process to the launch itself, which was obviously a big deal and through uh you know a major event all around the world, to the to the Rolling Stones deal for Start Me Up, which to this day I still get asked about all the time. So all those things combined to make it a a great, great launch event. And again, I want to you know reinforce that it was an incredible product in the and the product you can't you can't market a crappy product in the software industry and be successful. Right. But you can screw up a good product or help make a good product successful. In this case, we took a great product and helped make it successful with a really stellar marketing effort by my team and other people around the world. We kind of created a template and messaging and what we were trying to achieve, and then all the Microsoft people around the world built on that, and so they had a lot of creative ideas as well.

SPEAKER_02 23:50

Laura Jennings, vice president for Microsoft Network.

SPEAKER_07 23:54

So when did you take on the role of at MSN?

SPEAKER_15 23:57

I took on the role at MSN after launch. So, like everything at Microsoft, it was launched everywhere we were. So it was launched in whatever that was, 50 something countries, built on a proprietary X25 network right at the wrong time because that's when the world is moving to the internet. So I took over after launch, maybe six months after the launch of MSN, and really was responsible for then exiting all but seven of those countries that we had launched in and rebuilding the technology and our whole approach differently. So you just couldn't ignore that the the world was gonna change. And then I think about that as this kind of second fundamental shift in technology from thinking about distributed desktop computing to thinking about software as a service. And MSN was really the first place that Microsoft touched that idea of software as a service as we're we were rolling out MSN and really trying to figure out what are people gonna do on the internet? Should we take what we do in terms of productivity apps? And that's what the world is gonna look like? Is it gonna be entertainment-based? And so you see, in those early days of MSN, we're spinning out ideas that represent both of those elements, productivity tools and entertainment, and sort of trying to figure out what to do, and and what is the relationship of that to our existing software suites. So that was so during that transition, we lost four hundred and fifty million dollars, I think, that first year. And then and we wrote a lot of things off on a balance sheet at home at Hall Network. And so it was uh it was a time when you know I would present to the board every quarter what was going on. This was before losing a lot of money. Yeah, it wasn't something that technology companies routinely did.

SPEAKER_08 26:11

So, Laura, one of the themes that's emerged talking to all the folks we have talked to as part of the series was convincing the leadership team a change had to be made. They were well invested, they go down these paths, go down, and then you gotta make a full-on pivot. And uh What was uh because you having presented to the board, you were convincing Bill and the full all of the board, hey, we're going from this to this, which is not not is a non-trivial exercise. What what was your thinking as to convincing them, hey, this isn't gonna work and we gotta do this? Was there some some leading indicators that you used or was it obvious to them? Or how did they get convinced?

SPEAKER_15 26:55

Yeah, that's a it's a good question. Well, it was it was a lot of people, right? I mean, I I would say that I don't think Bill is the guy who needed convincing in terms of thinking about where the world is going. And we clearly had been behind. And it took us longer to launch MS. Again, I was still sitting over at interactive in interactive television land, but it took us longer to do that. We're looking at what other people are doing. I mean, the competitor we had gone to market thinking about was AOL. And the and the the world was changing. And we were losing a lot of we were losing a lot of money. This was clearly not going to be a successful strategy.

SPEAKER_02 27:28

Mike Fridgen, the managing director of Modrona Venture Labs.

SPEAKER_07 27:33

Can you talk about uh the kind of the architecture of the solution and how it was built?

SPEAKER_12 27:38

I mean, first, this is a very modern AI application. Uh so I think in a world where uh apps will be built in a much different way, where there it's multi, it's it's multiple agents playing a role in an orchestrated system to accomplish a complex task. So in our world, we are using large language models tied to different data sources. To give you an example, we have one agent that plays a role around calendar. It ingests calendar, as I said, itineraries, meetings, and it understands and makes sense of that to understand your preferences and your behaviors. We have another agent, which is an LLM tied to a traveling PI. It's pulling back real pricing and availability for flights and hotels and other and other travel components. And you can imagine multiple agents who are then all play different roles and have different prompts tied to them. We also have a personalized agent that understands not only long-term memory, the preferences of the past, but preferences and intent you've shown in the moment of indicating a trip or of directing a trip. So all these systems are coming together. And in terms of the models we're using, there's a lot of flexibility here. You know, we'll have different models for different purposes. In some cases, we might have a performance objective, other cases a cost objective. Other cases, we might need, you know, a stronger model with more capability around uh understanding uh, again, these these personalized data sets and in and and being stronger in terms of how we articulate what we know about a user. So we have flexibility in what models we plug in for different use cases, and that'll really evolve over time. I see it being multi-model, though. This is not a single model we'll use on the back. This will be a set of models that all play different roles in the system.

SPEAKER_07 29:29

And those models may change one day you may be using GPT-4.0, another day you may be using Llama or something, depending on how you need to change and et cetera.

SPEAKER_12 29:37

Well, certainly. With the rate of change, uh that's almost certainly uh the only thing we know is that the exact set of models we're using today will evolve and change.

SPEAKER_08 29:48

Mike, it's uh interesting. You you kind of said, you know, modern AI company. I thought you were an AI-born company, but you were already, there's already uh the three years that models have been around, there's already a you know, AI, we have to modernize how we think about language, you know, ML and AI models. It's very interesting. So let me start with this premise, right? Post-COVID, would you say the segment you're going after has a pro-leisure approach to travel, meaning I go to headquarters, but I also extend it two days, or I go to someplace, I extend it two, three days, and or stay two, three days to meet with my other colleagues beyond around a meeting. So it's somewhat not, it used to be you left on a Monday, came back on a Wednesday, but then your leisure trip started on a Friday. Seems like it's blending.

SPEAKER_12 30:44

Would you would you guys consider that? There's no question it's blending. There's a term in the industry that's called bleisure, a mix of business and leisure, and that's very common today. Uh and so, you know, those are things we're thinking about in the experience uh and the interaction model of um you know planning and booking these trips. But yeah, I mean, but you know, you you were speaking a bit about you know modern AI and and modern apps. I mean, we are, you know, inventing as we go here on the edge, what you know, as we're learning and and uh iterating on what what these what's possible with this technology. You know, some people think we'll have an agent everyone will have a personal agent, right? That'll help us accomplish tasks in our lives. Almost like everyone has this supercharged EA in their life, not just for professional reasons, but for personal reasons. And if when you think about these specific verticals, they're so complex. Uh and they they they tap into deeper uh data sources that need real time around. I mean, flights is in and hotels are a great example. I mean, those those that data is not sitting in the open in the models today. That's real pricing availability that's pulled down in the moment, that changes in the moment. And so these more complex systems just require um uh a more a more uh focused vertical execution. So I you can imagine that there'll be an orchestrator personal agent to every human, but underneath that are a set of vertical agents that have deeper capabilities.

SPEAKER_08 32:23

And every CISO, I'm sure you talked to the chief security officer of every enterprise, wants this balance between protection and AI implementation, right? At the end of the day, they're not measured on, they're measured on how well their enterprise is protected, right? As opposed to, and then you have all these various actors who also have happen to have the same access to the LLMs that you have. Right? So now you have this sort of a chess game that you're playing in the SECOPS, which has always been the case, but it's now heightened, so to speak, because they have access to similar tools. Where do you think is the pulse of the CISOs in terms of where they were pre you know, say 2024 pre-2025, to what you see as the uh chief product officer to say, hey, this is kind of where they're heading, or you are navigating them to this is the best path for them to protect their uh their assets. Yeah.

SPEAKER_09 33:24

So uh so here's an analogy, right? Like if investment analogy of if I keep my money as cash, I'm actually falling behind because inflation is just natural. So I have to make sure that my money keeps growing just to keep pace with inflation. Automation security automation is the equivalent in the cybersecurity. If I use yesterday's approach, if I'm not actively thinking, okay, I have I'm doing 10 people's worth of job along with my software. If I'm not actively thinking in terms of the job that 10 people are doing, how can I automate some of it? If I'm not actively thinking about it, I am falling behind. Because tomorrow's load will be more than today, day after tomorrow's load will be more than tomorrow. That's just the natural law of cybersecurity that their load will keep on increasing. So every day you have to keep harvesting, saying, What am I going to automate more? And that's not new. That's always been the case. What's new now is what you just observed, which is adversaries have access to the same models that you have. In fact, there are actually more. There are models floating dark web that only they have access to. And phishing attacks now, way more realistic than they were. You probably heard about the uh pretty famous incident of this finance guy in Hong Kong who got invited to a video conference where he saw his CFO and his colleagues all on the video conference, and he was told to wire uh somebody over, and it turned out they were all defective. So so with uh attackers using AI, uh this inflation uh insecurity is uh going up like a hockey stick. If you're not actively thinking about how to automate a helper scale, you are falling way behind. And so every CISO from that matter that we have talked to actually is like incented from that perspective. Saying, how do I how do I stay ahead of the curve? I don't want to be in six months regretting that I did not invest in AI. Um it sometimes like many of the customers that we talk to come with very specific problems saying that particular thing is painful for me, I need to automate that. But there are many that we are seeing who are just being forward-looking, saying, Look, I don't know whether it's this problem or that or that, but I know I need to get a good automation framework and automate more and more. So what do you have?

SPEAKER_02 35:55

Sabrina Wu, investor amadrona ventures.

SPEAKER_07 35:58

One of your predictions uh for 2025 uh was that it'll be the year of agents. We've actually spoken to two companies already uh who have built agent tech applications. I'd be interested in examples that you might be able to share that you've seen in your portfolio companies uh where you've been particularly impressed with their capabilities. Yeah.

SPEAKER_00 36:17

So I think AI agents, it's really interesting. I think this, you know, there's been a lot of discussion about it in the last year. This year we're starting to see even more buzz, more activity around it for sure. I think that agents fundamentally can help you with two things. One is as a company, it's one, you know, driving top-line revenue. And the second one is helping with cutting costs, right? And so you can think about agents as either a coworker, they they work with you or alongside you to help you achieve a certain task, or you can almost think of them as essentially an employee, somebody that you give work to, and they go and in the background are doing that task, completing that task autonomously for you, right? And so that's kind of how I, you know, think about agents. I think that a lot of people have different definitions of them and there's different types of agents, and there's autonomous, fully autonomous, semi-autonomous. So there's lots that you can dig into there. But I'd say within you know what I'm seeing, there's a couple of things. The first is around the vibe coding space. I think that's been talked about a lot of recent, but basically this idea that these agentic AI tools, you can ask them to spin up an application for you in natural language. So me, you know, as a non-technical person, I can go to a versatile V0 or cursor or Bolt. Bolt is one of Madrona's investments, um, and and you know, ask it to create, you know, call it like a fitness app for me that tracks my meals, tells me how much I should be working out based off of, you know, trying to achieve certain fitness goals and objectives, or maybe I'm trying to run a marathon, how do I create an app that helps me do that in a certain time? And it will go and spin up that application for me. Not only will it um create kind of like the UI UX layer, it also sets up the back end. It's a full stack application that you can have um coded in any language that you want, and it's it's all done for you in natural language. And so I think that's one thing that agents are really great at doing in the code gen space. You're seeing that a lot today. And that's probably one of the bigger use cases that we're seeing. In the B2B setting, you know, seeing a lot more vertical and horizontal use cases of AI agents. One company I call out in the more go-to-market space is a company called Clarify. Um, it's a next-gen agentic CRM system. It really unifies all your customer data and connects into different data sources, systems like your email inbox and also different sales tools that you have and creates a unified view of your customer. And what it does is abstracts away a lot of the tedious manual tasks that sales teams are usually tasked with doing, be that uploading a new lead or understanding like getting data entry into the CRM system, trying to figure out who's connected to this person. And then also just doing things like automatically doing follow-ups. They're all done by the agent itself. So the agent understands all the context and it's doing that in the background for you. And so really it's it's this it's this idea again of abstracting a lot of the low-level tasks and allowing the humans to work on more analytical and important things. And I'd say the agents really act as a, like I mentioned in the beginning, either a co-pilot or basically another employee that you can go and task it to do things for you.

SPEAKER_07 39:17

Do you see that there'll be a transition from semi-autonomous autonomous to fully autonomous over time just because people have to feel comfortable that the agent's got to perform the task properly and so forth?

SPEAKER_00 39:31

Yeah, I think we're still in the early innings of that. And I think it'll really depend on the use case. Like it depends on how customer-facing it is or how much sensitivity there is around it, or how much access you're giving the agent to do, if it's something more simple or for something more complex. And I think that there's there's issues today where we're not fully at a at a spot where the agents itself can reason. It doesn't have the best memory to understand specifically how to what it should remember to do or what you've told it previously, like long context memory windows, like those sort of things still block the agent from being fully autonomous without involving the human in a way that it's thinking like the human would or prioritizing like the human would. But I think once we get there, it will be able to go and do tasks fully autonomously for certain types of work. Um, but I think that we're still pretty early in that and we still need to see some more infrastructure unlock, if you will, for the agents to go and actually execute on a lot of the tasks that they're that they're Yashuag, co-founder and CEO of Gone.

SPEAKER_08 40:31

How what advice would you give those entrepreneurs, especially on the funding journey that you've been on and the support you got from everybody around you?

SPEAKER_10 40:41

There's no there's no magic pill that gets you funding. It is day in and day out of hard work and in general persistence that helps you with funding. Uh unless you are fully into it, you have burned your boats, that there is no fallback option. You have gone into it all in and you're trying your best every day, is what eventually gets noticeed, is what I feel. It wasn't it wasn't a year plus at a year and a half of building gone, and then we like a year and a half of no paychecks, just building, just trying to figure out what we can do, and then eventually pivoting many times is what led to the June 2024 funding. So getting it bootstrapped, trying to figure out how many pieces that you can put in place, connecting with the right people, and at the end of that it like being extremely patient and extremely, for lack of better terms, resilient because not every day you're gonna find an idea, not every day someone is gonna come and help you. Just trying to meet as many people as possible was what led us to this. And I'm very grateful to Thai Seattle for that instance, because my entire journey, entrepreneurship journey, also started with me getting involved with Thai Seattle. It is you folks who introduced me to many people. Obviously, we take the coffees, you take and take the meetings, you understand from them, you learn from them. So we did all of that, but I'm extremely grateful to organizations like Thai Seattle that help us actually get connected into the ecosystem, and that that makes things happen. And the biggest example, I met Kellen at the holiday party at Thai Seattle. I met Kellen because I got introduced into the charter member group, and that's how we reached out. So getting the connection, the networking going while building with your head down, both of those things come together.

SPEAKER_08 42:41

That's excellent advice. That uh it's a journey, and you have to be resilient to uh withstand it and get up every morning and do it all over again. Seems like that's uh and I'm and we are glad that uh we see one of our own, you know, get funded and get get their vision into real execution. Now, you talked a little bit about the pivots that you went through even as you were heading to your launch, right? So that means there was morphing of it. A lot of entrepreneurs get stuck in saying, hey, I have this idea, so this is it, right? You seem to have struck that balance between pivoting and staying core to your idea. How would you say that you struck that balance and were there times in which you felt like you're moving away too much from the core, or you're not incorporating enough of the suggestions you're getting?

SPEAKER_10 43:37

So the two things that you're probably talking about is how do you balance purpose with the market signals? Right? My purpose stems from my daughter. What kind of a world do I want to leave for my daughter as we pass on back into the next generation, right? And that meant that we wanted to create a sustainable world, the carbon scope. Are going to the roof where we are just consuming and consuming and we're becoming a consumerist nation, consumerist people. And my thing was like, we got to stop this cycle somewhere. So the core purpose that I have gone behind is a sustainable, mission-oriented thing where we stop the consumerism and actually are happy and content with what we have. And that means making sure that the second hand goods that have life left in them get used to its fullest rather than trashing and making sure that we don't consume new things. That's the purpose. So if that was the purpose, we started the idea in many different ways. The first way was like actually consignment. Like we started a company 1800 consign where we were able to take things in, store it, and actually give the customer some money back. Once you did that, we we figured out that as we learned from the market signals, the idea pivoted completely and it's a new business again. People did not want money for it. People wanted stuff out of their lives and gone out of their lives. Okay, fine. So we tried doing that. From there, an idea of treasure truck also struck us. Like, can we do it in a treasure truck where we'd come pick it up and just distribute it out there? Eventually, when you hear your customers enough, eventually, when you get the signals right and you match the signals from what the customers are telling us to the purpose, that's how you do it. If you are hell-bent on one idea and if you just want that idea to be proved right, you don't listen to the market, I think it is doomsday. You have to listen to your customer, you have to ask your customers constant questions, and that too when you're with them in a very friendly way without indicating to them that this is what you're gonna ask. So we made sure that we are solving a problem that the customers wanted while staying true to a deep, innate purpose that I have.

SPEAKER_02 45:56

EJ Kamar, director of AI Frontiers Lab at Microsoft.

SPEAKER_08 46:01

As a researcher, what advice would you give entrepreneurs out there to go look for so they can, you know, advance their uh their own startups?

SPEAKER_01 46:12

Yeah, that's a great question. So I have two particular thoughts about that. One is it feels like we are at such an inflection point right now. Because particularly on agents, many core technology pieces are finally at a level of maturity for them to come together and build something integrative and cohesive. Just to give an example, right? We talked, like even in this conversation, we talked about reasoning models. Reasoning models are likely to become a brain of these future agents. Really bring in the tool calling and being able to retrieve information in active ways. The reasoning models is going to have a big role in terms of enabling the agents. We also talked about protocols and tool access and API access as a way for these agents to be able to access the wider world. And of course, there's a lot of innovation happening in specialized models and small models that we do as well that is helping developers to create these agentic systems that are powered by a diverse set of models to really like enable new experiences. One of the things I think about all the time is where is all of these technologies converging to and how are they gonna be composed together to create a future that is much bigger than what we are seeing today? Where are we really headed when all of these pieces come together? And I think asking that question is very fundamental for any entrepreneur to build something that's not only gonna be relevant today, but it's gonna be relevant going to the future as well. So my bat on that question is really thinking about ecosystems, thinking about like how agents, people, what kind of a network they are gonna be able to create, and how we are gonna be able to create outcomes that are much bigger than the sum of its parts as these pieces are gonna come together. So I would really advise any entrepreneur to really try to imagine, like I'm I'm trying to do that. It's not easy with how fast the technology is changing, but like, what are our bets for where things are headed? And can we make any bets on that state of the world where some of these technologies are gonna be converging? And for example, what are the future models of incentives and economics and how some of these benefits are gonna be shared? It just feels like we are just there, but we are not able to see that picture just yet. So that's I think a very important piece of it. The second is these technologies are still very much maturing. We are seeing like, yes, reasoning models are gonna have a lot of importance. Models that are gonna be able to do specialized capabilities are gonna be important, these communication mechanisms are gonna be important. But it still takes, like, it's still early to build something that is just gonna work in every single domain. But being able to specialize on a vertical and show something working in a particular scenario where some entrepreneur can go and understand what the current workflows look like and be able to integrate these technologies into flows and processes that people are already familiar with, has a much easier time for adoption than changing everything. So that would be the second thing I would recommend. Like, is there a particular domain where the entrepreneur has a unique understanding of the processes and problems and the gaps and can use that know-how to be able to really like build something that work?

SPEAKER_08 50:07

The advice that you give entrepreneurs is uh quite phenomenal because for the first time, generalist, right, your understanding of the space and then the workflow that you describe, right, within that space becomes as important as the depth of understanding of saying, I have this much knowledge and I can go this deep, because some of the depth could be solved by these foundation models. You don't have to solve for, but the foundation models, if with the agents can connect, say a workflow or the way uh things move within a workflow, or how people make decisions and how interactions should happen, that could be as powerful for an entrepreneur. Seems like the generalist is as important as the specialist. It used to be that you had to be really good at say database, or but now you almost know how things work. It's very, very fascinating. Thank you for listening to our podcast from Startup Biggsit, brought to you by Dai Seattle. Assisting in production today are Isha Jain and Mini Varma. Please subscribe to our podcast and rate our podcast wherever you listen to them. Hope you enjoyed it.