#497 From Startup Grind to Mozilla Labs: Raj Singh on AI, Vibe Coding, and Building What Matters

[00:00:00] 

Mehmet: Hello, and welcome back to a new present of the CTO Show with Mehmet today. I'm very pleased joining me from the US Raj Singh, Raj is the VP of product at Mozilla. But in a way, as the audience know by now, I like to give it to my guests to give us input [00:01:00] about, you know, their background, their journey, and what you're currently up to.

So Raj, thank you again for being here with me today. The floor is yours. 

Raj: Uh, sure. Um, I'm Raj, um, product at Mozilla. My journey, um, I'll try to keep it to a minute. Uh, it's been, uh, a series of startups beginning with, uh, uh, dropping out a master's effectively in 2000, or not finishing at least. Pursued my first company.

Uh, never really looked back. Um, I've had, uh, multiple startups since then. What I would say is, um, half of them failed and half of them did. Uh, some, some, some did well and some did mediocre. Uh, landed at Mozilla through an acquisition of my last company that was focused on meeting summarization. Uh, this is kinda like pre ChatGPT era.

Uh, I know it's a very hot topic. Uh, and now, uh, my focus is on new products at Mozilla, so I've been working on. Uh, a variety of products in and related to small business. Um, so whether it's, uh, setting up your [00:02:00] website, invoicing, accounting, growth, all the sorts of different facets that are involved there, uh, everyone wants to be their own boss, ownership of the economy.

So it's kind of a, you know, the freelance economy, et cetera. So it's a, it's a, uh, it's the fastest growing business, at least in the us. Uh, the side hustle. Uh, and so we're trying to help, uh, enable and support them, uh, otherwise, uh, uh, interest, uh, span, you know, product management, gen ai. I've been at that intersection for a while.

But yeah, that's, uh, that's kinda my, my quick two sense about me. 

Mehmet: Great. And thank you again Raj, and what a fantastic journey. I would say we have a lot to learn from you today. Um, so starting, you know, a little bit from what you mentioned about, uh, starting and exiting. So you've built multiple, uh, startups from mobile browsers to recipient communities to calendars, um.

What's your personal pattern recognition when something is working? Because you said like some stuff worked, some stuff didn't work. [00:03:00] What do you think, you know, today in, in, in this gen AI era also as well? Um, some of the things that we need to focus as, of course entrepreneurs want to start something, how we can measure what is working really.

Raj: Yeah. So, uh, in my opinion, um, companies follow S-curves. Um, and, uh, if you kind of think of an s it kind of goes like that. And within what, what they don't really think about is within that scur. There's like thousands of little kers, right? Uh, and so, uh, sometimes you hear terms like product market fit, uh, you know, um, have achieved, uh, uh, hit, hit, hit a certain, uh, LTV or hit a certain, uh, a RR or whatever.

But the reality is, I. Like you're going through another form of PMF, uh, you know, every three months, every six months, it's a different set of issues, right? And so, at the very beginning, at least for me, thematically, I tend to focus, and again, this is me, uh, and, and different [00:04:00] entrepreneurs. I. Uh, uh, you know, resonate differently and do it differently.

Uh, uh, for me, I focus on personal pain points. So I've always been in the bucket of, um, is this something I'm experiencing that I wanna fix for myself? Uh, uh, you know, uh, how can, how can I, how can I resolve this for myself? And if there's me, then there maybe 10 of me. I know a lot of people, uh, uh, caution against that sort of advice.

Uh, but most everything I've done has been consumer. Uh, and thus, why? Uh, if you look at my background, it spans, you know, ringtones and music to live video, uh, to, uh, uh, recipes to calendar. These were all various pain points each have their own sort of story arc in each of their own sort of times. Uh, how did I know if things were working?

You know, you don't, this is like the hardest question. What I tell most, uh, founders is. There's, there's like a few percent of startup maybe let's say like startups, let's say like less than 5%. They're just rocket ships. Like, like if you get on one of those trains, like you're lucky you got there [00:05:00] and just hang on.

Like you're growing no matter what. You know, you joined a PM early in Google or Facebook or whatever, it didn't matter what you did. You could be the worst morning pm The company was still growing gangbusters and so just amazing to be there, right? Uh, and then there's like 10, 20% of startups that really.

Are just totally failing. Like it's very clear from the, you know, from the outside, uh, and honestly should shut down. It's a difficult situation and or it has to, you know, the founders have to have the perseverance to say, I wanna pivot this or move in a different direction, whatever it's. But the vast majority of startups are somewhere in between.

And it's super hard because there's no clear signal saying this is working. Maybe you, you have figured out like, okay, I have a certain NPS score, uh, and, uh, you know, we're acquiring a certain amount, uh, organically in our blended CAC has come down, you know, quarter over quarter and, and we're retaining and what, whatever it is, who knows?

Uh, you know, or generating a certain amount of revenue or a certain amount of engagement or session time, but like, you just don't know, everything is sort of arbitrary in some way. And then you hit the next issue like, well, can I [00:06:00] make this repeatable? Can I scale this? Can I go faster? Uh, as it, as it become linear growth.

'cause in startup land, linear growth is effectively failure. So, uh, you know, uh, as I said, the vast majority of startups are in this bucket. And, uh, I didn't know in each of my times. Right? And if you look at why a lot of these, uh, startups ultimately fail. Um, I don't think it's necessarily, they didn't crack that code.

Maybe they got to a point, they're at 2 million a r or 3 million a r and that's great, but they're like stuck in this purgatory. It's because the founders get tired. Uh, right. You know, like you're doing this for five years, you're grinding. I mean, you may not be at work all the time, but you're thinking about work all the time.

And, and that really, uh, uh, I think that's really the number one reason, uh, outside of just simply running outta money. 

Mehmet: Right now I gotta ask you something which is related to AI because you know, we're gonna talk a little bit about ai, uh, here today. And you mentioned you've built like AI products before, you know, um, you know, chat g PT WA was [00:07:00] something and you know, it was so cool.

What have changed, do you think, you know, and is it, do you think easier now for building product from scratch, leveraging the different AI tools, whether, you know. From ideation to validation to even writing, you know, the m you know, the code for the MVP or Vibe coding. How have you seen this? Is it like getting easy?

Is it getting, I mean, of course it's getting easy, but is it like more challenging from differentiating yourself in the marketplace because it's available for everyone? 

Raj: I think things have, uh, uh, rapidly changed. Uh, you know, my first, uh, attempt at some AI stuff was like, I think 2000. Four, 2005. Um, I was building a, uh, desktop app to do, uh, photo classification, uh, by face.

Uh, 'cause I had all these photos I was collecting and I was like, Hey, it'd be neat if I could search for just garage or my brother or my mom or my dog, or whatever. [00:08:00] It's, and, uh, uh, I had to use open cv, uh, and everything is c plus plus. It's just much harder. Right. And you don't have the, the cloud hasn't really formed, the tools weren't there, you know, fast forward.

Um, I had a company I spun outta Stanford Research Institute called Tempo that was acquired by Salesforce and we were doing, uh, meeting preparation, like it was a smart calendar. And in that case we were doing semantic analysis and we were doing, uh, entity classification or named entity recognition, whatever you wanna call it.

Uh, ranges of things like that, like building sort of an ontology of you and your data. And then we could. Surface and associate and, and just, uh, show things, uh, related to your upcoming meetings. Um, it wasn't machine learning, so to speak. It was a bunch of rules and heuristics. Uh, machine learning was hard.

I mean, it was expensive. It was compute expensive. Uh, it, uh, required, uh, multiple cycles of training. It just wasn't there. When I look at like my last startup doing meeting summarization, we were finally there, right? We could train our own model annotation [00:09:00] pipeline. This has got pre error, but it was still foreign to a lot of companies and many companies doing any sort of machine learning.

We're looking at very narrow sort of cases, right? And uh, and now fast forward, we have generalized AI and we have it in the cloud and it's available through like two lines of code. And this has been. The most dramatic, uh, disruption of productivity I have seen in my career. I mean, certainly the internet and mobile are, are up there as well, but this is just amazing and I think, uh, you know, the way I think about it is I.

Usually like when there's a new technology. So like back in 2008, I remember on the app store there was a location category and these were apps that used, uh, location. And uh, it was just kind of silly if you think about it because like what app doesn't benefit from location? And so we kind of see how we went from location as a category to location as a layer.

Uh, the same thing is true now with ai. You know, uh, if you look at pre gen [00:10:00] AI era, there was actually a category like in a market map that maybe an investor would put together that these are AI companies. But now it's like, what company isn't using Gen AI at some point in their workflow? It's almost silly to say, oh, this is an AI company unless you're doing just deep research.

Um, and so, uh, the fact that, uh, we're now at this point, um, you know, and we have fractional experts, which is basically Jenny High at our fingertips, I think you're seeing, uh, rapid. Uh, rise in what I call vibe creation, whether it's programming or whether it's like writing or content creation or video, or you know, an architectural diagram or an interior designer.

And it's really unlocking, uh, the massive creative energy. Uh, across all of us. Uh, and so therefore you're just having much more creation, uh, uh, which kind of, uh, points at what you talked about with the signal and the noise and just the volume of things being created. But, but I, I just think this is just the, this is, you know, people talk about Neuralink, I think of this as neuralink.

It's Neuralink without [00:11:00] needing to put something in your head, right? Like you're getting the, you're getting the, uh, the neural linkin and software. 

Mehmet: That's interesting, uh, point of view, Raj, now beside, you know, designing the product like coding and so on, how are you seeing also the AI impact on something like, you know, what we call growth hacking and user acquisition.

Um, like do you see it like overused? Do you see. Do you think like there is still room for, uh, improvement on it? Or we can say, because you know, I'm seeing a lot of people talking about now, you know, relying on AI in some startups completely to, to, you know, have this, I would say engine for customer acquisition.

While some people they're saying, no, it's so early, we can't a hundred percent rely on ai, so let's slow down things. Let's. Keep doing it to the way we were doing before. What are you seeing from, from, you know, what you see today? Yeah, so 

Raj: I [00:12:00] would say I, so sort of many points. Um, I think, uh, certainly the, the role of the people's roles are changing, right?

If you think of a, let's take an engineer because it's just kind of an easy example. Historically, your writing code, your testing code locally or whatnot, um, you're reviewing code, uh, and your merging code and resolving conflicts. And you, you might be writing some technical specs and things like that, but that was loosely, you know, that that was kind of, and obviously communication and whatnot, but that was kind of your broad sort of role.

Their day-to-day is shifting now to not necessarily writing the code, but writing the instructions to generate the code and then reviewing the code that AI is generating. To then catch errors or mistakes to write better instructions the next time around. They're still involved in obviously creating the initial architecture, uh, but then AI is generating the scaffolding.

They're still involved in writing, maybe like the requirements, how to approach this, but then AI is then filling in those. Uh, gaps or thinking about those edge cases. And so this is a dramatic shift [00:13:00] in how people work. Uh, and this is happening everywhere. Uh, you could think of, let's say just a, a writer maybe.

Historically, they write all of their own articles, but it's not uncommon now to sort of, I. Verbalize all your thoughts maybe to an ai, see, I'm thinking about this, I'm thinking about that. I wanna make this point. Here's some interesting quotes, blah, blah, blah, whatnot. And then AI writes the first draft, and then you review and edit, uh, and improve upon that.

And then AI may say, oh, no, that's good. And then you can ask AI to simulate how people would respond to this. And then you make further edits and then you submit and publish. And that, that's, again, that's a dramatic shift in the workflow. And so I think, uh, I think we're seeing, uh, we're seeing this, uh, you know, across the, across all sort of occupations.

And then, so I sort of see, I sort of feel, uh, uh, this is reinventing, uh, everyone's sort of work and uh, you know, to your comment like, what am I seeing at least with. Uh, gen AI in terms of growth, [00:14:00] in terms of other sorts of, uh, functions. Well, now that we've democratized access or given everyone access to these fractional experts and unlocked all this sort of creative energy, I.

Um, you're seeing all kinds of things being created and, uh, you know, from a growth perspective specifically, you know, this is a category that's always been very heavy ai. If you think about where AI was probably used the most historically from a tech perspective, it was ad serving, uh, and the ad ecosystem, right?

So there're always, there're always been playing with AB testing and how to improve different sort of outcomes. But, uh, you know, I can speak to it Mozilla. You know, we've looked at, or we've built, you know, tool, vibe, coded tools, uh, to help with, for example, uh, making it easier to publish a corporate blog.

Um, so if you need to publish a bunch of posts because you know, we shouldn't have to human author all of those things and any of these companies. Anyways, we're outsourcing these to basically human writing farms, so they're pretty much already low quality content. Uh, but like we, you know, we've kind of reinvented that workflow.

Uh, we, we looked at, uh, what now people call ai, SDR, like cold [00:15:00] email drip campaigns. Like how do you generate the first email If you have a lead, how do you send it? How do you follow up? How do you personalize it? These are things that you could rapidly vibe, code. And get like 80% of there away there. I don't see the human in the loop going away right there.

You know, I have this sort of rule with ai. I think of the, I call it the 95% rule, which may or may not be accurate, but basic, the general, uh, uh, contention is that AI is 95% accurate. It's good to use as a litmus test. Maybe it's 98, maybe, whatever, but, but given that it's, let's say 95%, how does that affect how you approach the problem?

Or how do you, uh, approach the user experience if, uh, a good example is. If you're thinking you're gonna rely on AI to schedule all your meetings, but it gets it wrong like 5% of the time, that would be unacceptable. Right? And the same thing is true, for example, with the self-driving car. Like if it dries, but 5% of the time it makes a mistake, that's unacceptable, right?

So there's certain applications and workflows where you need 100%. Uh, and so in that case, uh, how do you bring a human in the loop so the AI is doing more of an assistive role or maybe doing the [00:16:00] foundational work or whatever it might be, right? So it's just a different way of thinking about how you approach problems and where the human might play.

Mehmet: Cool. Now speaking about, you know, this shift also, Raj, and because you mentioned about you developing tools that helps, um, you know, freelancers, solopreneurs, um, I've seen, you know, AI is allowing everyone. To kind of building their own businesses much easier than before. Now the question is, if we have this community of builders, let's call them this way, or like solopreneurs.

Um, so are we going to have really what a lot of. You know, leaders in the ai, like some Altman mentioned it and some other people like about the, uh, unicorn, uh, one person company. Yeah. Is this something which feasible in the near future? I mean, maybe in one [00:17:00] year time? Or is it like kind of a long vision, um, that we still need to develop a lot of other AI tools or leverage what we have done so far to make sure that we can have the full fledge, you know.

AI powered company, whether it's Agentic AI or maybe some other new concepts that allows really one person to run the full thing by themselves. So how far, like my question is how far still we are from, from really having, you know, these. Massive businesses running run by one single person. 

Raj: So I, you know, when I hear those comments, I think of it as a very aspirational kind of goal.

Yeah. Uh, you know, can I, uh, uh, can one person build a billion dollar company? Um, and I think it's fine to think like that in tech. You know, the being working in tech and being in the tech community, we. We like to think in hyperbolic ways. Like, oh, we can get to a point where we can replace everything we can.

We can automate everything. We can automate, you know, the HVAC repair, which I, [00:18:00] I assure you is not happening anytime soon. Uh uh, you know, uh, so, uh, uh, no, I don't expect, uh, you know, multiple billion dollar 10 person companies. However, the friction that's involved to get there. Uh, has come down substantially.

No different to the laptop, no different to the internet, no different to mobile, uh, which is amazing. No different to the cloud, you know, enabling it, making it easier to build. Um, and then two, uh, having fractional experts at your fingertips means the friction to get started, you know, way easier, right? And you can automate some of those functions, but are you going to entirely rely on your chat GPT attorney, uh, to draft all your legal documents?

No, you're gonna certainly use it for perspective. You're gonna get a first. Form or first draft or whatever, but you're still going to go and speak to a real expert because there's a number of things that need to happen, uh, and assurances you need where you just simply can't be wrong, right? You can't, you can't be wrong with 5%.

And then on top of that, there's still all this other stuff [00:19:00] that just has to be done, whether it's operations work. Sure, you can automate some of your customer support sometimes. Then you get on the phone and your voice bot isn't gonna do well enough or maybe, uh, they're sales and you have fine, you have your AI sales chat bot.

Uh, that's sort of answering questions, but then they wanna talk to somebody physically or like, whatever. It's right. Now. Do you need as many people, uh, as you may have needed before because some of the operations can now be done by others? Uh, probably not, but at the same time, I. Whatcha gonna do with that extra revenue that you're now making, you're just gonna go faster, you're gonna build more.

It's not like you're gonna stop. Right. I think the, I think the fundamental issue is just we're all thinking too small. We're thinking like billion dollar companies. When you start needing to think about trillion as the new baseline, and we start thinking about trillion as the new baseline, then you're like, well, we're just gonna need 10 times more people.

Right. Even with the ai. Right. So, uh, um, I, my, my general opinion is. Capital creation and just productivity is just gonna go off the charts. Uh, we're sort of seeing 10 xs [00:20:00] across individual contributor functions in organizations everywhere. Uh, certainly different flows are, um, getting assisted or augmented with, uh, agentic, uh, agent cap, you know, agentic internal tools or what, whatever it might be.

But that doesn't mean the companies are gonna do less. Now the company's gonna do more. They have product roadmaps to take over the world. Now that certainly means that, um. You know, signal noise, who do you trust? It's becomes a more competitive environment. But, but, uh, all those things are still gonna matter.

But, but, uh, uh, I think this is like single-handedly probably the best time in my career, uh, to build. 

Mehmet: Absolutely. It's, it's the best time to, to build. Um, going back to what you're currently doing at Mozilla, um, uh, Raj, so you're leading what's called the zero to one initiatives. Right? And usually when we hear the term zero to one, of course, startups comes to mind.

Now, how, as a product leader do you [00:21:00] approach the zero to one in. A very, you know, well established legacy company like Mozilla, like how you can apply what you've done in different startups back in the days into that setup. Uh, and I'm asking this because, you know, one of the things that happen usually is a lot of startups, you know, they get acquired, right?

So by bigger players and. I've read like some stories where, you know, this zero to one concept can be applied easily in bigger setups. So what can you tell us about what you're currently doing, how and how you are doing it at Mozilla? 

Raj: So I'll share, um, what I can, and I'll also add. Um, you know, zero to one has many meanings, like new product development differs from company to company.

I'll start with a macro statement. In general, [00:22:00] large companies suck at building new products. Um, and I don't say that from experience, meaning I've not worked at very many large companies. I say that from anecdotal data that I've gathered in talking with a number of, uh, companies that have various sort of zero to one or new product sort of initiatives.

And, and the question is why, right? Like you take like a labs group and a Fortune 500 company and like why isn't it working? And uh, and the real trick here is figuring out one, can you create the necessary changes that are needed to sort of address those whys? And I'll go through those. Um, and, and is, is there enough conviction and team and culture and board that's on board to do that?

Um, for a long enough period, right? So let's go through some of the reasons. So like, uh, one, a lot of the pe you know, a lot of new product development, um, means you need to hire, you need a team, uh, that's comfortable in am [00:23:00] ambiguous amounts of data, right? You know, uh, if you think about different archetypes of roles that people hire within an organization, um, a lot of mature companies hire operators, people who know how to extract another 1%.

Run an AB test, get another 0.2% here. Um, decisions are data driven. Uh, in many cases they're more consensus driven because they're driven off data. Um, when you're dealing at zero to one, you don't have data, right? And so you have to sort of, uh, make decisions because you have strong conviction and you don't have consensus.

You have 10 different opinions around the room and people like, I disagree, or whatever. And you have to be able to kind of plow through that, which is really, really hard and very atypical. For folks who primarily work in operator, product management or other kinds of roles. Uh, right. And, and one of the reasons it's so hard is it feels uncomfortable because you now actually are accountable to a decision without data, with people with different opinions.

And then what if it fails? Right. So, which kind of leads to my second comment is [00:24:00] failure. Like failure is not, uh. Uh, uh, when you, when dealing with mature products, there's so many strategies to deflect failure, right? So it's like, well, the data suggested that we should do this, or we used a series of outside consultants and they recommended we do this.

When you're dealing with new products, like you're pretty much accountable to your decisions and your failure, and you have be totally comfortable with that. And most people are not comfortable with failing because the way the incentive structures are created, uh, in large companies, it's. You know, they almost mask this, right?

Because it's about like being on the hot product and growing at the right time and all these sorts of things. And so you don't want, like you, you've never really experienced this. And this is something that's very different than startups because you're failing publicly all the time. And so a lot of people are just not comfortable, uh, in that kind of role.

And this kind of leads to. You know, an orthogonal topic, which is just the entire career level guides and how you think about that with new products and zero to one development, totally different. Right? Right. So let's go to like number three. When you think about, uh, mature products, the [00:25:00] people you hire tend to be very specialized.

You might have, uh, an SEO team, you might have. Uh, a go-to-market team. You might have a product marketing team, you might have a market research team. You have your pm, you have your program management team. You might, you might have engineering, you have testing, whatnot. All of those sorts of functions are kind of rolled into just a single pm uh, in the new products team, which is I.

Which is much more of a generalist. It's more of like, I have a breadth of experience across a range of things. I'm a super curious person so I can learn and figure it out when I, when I need to go super deep. But at the same time, like I'm not like the top 1% expert on anything. Uh, but I can certainly learn faster than most people, and I, and I can do that.

And that's very different to the types of folks that are often hired within. More mature products. Typically they're hiring people who've done the same thing for many, many years. Right? Maybe they've been, uh, senior, uh, leadership, product management leader, whatnot for many, many years. Uh, and so they're really good at organizational design, right?

And [00:26:00] that's like, that's kind of their expertise. Or maybe they've been, I. Working in a particular domain for a very long time. And so they want them in that very specific domain. And so those sort of things go sort of out the bucket, right? So these first three things I just talked about are all related to talent, right?

So then you have organizational things, right? So like number four here, um, legal administration and not administration, administrative, uh mm-hmm finance, all that sort of stuff. Those things can massively detriment new product. When you think about new product and startups, we have to operate in gray areas.

There's not a. Startup in the valley that didn't, at some point do things that were slightly questionable, like whether that's cold email marketing, whether that's, uh, you know, uh, some tactics that may be a little bit frowned upon. But these were necessary to drive growth. And then once they achieve that certain growth, obviously, uh, they change and they become uh, you know, uh, let's say more, more comfortable or whatever you wanna call it.

But until you get there, it is, it is a lot of hacking. It is a lot of [00:27:00] sort of gray area decisions. You know, there is a lot of like little tactics around your drippy mail or whatever it is. Who knows, right? And when you have a mature company. It's very hard to sort of deal with this, right? You have legal saying, no, you can't do that, and legal's creating roadblocks.

And so my sort of fourth point here is, and again, not all companies can do this, but can you adequately create some kind of firewall? Where your teams can go use whatever tools they want without having to go through legal review because they just need to move fast and try a lot of things. Your teams can expense whatever they want because they need to try and test things 'cause and that's not typically how things work.

When you deal with mature products, they run through processes and they have to get reviewed and they have to go through security reviews and all these sorts of things. And your teams, if they need. You know, to do things that are a little bit more gray, that's fine, but maybe like we figure out a way to shelter the brand.

Uh, right. And the way legal in terms of services and privacy policies and all those things are created, are much more form and more generalized, uh, as opposed to narrow. Right? So that's kinda like the sort of fourth bucket. The [00:28:00] fifth thing, which I think is part of the reason that a lot of these, uh, new products fail within many organizations, um, is what I call the conviction window.

Right? And so you typically have a, you'll have some leader. And, uh, they support these sort of initiatives. And then there's turnover in leadership or organizational design. And you lose your kind of sponsor, right? And you're no longer the shiny object. They've moved on because startups and building new products just take a long time.

You know, the average, uh, time to liquidity in the valley for a venture backed, like a seed stay startup is at least seven years, right? Um, so like not a lot of people are on for that ride. And if you look at a lot of these great growth stories, like, oh, look at them and look at that. You know, there was a grinding for five years.

I mean, here's lovable hitting. Or Bolt or one of them living, like hitting like a hundred million a RR. But there was like five or seven years before that where it was like a zero, right? Where they're just like trying to figure this out and then they pivoted. Look at Slack, you know, look at so many of these success stories went through such long grinds and many large companies don't have that kind of patience 'cause they're so focused on immediate dollars [00:29:00] that this conviction window is short.

And so there's some tactics on how to work around that. Um, you know, and I think, you know, I think the sort of the sixth thing. Um, you know, in a lot of companies building new product, build it as an adjacency. Um, so you're basically saying, here's my core product. I can message to this audience, I can upsell to that audience.

And that's perfectly fine. But for some companies, these are truly new products. It's new go to market, it's new audience, and a lot of companies don't know how to deal with that, right? They have a, a, a marketing team or a product team that's always had users and now they have to start with zero. And like, how do we acquire new users and how do we think about new audience and all that?

So. You know, these are ranges of things. I'm probably, I probably could speak on this topic for another 30 or 45 minutes, but, uh, there's a lot of reasons why new product development is so hard in large companies. Now, I've been fortunate. I've, I've worked through multiple incubators and accelerators. You know, I'm running this, you know, uh, at least for small business within, uh, Mozilla itself.

And I've had multiple startups, and so I've got this kind of rounded perspective. I feel like, um, uh, uh, I can see where, where things are [00:30:00] breaking down. Uh, but, but I think for many, uh, leaders that try to enter this. Um, they come in fresh and they're smart people, but the thing is, they, they try to operate like it is there, uh, even when they're trying to be intentional about it, because they just haven't really seen how scrappy the other side is.

Like, we didn't even talk about incentive structures, right? How do you do payouts? How do you do promotions? It's totally different. It shouldn't match the, the guides on the other side of the company, right? So, various reasons. This is kind of the quick, uh, you know, the, the quick two sets. 

Mehmet: Absolutely. It's make a lot of sense.

Uh, Raj to me, and I think we discussed it with, uh, uh, you know, multiple guests on the podcast, usually I tell people to understand why big companies, they don't usually, usually I'm saying innovate. So there's a great book called The Innovators Dilemma by Clayton Christensen. So it's a good book that gives, you know, people understanding why this happens and.

You know, the author gave some tips, let's say, like, even if you are like [00:31:00] the well established company, how you can try, um, to, to, to create new markets or like, let's say, uh, try new ideas, which, which is a great book, but, uh, to your point, you know, this is, this is the true story of, of every startup and how, you know, it can become in the future.

You mentioned something about pivoting, which is. I tell people, you know, pivoting is one of the, sometimes let's say destiny, it depends of course, on market conditions, depends on the idea they're building. But people, majority of them, they say, we are afraid of pivoting. Like, because we don't know if we should wait a little bit more or like, we should really go and pivot.

What do you think, you know? A main kind of, I, I would say it's multiple things. Multiple points I know, like, but what the major points that they [00:32:00] need to consider as signals that need to push them to iterate on the product or maybe even on the whole, uh, business. 

Raj: I think the answer to this question depends also on the stage of the company.

Like if you're a startup and you're a pre PMF and you're pre-revenue. The, the, the sort of penalty, uh, or sunk cost, or whatever you wanna call it for pivoting, is substantially less than your company making, you know, 2 billion a year, uh, mature a thousand people or 5,000 people or whatever it is. And, uh, uh, now, you know, it's a much bigger ship to steer, right?

It's not a, it's not a motorboat, it's a aircraft carrier and it's hard to turn if you get more people aligned. Um. I think, uh, the other thing to call out here, people think of pivots, they think of product pivots, but pro pivot, pivoting doesn't just mean product. You might pivot your go to market. Mm-hmm.

You might pivot your target audience, you might pivot your business model, right? There's many forms of pivoting. Uh uh, so I think it's, uh, you know, in many cases you built the [00:33:00] right thing, but for the wrong user or you built the right thing. But you're not the right model, wrong model. You're not charging enough and you can make more money.

Right. Google didn't know, uh, that ads was the model. Right? Uh, they probably, maybe initially thought maybe it'd be, um, we charged for it in some way. I mean, I don't know. Right. Or it'd be sold to companies or something. Who knows? Right. So, um. I think, uh, um, I think now to the core of the question is when's the right time to pivot?

Now you hear a lot of stories about startups or founders, um, spend too long maybe on an idea, uh, and you're like, man, they should have pivoted sooner. Or, or maybe they pivot, they're pivoting too much, and then they run outta money. And now, uh, they've exasperated their team, each pivot. Is not only, uh, physically, you know, it's, it's not only physically exhausting 'cause you gotta kind of run through it, but it's emotionally exhausting, right?

And that exacerbation can really hit teams. And so I think, uh, you know, I think the, uh, um, the way to think about this and what, and what I often tell people is I used to have strong opinions here. Like, when's the right time or wrong time to pivot? [00:34:00] Now I ultimately say it's really up to you. Uh, you know, you gotta obviously look at all these sort of different considerations.

And the reason I say that is I've come to realize that. I've seen every scenario at this point work. Like I've seen people, uh, do something. I'm like, man, they should pivot. They go seven years, then suddenly it starts working. They cracked it, the market shifted or whatever. Right? I've seen examples where people are pivoting every month.

Uh, classic YC example companies, uh, and then like suddenly they found something two years later, which is insane to me too, right? So I. Uh, I've come to the point now where I'm like, I don't know anymore. When's the right or wrong time to pivot? I certainly have my own rubrics, uh, where I think about like, huh, this doesn't feel right.

A lot of it's gut call. It's not something AI's gonna answer for me, but I think the key is being as informed as possible and making sure you have as much data as possible and making sure you've done enough repetitions, um, so you can kind of see it, right? Because at the end of the day. Uh, pattern matching is a big part of this game, but I would also say there's plenty of [00:35:00] anti-patterns, uh, that turned out to be crazy success stories.

Mehmet: Yeah, and to your point, like sometimes, um, we call it the gut feeling slash the data, so, um, it's a tough decision. But yeah, like, uh, I think it's becoming more and more complex or let's say, um, you know, critical, uh, moment that the founder or the founders. They need to, to, to think about. Now, shifting gear a little bit, Raj, and you have raised before capital, um, for, for, uh, for your startups.

And usually it was considered kind of a. Like minus point I would say if you are a solo founder or let's solopreneur, whatever you want to call it, especially in early stage products. So have you seen this? Um, you know, I. [00:36:00] Kind of becoming more acceptable. Uh, is there any changes, especially with the AI wave that, you know, started it like with the chat GPT and all the other tools that now VCs are accepting more and more the idea of solo founders building great products?

Raj: I don't, so I don't have the data, so I don't know. Um, if the number of solo founders, uh, raising venture money has gone up or down. What I will say is VCs have always invested in solo founders, but it was less common, but not for the reason. People think it was less common because statistically solo founder companies did worse.

Why? Because founders get tired. Um, and founders being a founder is probably. One of the most lonely jobs you can imagine. And the reason is nobody's [00:37:00] totally honest, right? You're, you're, you have your board and you need to, um, you need, obviously you wanna authentically share your information, but you're certainly spinning it in a certain direction to exude a certain amount of confidence, uh, to show that you have an idea of where we're gonna go, uh, and what we need to do.

And there's a strategy and not all is lost, right? You have your team. Uh, that some members of your team can handle the emotional swings of startups where you have like super high days and super low days, but most people can't. They're not built in that way. Right. And. It's not a, and just to be clear, when I say not built in that way, I don't mean that as a negative.

It's just, you know, we're different archetypes, we're different personalities. We have different strengths, right? And so some people don't do well with like large swings and that's just not the way they work. And so, uh, and so you don't, you need to shelter that information, right? And so when you have your all hands on a Friday or team gathering or whatever it is.

You have to, you know, present in a certain way. And so you're authentic, but you're spinning in a certain way, right? Um, you know, when you're talking, uh, when you're at that CEO [00:38:00] dinner that your law firm is putting on or your bank is putting on, right? And you're talking with other founders, you don't wanna be totally honest.

I mean, there's a risk that, uh, you know, uh, information spreads and it affects your future fundraising, right? This stuff happens all the time, right? So you have to like, spin it in a certain way, like, yeah, things are going fine. You know, we're growing and we're trying to solve this. And sometimes you might be a little bit vulnerable, but it's guarded.

Right. And so you start going through this and you're like, wow. Like who can you even have your own family? It's like, Hey, Raj, you're like always working or thinking about work. You're like, never like focused on me. I could tell you're thinking about something else. And like, well, no, no, things are going fine.

You know, like, it's like who can you actually be totally vulnerable with, right? Uh, and so, uh. The challenge is the job itself is so difficult and so lonely, and you know, founders is very different to a hired CEO. Just to be clear, I mean, the hired CEO is a stressful job, but hired CEO doesn't have the same gravitas because in many ways hired CEO can blame the company if it doesn't work out.

But a founder. Gets [00:39:00] all the, you know, if it succeeds, everyone takes credit, but if it fails, they blame him, him or her. Right. And so, uh, uh, uh, I think, uh, um, uh, given all that, uh, you know, I'm trying to, I like lost track of the original question here, but, uh, uh, you know, I, I think this, oh, the solo foundry, and I think, I think that is why.

Uh, you see a bias towards investing in group founders and so much a venture. You know, they are looking at data and statistics, but again, as I mentioned, uh, you know, a couple earlier in the call, uh, earlier in the podcast, um, uh, there's plenty of anti-patterns that buck that trend, right? Right. So nobody really knows.

Mehmet: Yeah, so, so I think, you know what I've been noticing, it's completely fine being solo founder. As long as you surround yourself with people who can help you in the things that you know, you, you can't keep doing it by yourself, the things that you can't scale, right? So, so being able to delegate some of the stuff, like [00:40:00] you don't need to go and do every single thing by yourself, even.

Maybe they are not co-founders, but kind of supporters. Let's call them this way, that they can do the, let's say the nitty gritty things, right? So, um, you don't need to go yourself and do everything from building the product, sales, marketing, everything, unless maybe it's something. Which is maybe it's a kind of a product led, uh, uh, product led, uh, uh, kind of, of product that yeah, people can go download or it has network effect.

Yeah. People gonna go and discover and so on. But yeah, to, to, to your point, uh, I, I, I think this is also what I'm seeing now. And I, I'm sure I, 'cause I follow, you know, the, the VC landscape both in, in the US and in, you know, the me region where I, I am based. So seeing the same pattern over there. Now, going back to, you know, having kind of this your, I.

Uh, expectation or let's [00:41:00] say vision to the future. Where do you think there's still, um, bunch of opportunities in consumer AI beyond just like productivity and creativity tools? Like what, what are still some of opportunities, uh, over there in consumer 

Raj: ai? 

Mehmet: Yeah. 

Raj: You know, I think I would answer that differently.

Um, sure. So, and, and, and the reason I say that is I think the minute you inject AI into that, uh, question that you're kind of putting the tech before the problem, I think consumers don't care whether you use AI or not, right? Mm-hmm. They're just a core set of problems, like, can you solve this problem for consumer?

And when I think about. Very successful consumer products, they usually tap into a consumer vice. Um, and so think of like Maslow's, you know, uh, hierarchy, uh, think about, uh, core needs, right? And it is wild, right? How many consumer products are really focused on really [00:42:00] only addressing four or five things?

Like, like, uh, does it help me make, does it make me rich? Um, does it, um, does it help me meet my partner, my future girlfriend, boyfriend, whatever. Right. Does it, does it, uh, uh, um, does it solve boredom? Like, does it entertain me? Right, right. You know, uh, you know, so like, if you start looking at it from that perspective, like where you kind of break it down like.

All consumer products have to kind of map into one of these buckets, right? Uh, then, uh, uh, then everything becomes more clear. And then the question is like, okay, well whatever I'm building. Can it benefit from using AI or not? Right? It doesn't have to use ai, right? So let's take, you know, let's take the FANG companies, take Facebook.

It's an entertainment company, right? Right. Like, like, I'm bored, right? It helps me connect with people, see what's going on, right? Like it's an entertainment company. LinkedIn, like, it helps me get a job. It makes me rich, right? Like, you know, like, you know, it's like, you know, it doesn't make me safe, right?

Like, I mean, these are like, like, which has really driven off [00:43:00] fear, right? And so you start mapping back to these core vices. Uh, you know, like, um, uh, you know, Netflix Entertainment, right? Like consumer products always fall into these buckets. Uh, uh, at least, at least when you're talking about the unicorns, right?

You know? Um, and so when I don't see, when I see a new consumer product, and I can't quickly map it into like some kind of core consumer psychology need. Then I start asking the questions like, okay, how are we gonna drive growth? Right? Because word of mouth becomes a lot harder because there isn't this sort of, uh, motivation to go and then just install and like, activate and try this app out, right?

Like, uh, but if you really wanna tap into like high motivation, then you have to play on one of these coordinates. 

Mehmet: Absolutely. Make, make completely sense. And to your point, like, uh, I like when you said, uh, you don't need to tell. People that it's ai, like you need to just be solving a problem or like adding some value to them, um, without mentioning the ai.

So I love this because [00:44:00] people, I think they are doing this mistake. I see it more in B2B, you know, not in in the consumer space where, hey, like we have this AI power. X, Y, Z and they tell people why you, why you feel the urgency to mention the AI if it's not needed. And they said, yeah, it's just the hype. So we're not, we just need to, to follow the hype.

I, but I think it's also applied in the B2B, because at the end of the day, what they look at is, are you solving a real pain for them? Uh, what value are we adding? Maybe you can go into the AI piece later on, but in my opinion, it's not necessarily, you know, to to mention unless you know. Um, there, there is, uh, a, uh, I would say very, uh, kind of, uh, need to, to, to, to, to mention it.

But anyway, yeah. You, you want to say something, Raj? 

Raj: Uh, I was just gonna say, yeah, I don't think of, when I think of, when I hear technologies like cloud ai, you know, mobile, you know, whatever it might be like, to me, those are just enablers, but they're not really you. Right? [00:45:00] When you find a problem you wanna fix.

Or dress and build a product around it. Then you, then you ask the question like, what's the best, fastest way to build it? What's the best way to build it? You know, how do you create a delightful experience? How do you create an experience that's sticky? How do you lock in data? All that kind of stuff, right?

And then if you need to use those different technologies to make that happen, that's great. I. But, uh, um, I think there is a lot of this sort of what I call putting the carriage in front of the horse, like tech, tech looking for a problem, which, right, which is a super common early stage founder trap. Um, you get enamored with the shiny new, and it happens in large companies too, right?

They're like, oh, 

Mehmet: absolutely. Look at 

Raj: this gen AI thing, like our product needs gen ai. I'm like, does it, um, like what is the problem you wanna solve? Okay. And can I do it? Is really the, the reframing of the thinking that needs to happen. 

Mehmet: Absolutely. And you know, all, all the, you know, the, the studies and, uh, I would say the surveys that are being done today, so customers are looking, they know that actually these technologies.

Kind of being [00:46:00] commoditized. So it became a commodity. So AI is becoming a commodity. Cloud became a commodity. It's not that there's no edge in, in the sense of business edge. Yeah, it's a great technology. But yeah, so everyone is using the cloud to your point, everyone gonna be using the ai. So again, um, great insight here now Raj, as we are almost come close to the end.

Um. I can ask it in a different way. So if you want to advise someone who is planning to launch a project this weekend, what do you tell them? 

Raj: I would probably launch, meaning starting from scratch. Yeah. I. I, you know, I, I think I would ask them like, what are their goals? Like, you know, I think that's the first question here.

Like, uh, is your goal here just to learn a new technology? People are always playing with stuff over the weekend for fun. 'cause they're curious, they wanna play with some cool new, cool vibe coding tool or new design tool or whatever. Right? Um, and if your go, if your goal is just to learn, like go have fun, [00:47:00] whatever, if your goal is to like.

Like build the thing, right? Like, you know, it's gonna be a much longer journey than a weekend. I think you get 80% of the way there, but the 20% takes 80% of the time. Um, but, but, uh. More power to them. Right? And I think, uh, uh, uh, presumably they're pursuing some sort of passion or some sort of problem they've identified, which is great.

Uh, they, uh, either they're the user or they've identified somebody who really needs this and they wanna build it for them. And I think that's awesome. I think the, uh, you know, that person, whoever that is, who's doing that, is already 99% ahead of most of the people out there. And the reason I say that is.

Most people I meet, they tell me all the time, I have this idea, I have this idea. But then they don't really do the second step. If you actually go and build it, you're already ahead of all of these people, right? Most people never get past the idea stage. So, uh, uh, you know, kudos to them. 

Mehmet: Absolutely. Yeah. So this is the first step is like you to go start executing, uh, Raj, where people can find about the work you're currently doing and where they can, [00:48:00] uh, you know, follow you or get in touch with you.

Raj: Sure. Yeah. So I'm pretty active on Twitter. Uh, I read about startup vc, uh, it's under Mobile Raj. Um, uh, startups, vc, product Management, et cetera. Uh, LinkedIn, uh, you could find me there. Um, link Tree, if you're interested in my product portfolio. That's also Mobile ra. And then, uh, you know, Mozilla, obviously Firefox.

Uh, great browser. And then we, uh, you know, one of the products outta my group is solo, so do check it out, uh, if you need a website. Um, so I think, uh, any of those places should be great. 

Mehmet: Great. So I will make the life easy for the audience. All the links will be in the show notes if they are, uh, listening on their favorite podcast app, or if they are watching this on YouTube, it'll be in description.

Raj, I can't thank you enough, uh, for this amazing discussion with you today. Very eye-opening. We discussed a lot of things between AI growth ideas, products, um, VC and raising as a solo [00:49:00] founder. And bunch of other stuff. So really it was an enlightening, uh, you know, episode I learned a lot myself. So thank you for sharing your time and this is how usually I end my, uh, episodes.

This is for the audience. If you just discovered this podcast by chance, thank you for passing by. I hope you enjoyed it. If you did, so please give me a favor and share it with your friends and colleagues. And if you are one of the people who keep coming again and again, you are one of the loyal fans and audience, thank you very much for what you're doing for the show this year.

I know I'm repeating this at the end of each episode, starting, you know, maybe April this year. But without your support, we couldn't rank in the top 200, uh, apple Podcast, um, you know, in different countries. So this is happening for the first time. We used to rank in the one or two countries, but last I checked, were actually in eight countries, which is something I never seen before.

Thank you very much for your support. This cannot happen without. Two, I would [00:50:00] say one, you the audience, and second, including Raj, my guest. I'm very grateful for all the support, for all you know, the time you give me. So thank you very much, and as I say, always stay tuned for a new episode very soon. Thank you.

Bye-bye. Bye