412: Superhuman Mail責任者ラフル・ヴォーラ氏 ― 手作業のオンボーディングからAIネイティブな生産性スイートへ

Where do we spend all of our time? This is a little bit of a trick question. Can you guess the one thing that professionals spend the most time doing? That's not scheduling. It's a very good guess because a lot of email actually is scheduling. It is, in fact, sleeping. We all spend more time sleeping than we do basically any other activity in our life. And after sleep, it, of course, turns out to be email. In fact, there are roughly one billion professionals in the world. And on average, we spend three hours a day reading and writing email. So that's three billion hours every single day that go into email. And that's how I knew it was a problem worth solving. Good morning, good afternoon, or good evening, depending on the way you're listening.

Welcome back to AI and the Future of Work, Episode 412. I'm your host, Dan Turchin, CEO of Peeperain, the AI platform for IT and HR employee service. If you enjoy this episode or any other, you owe it to humanity. Pay it forward. Share a link to this one or anyone you enjoy with just one friend, and they'll know that you care. About making humans better with AI, if you leave me a comment, I just may share it in an upcoming episode. Like this one from Sonia in Rolla, Missouri, whose favorite episode is number 339 with former Slack exec and future of work authority, Brian Elliott. It's a great conversation if you haven't heard that one. About AI anxiety, burnout, and building trust in the workplace. You know what? I'm going to call an audible. There's so many fun facts I could be sharing with you today. But we've got an amazing guest, and I don't want to sacrifice any of our time with today's amazing guest.

So you know what? We're going to jump right into the conversation. Rahul Vora is the founder and CEO of Superhuman, the most loved email client on the planet. He launched it in 2014. Superhuman Mail has now teamed up with Grammarly and Coda as one powerhouse productivity platform. Prior to Superhuman, Rahul founded Reportive, the first Gmail plugin to scale to millions of users, which he sold to LinkedIn in 2012. He's also a prominent angel investor, and his fund has invested $50 million in over 120 companies, including Placer, Superbase, Mercury, Zip, Class Dojo, and Writer, home, of course, of prior guest Mehabib. Rahul received his BA in computer science from Cambridge, which is notably the second oldest university in England, founded in 1209, more than 100 years after the founding of Good Arrival, Oxford. One thing that Rahul has not yet accomplished in his life is having been a guest on AI and the Future of Work that's going to change today.

Hey, Raul, it's absolutely a pleasure to welcome you to the podcast. Let's get started by having you share a bit more about your background and how you got to where you are today. Well, hey, thank you for having me and allowing me to accomplish this bucket list item today and finally coming on the podcast. So in terms of who I am and how I got here, I think you did an incredible job. I can add a little bit of color. I've always been drawn and passionate about making things. And I studied computer science at Cambridge, as you know, started a PhD, in fact, dropped out, which I think is the cool thing to do, and then ended up getting into entrepreneurship. My last company, Reportive, took part in Y Combinator in the summer of 2010. That's when I moved over here to California with my two co-founders and their partners. All in, I think I've moved about 20 plus people to California over the years and sold that company to LinkedIn.

Took a year off or so and then started Superhuman, the fastest email experience ever made. That company grew very quickly. And as you mentioned, we were acquired by Grammarly in the middle of last year. We then renamed the entire company to be Superhuman. And we're now building the AI native productivity suite for apps and agents. And this is an entirely new kind of AI. We're bringing AI to everywhere you already work. That's in every app and tab you might be working in today. And we're rethinking the core application services that you work from, like email. So today, the suite starts with Grammarly, which, of course, is the most trusted writing agent in the world, with north of 40 million users and 50,000 organizations that use it. Superhuman Mail, the company I founded, which is the most productive email app ever made. Coda, like he mentions, the all-in-one AI workspace for teams. And a new product that we're making, Superhuman Go, which is a proactive AI companion that works in every app and tab you use without you having to ask.

And so that, of course, is just the beginning. We have big plans for the whole suite, which I'm excited to get into. Rahul, I had the pleasure, back in 2019, having received a personal invite to Superhuman, which at the time was just called Superhuman, and I kind of on a whim registered to be on a wait list. Of all bizarre things, I got a call from someone on your team to see if I could clear the wait list. At the time, I thought that you were building a rocket ship, a literal rocket ship, and not an email client based on how sophisticated that onboarding conversation went. Since then, you've really become renowned for kind of coaching others through how to find product market fit. Just if you could share a little bit about the backstory, how you thought about the onboarding process and how that contributed to the early success of Superhuman. So there's a lot of ground that we could cover here. Are you more interested in how we define products market fit,

or are you more interested in the genesis of the onboarding program? Where would you like to go? Give us a case study in how you used the output from those onboarding conversations to deliver early versions of Superhuman. Got it. Okay. So to begin with, I was insistent that we do every user, We put every user, I should say, through an onboarding that we actually sit down and we watch them use the product because nothing approaches that in terms of the amount of information you get and how much you learn when you're watching those users. It's actually relatively easy to get people to try a new piece of software, especially if it's in a well-known space like email. I mean, there are any number of upstarts that come and go. Any company can at least try to build an email app. It's, of course, incredibly difficult to build one that people fall in love with. And say, hey, folks, come try this email app.

And overnight, you'll get tens of thousands of people easily to come try your thing. However, imagine what most of those people go through. For most of those people, they will download the app. they'll put it through its paces, encounter a few bugs maybe think to themselves, okay, I'll report one or two things but they probably won't even do that and then as the bugs they encounter increase this is inevitable for early stage software they end up giving up on the product sometimes it's bugs, sometimes it's feature parity sometimes the secret sauce didn't quite land but that's the normal experience of a new user with a new piece of consumer software. That is why churn is such a big problem in any new piece of consumer software. So I've seen this happen with any number of companies. And I knew that that would happen to us unless we did something different. So I don't really believe

in doing a traditional launch at this point. What I believe in doing is putting your products in front of increasingly large sets of customers and then listening to their feedback and tweaking and iterating for them. So we initially did this, and we were initially onboarding as low as five users per week. And I would actually go to the customer's office. I would take a bottle of wine or some kind of gift to say thank you. And deliberately, also at this point, these were also the most influential people you could find in Silicon Valley, well-known CEOs and investors. and I would, this whole thing would take an hour or two. We would sit down and I would ask them to show me how they're doing their email today. So they would pull up Gmail and they would show me how they do it. I would then give them a demo of Superhuman. I would then say, okay, this is the point where you have to pay. And so I would ask them to pull out their credit card and put it in,

like type in the number as I was in the room, which led to some socially awkward moments. but I'm also a strong believer in charging from day one for your product because price is actually part of packaging, is actually part of product. All of these things are inextricably linked. And I would then spend the next hour or so actually onboarding those users, showing them how to take their workflow, dial it up to 10, and get through their inbox twice as fast as before, responding to important messages faster. and about half of those users actually hit inbox zero within that session which for them was life-changing like that this was something that was just impossible before using superhuman of course during that process though we would find 10 15 20 25 bugs and the fact that i was there watching as opposed to the user having to spot that something had gone wrong and report it meant that we found way more than if the user was just trying to find issues themselves.

So I'd be furiously scribbling down notes in my notebook of all the things I'd seen that had gone wrong. And here's the part that made this magic. I went back to the team and I said, there is no point onboarding the next five or ten users until we've fixed all of the things reported by this week's five or ten users. And the reason for that is we'll never start to get to the next layer of issue until we fix this layer of issue. There are only so many issues you can actually think about an encounter in a half-an-hour or a 45-minute period. And so initially there were some grumblings, but the logic made sense, and we started to work on all of these issues. And sure enough, that uncovered the next set, and the next set, and the next set. Until what you actually have is a finely-tuned, highly-polished product experience that most people are just unaccustomed to today. And this is how, unlike almost every other, I think actually every other email startup that has come and gone,

we were actually able to, in a very high-quality fashion, reach parity with something like Gmail or Outlook, and actually, not just parity, a polish that goes way beyond those products. Then we started looking at the metrics of those users. And on every single metric that matters, whether it was churn, retention, net promoter score, product market fit score, virality, we were seeing best in category numbers. And I was like, wow, is this the product? Is it the onboarding? Is it me? Am I just that charming in these meetings? And so we started to change each variable one by one to disambiguate. And the first thing we did was I no longer did the onboardings. So I asked how I head of growth. Actually, my brother, Gaurav Oro, who's well known now for doing growth consulting to many amazing startups. And I was like, hey, Gaurav, you go do the onboardings. In fact instead of 5 or 10 per week why don you do 20 per week And I started sending him around San Francisco and he was doing the exact same playbook I was doing And his numbers were great too And I was like okay it not just me Maybe it him

We should hire a team of people to do this. Not like go crazy, just like a few people. And whenever you're doing this, you want to hire more than one person, just in case there's a false negative due to the person not being particularly good. So I think we hired two or three people simultaneously to start doing these onboardings. One of them didn't work out, but the other two had phenomenal numbers, almost as good as the numbers that we were putting out to begin with. And then I was like, well, we've got these calls down to half an hour each. I think we can design a schedule where folks do 44 or 45 of these calls per week. And I did the back of the envelope napkin math. And I was like, wow, this actually works out. This could be a whole new go-to-market motion that no one has actually ever really come up with before. We have this huge wait list of people who want access. So why don't we hire incredibly compelling people who are experts in productivity,

who know email and productivity products inside out, and can help our users become fully onboarded in a matter of minutes, something that no other software to my knowledge has tried to do at this scale. And so we hired a whole bunch of people and we scaled that team up and that became known as the superhuman onboarding, which is the thing that you experienced. And that is the thing that, as you know, many companies since then have been copied. And to this day, I still get founders asking, wait, what are all the tips and tricks? How exactly did you do the superhuman onboarding and how can we make that work for us? So it's an incredibly powerful technique. It doesn't work for every product, but for the products where it's well-suited, it can have an outsized impact. So you didn't get your PhD in computer science. You famously dropped out, but I'm going to declare right here that you got your PhD in another CS, customer success. Pretty much, maybe. I'm a huge believer in the customer is always right, or at the very least, they always have something to say.

They may not be right. and it's our responsibility as founders, as builders, as leaders to take that feedback, whatever it is, and to figure out what to do with it, to interpret it. And so we built so many systems around doing just that. Early on, you sought the advice of arguably the founder of the growth hacking movement, a gentleman named Sean Ellis, who told you about what I think later became like the Sean Ellis rule, how to evaluate the success of customers with products. If you wouldn't mind, I think we've benefited our audience just to hear a little bit about how Sean thought about measuring growth and some of the lessons that you took into Superhuman. For sure. So this is where our product market fit engine came to be. And at Superhuman, we built our very own product market fit engine. The engine gives you a way to define product market fit, a metric to measure product market fit, It also even gives you a methodology to increase products market fit, and it can even write

your roadmap for you, and that's a roadmap that will take you to products market fit over time. But in order to understand how it works, it's probably worth starting with what even is products market fit. It's a phrase that gets thrown around, but what actually is it? Well, it turns out to be the number one reason why startups succeed. And the lack of products market fit, on the other hand, turns out to be the number one reason why startups fail. It was clearly super important. Paul Graham, who's the founder of Y Combinator, would say that it's when you make something people you want. Sam Altman would say it's when users spontaneously tell other people to use your product. But in my opinion, it is actually Marc Andreessen who has said the most vivid definition. He would say, you can always feel it when product market fit is not happening. Customers aren't quite getting value. Users aren't growing that fast. Word of mouth is not spreading. The press reviews are kind of blah, and the sales cycle takes too damn long. But you can almost always feel it when product market fit is happening,

because customers are buying as fast as you can add servers. You're hiring sales and support as fast as you can. Reporters are constantly calling you about your hot new thing. Investors are staking out outside your house, and money is piling up in your checking account. And that is indeed a very vivid definition. It was one that I was staring at through tears in the summer of 2017 because it seemed so subjective, so unactionable. What do you do if, by that definition, you don't have product market fit? Indeed, can you measure product market fit? Because if you can, then maybe you could optimize it. Maybe you could systematically, even numerically, increase it. so in the summer of 2015 we started much like any other software company by writing code and in the summer of 2016 we were still coding in the summer of 2017 we were still coding and i felt this incredible intense pressure to launch both from the team but also from within myself because after all my last company that we talked about reported that had launched scaled and

been acquired in less time and here we were two years in and we still hadn't launched But deep down inside, I knew that no matter how intensely I felt pressure, a launch would go very badly. In other words, I didn't believe we had product market fit. And although I knew it, I couldn't just say that to the team because these are super ambitious, hyper intelligent engineers. They put their hearts and souls into the product. So I needed a plan. And that's where Sean Ellis came in. In the April of 2017, I started my search for that holy grail for a way to define product market fit for a metric to measure it and for a methodology to systematically increase it. And I searched high and low. I read everything I could find. I spoke literally with all the experts. And then I came across Sean. Now, he ran growth in the early days of Dropbox, Lock Me In and Eventbrite. As you noted, he coined the term growth hacker. And he found a leading indicator of product market fit, one that is benchmarked and predictive. You just ask your users the following question.

how would you feel if you could no longer use the product and measure the percent who answer very disappointed? And after benchmarking hundreds of startups, Sean found that the companies who struggled to grow almost always got less than 40% very disappointed, whereas the companies that grew most easily almost always had more than 40%. So in other words, if more than 40% of your users would be very disappointed without your product, you have initial product market fit. And with that engine, we now have a methodology for systematically increasing product market fit. It even automatically generates our roadmap for us. The metric is even more objective than a feeling. It predicts success better than net promoter score. And it's not only the best way to measure product market fit. We use it to develop our very own product market fit engine. And with this engine, we now have a methodology for systematically increasing product market fit. and it even automatically generates our roadmap for us. And if you look through the article, which goes through this in depth,

it's basically guaranteed to increase product market fit. So I rather infamously wrote this up. I think it was in 2019 or 2020 on first round review. And then since then, it's become the most shared entrepreneurship article of all time. So maybe we can put the link in the notes for our listeners here. but I go into a lot of detail how you can build on Sean's question, build a whole algorithm around it, which, as I said, actually takes you to product market fit. Brilliant. That's what I wanted the audience to hear and thank you. We will share that link. I've heard you talk about, after Reportive, your mission was to pick an impactful problem and go about solving it. You already had a lot of success. I'm sure, you know, shortage of things you could do. I would say probably on a short list of technologies that have impacted humanity, maybe after fire

in the wheel, I'd say email is a close one. Obvious why eventually you decided that that would be the broken thing that you'd fix. Something I've never heard you talk about is what that process was like and if you could share any examples of other big problems that you chose not to fix in order to focus on email. Yeah. Well, as you noted, when I started Superhuman, I wanted to solve the biggest possible problem. And email, of course, turns out to be a way bigger problem than most people realize. And it hit me just after I'd sold Reportive to LinkedIn. I was commuting from work in Mountain View, that's where LinkedIn is headquartered, to where I live in San Francisco. And that's, in case you haven't done that, roughly a one-hour journey. but this was 2014 and both Uber and Lyft were having their moment. So I happened to be in the back of an Uber and on this trip I finished writing some documents, I cleared out my email, I checked my calendar for tomorrow, I made dinner plans with my girlfriend and I even called my mother. So I was simultaneously being a productive employee, a proactive boyfriend and a present son

and I realized in that moment that the real magic of Uber was not commuting, it was not about getting from A to B. It was about time, being able to do these other things. So I asked myself the question, where do we spend all of our time? And this is a little bit of a trick question, but I'll ask you, can you guess the one thing that professionals spend the most time doing? That's not email. Scheduling. Scheduling is a very good guess because that's up there with email, a little bit less time than email because a lot of email actually is scheduling, but it is in fact sleeping. We all spend more time sleeping than we do basically any other activity in our life. And as a lifelong sufferer of sleep apnea, it hits me particularly hard. And I thought a lot about that, but I also had no real idea how to fix it. And after sleep, it of course turns out to be email. In fact,

there are roughly 1 billion professionals in the world, and on average, we spend three hours a day reading and writing email. So that's 3 billion hours every single day, or not a trillion hours every single year that go into email. And that's how I knew it was a problem worth solving. I couldn't imagine something bigger than a trillion hours per year. Now, alongside this, I could see Gmail and Outlook getting worse every single year, becoming more cluttered, using more memory, consuming more CPUs, slowing down your machines, still not working properly offline. And on top of that, people were installing plugins, like my last company, Reportive, but also Boomerang, Mixmax, Clearbit, Yesware. You probably remember some of these email products. You name it, people had it. But the problem with these plugins is that they took those problems of cluster, memory, CPU performance offline, and they made all of them dramatically worse. So we decided that it was time for change. We set out to build the fastest email experience ever made. And today, most of our customers get to their inbox twice as fast as before.

They respond faster to the things that matter, and they save four hours or more every single week. Today's amazing episode is brought to you by the good humans over at Guild AI Academy, which builds training around your company's own AI tools, workflows, policies, and leaders, then deploys it with small peer cohorts to make sure those habits actually stick. If you've rolled out AI tools, and I assume you have, the real question everyone's asking is, is anyone actually working differently? For most teams, the honest answer is not yet. Tools are there. The payoff isn't. Guild AI Academy changes that. It's the AI training program customized for your company. It's delivered through a cohort-based learning experience so employees practice together, and they actually get better at doing the skills you need them to be learning. A Fortune 500 company had an AI tool sitting mostly unused for a few years After Guild AI Academy program over 90 of its leaders were using it regularly

And unlocking new levels of efficiency benefited the whole company. See how it works for yourself. Go to guild.com slash AI. Again, that's guild.com slash AI. Now back to today's amazing episode. So, at the time when you were launching Superhuman, nobody realized they had an email client problem until they saw Superhuman. Today, I would say, well, it's a question, it's not a statement, but AI entrepreneurs that are looking for the problem to solve, like email, people not realizing they needed a Superhuman, And just to me, like the calculation, the analysis that you would do if Rahul was looking for the equivalent of superhuman in 2026 would be different because kind of everyone's sitting on top of this substrate of insanely intelligent reasoning models and dot, dot, dot.

How would your process have been different if you were looking to do something as impactful starting today versus 2014? In many ways, it would have been the same. The thing that makes Superhuman a fascinating company, highly investable, and why it's been as successful as it has been, is the market is absolutely ginormous. For example, one way to think about it is the top-down approach, which we've already talked about, which is a billion professionals spending three hours a day on email. That's three billion hours a day, a trillion hours a year. it is really hard to think of something that we spend as much time doing. Slack is up there. In other words, team chat. So Slack and Microsoft Teams. Scheduling is a second, actually a distant second. And then you have a bunch of the other work activities that we have. So deliberately, I was going after this thing that just impacted so many people and which no one loved.

Everybody hates doing their email. it's also a really great edge of a wedge because you can think of email as a central and a key component to an ai native productivity suite or an ai native collaboration suite and of course that's what we're now building here at the wider superhuman one of the most exciting things for me about the acquisition is we basically get to do all of our roadmap but in parallel the thing that I thought would take 20, 30 years to build out, we can massively compress the timeline to that. So what I still look for in founders with companies today, my Vice 20 founder, is just pick an absolutely ginormous market. Now, it doesn't necessarily have to be ginormous today. It's okay to pick, sometimes in fact better actually, to pick something that seems a bit weird and silly today, almost like a toy, but that is going to grow incredibly rapidly. That is a perfectly fine alternative approach. But ultimately, the market has to become very large. Either it's very large now or it's growing very quickly. Some things, by the way, are easier than

ever about creating companies. I would say, number one, writing code. The cost of creating most software projects, of course, is rapidly reducing with all of the coding agents that people now use on a regular basis. Number two, raising capital. There has never been this many dollars chasing the next big thing. Number three, learning from experts. The art of creating startups is becoming increasingly a science and very widely known. I remember when I was first coming up, there was basically nothing online about how convertible notes work. The safe hadn't been invented yet, just to pick a random example. And so trying to figure these things out from first principles working directly with lawyers was very slow. And number four, getting early traction. You know, just a few years ago, a good growth rate for a YC company was 5% to 7% per week. 10% per week was exceptional. Whereas last year, the average growth rate for a YC company was 10% per week with plenty doing 15% to 20% per week. And it's not uncommon nowadays for a Y Combinator

company to get to $10 million of ARR in their first year. That used to be completely unthinkable. That was like once in a blue moon. Now, again, it's not uncommon. But some things are harder. Taste, for example. When anyone can build anything chiefly, the real skills actually become taste and restraint. And it's an old example, but one of my favorite examples of this is actually Instagram. And maybe you'll remember this. Before Instagram was Instagram, it was actually an app called Bourbon, and it was a lot more than just photos. It was actually in that era of applications that I think we called social mobile local, SOMOLO or SOLOMO. It was this weird acronym that we had. And then there was like a hyper social mobile local set of apps. And it did, it was a social network. It did check-ins. It did location finding. It was like, where are my friends on a map? And it also did photos. And it didn't really take off. It didn't take off because it didn't have a point of view. It didn't do one thing incredibly well.

It didn't channel taste. And when they narrowed down to the photography piece and did all of the things that had been very well documented, that's really when it blew up. That's a really good, simple example of taste actually making the difference. The other thing that's harder is competition. There are a lot more smarter people and more venture dollars now going after company building than there ever were before. And then the first thing that I think is actually harder is distribution. And in a world where building products is increasingly easy, it becomes increasingly important to figure out how to get those products in front of users. And in such a world, leverage actually shifts a little bit back to incumbents. Startups need to find their edge of a wedge and then a durable business before incumbents leverage their distribution to do the same, basically before incumbents build or buy the same thing and put it in front of all of their users. So I would say in many ways it's the same.

Pick gigantic markets, either those that are big today that are going to be big in the very near future. But ultimately, it's never been better to build a company. We have never been in a better time. And the proof is in the pudding. Companies today are growing bigger than ever, faster than ever. The bar to be one of those companies might be harder than ever, but the prize has never been bigger. On this podcast, we talk a lot about the complicated relationship between humans and machines. And unlike, say, getting through your email inbox faster, AI is forcing us to wrestle with some more complicated questions, like existential ones. Like, literally, like, why am I here? Because it's feeling like this artificial entity does a lot of what I thought was uniquely part of the human condition better than I can. I know you're a deep thinker so I get to ask you this question I haven't heard you talk about your perspective on the incursion of AI into everything we do

what's your thought about how does team human how do we as humans achieve a balanced relationship with AI it gets us beyond the current state where it's a very polarizing topic in general particularly in the United States, as you all know, there's a great mistrust of AI, and yet we're increasingly dependent on it. How do you think about those complicated topics, both at a raw level but then at a macro level? I think that AI is going to deliver so much improvement to our quality of lives and the economy and our national security and America, as well as the world in general. we are only beginning to scratch the surface with what we're seeing today. At the same time, we're also, and this is me as a person and also us as a company, where I am very aware and we are very aware of the anxiety and the fear

and the real impact in some cases that it's having on people's lives. And we think that what we can do is build AI in the most responsible fashion. and I can talk about what that means to us if that's interesting. But I'll say the high level just for the moment in terms of the overall impact and why I'm excited. So maybe starting with the name, Superhuman. Like why did we rename the whole company to be from Grammarly to Superhuman? And as you can imagine, that's a complicated choice. It's difficult. Like we have to change websites. Superhuman is known for being an email app. not the whole suite. So there's a lot of work. In some cases, it's an uphill battle to make that work. But the reason we did it is not only does superhuman stand for what we're trying to do, which is to give you superpowers everywhere you work, but it also has human right there in the name, front and center.

The way that we want to build AI is by allowing you and me and humans to do what we uniquely can do, be our most creative, our most strategic, and our most impactful. and as amazing as AI gets there's something special when you and I get in a room together or we get on a call and my unique lived experience combines with your unique lived experience and there's a certain frisson there coming together of decades of somebody's, in this case two lives and obviously the commentorial explosion when you have a third person or a fourth person is kind of mind-boggling that AI simply will never be able to do, or at least not in the foreseeable future. Theoretically, if it recorded everything that happened in my life and everything that happened in your life, maybe. But that is not going to be a problem for many decades. So there is something still very special and very unique about teams of people coming together to combine their lived experiences.

And that's superhuman. And that's one of the things that we're excited to deliver. But as I think more broadly about the overall economy and the impact on jobs and so forth, I'm sure that other guests have spoken about Jevons Paradox, but I'll redo it if they haven't. I think it's very germane and more people need to understand this. So Jevons Paradox came from the Industrial Revolution in the United Kingdom way back when. And, of course, back then, all power was derived from burning coal. We'd burn coal to create heat, to heat up water, to create steam, to move engines. And that's how power used to be generated, right? And then James Watt came up with a much more improved steam engine that made the energy extraction from coal way more efficient. And there's a lot of analogies between then and now in AI.

Back then, there was this widespread panic that the demand for coal would go down because we need far less of it. And therefore, the coal industry would fall apart. And the coal industry was all of everything back then. Everything kind of rode upon it. And therefore, the whole economy would fall apart and there'd be widespread unemployment and everything would be terrible. But it turns out the exact opposite happened. total coal consumption surged as engines became cheaper to run and were adopted much more widely for a greater set of use cases before as new things became profitable that previously could not have been imagined. Let's do a more recent example. So computers and digital tools, if you go back to like 1960 70 80 90 there was actually a panic at that time from paper companies and lumber companies saying oh as we digitize everything the paper industry is going to fall over

It's going to die. But actually, it turned out that global paper consumption tripled during that period because printers and printing became easier and cheaper. when spreadsheets, when Lotus 1.2.3 and Excel started to become really popular, people were really worried that the accounting industry would go away. And actually, the opposite happened. It led to a whole new line of work, which is FP&A. Now, as we catch up to today, a lot of people have said, well, with coding agents and software factories, does that mean we need fewer software engineers? And hopefully you can see the pattern here. This is exactly the same pattern of argument. Actually, it turns out you need way more software engineers because way more kinds of software are now more profitable. I had a really great conversation with a not-for-profit founder the other day who is building technology for certain parts of Africa,

and she was unable to do so profitably. Profit is the wrong word. She was unable to do so in a way that was sustainable before coding agents because it was too expensive to afford to pay everyone to do all the things, whereas now she can create the software and hire the software engineers and deliver it to market much more cheaply, thus enabling a whole new category of industry that didn't exist before And in this case, it's a really great not-for-profit story. So it turns out there's way more demand now for software engineers than there ever was before because there's so much more software to be written. And by the way, when you have more software to be written, you have more security risks, more complexity. So there's more roles in cybersecurity and compliance and governance. I'm hearing the same thing in legal work, by the way. Folks have said, oh, aren't we going to need far fewer lawyers now that you have Harvey and Lagora and so many other of these legal companies that are automating so much legal work?

Well, it turns out that when legal work becomes really easy, the demand for legal work increases. Like a lot more cases get brought. There's a lot more demand just to do things. And so case volume has massively exploded. I was talking to a lawyer whose firm has now like 10x the case volume they used to have. And when you have 10x the case volume, sure, AI is going to do a whole bunch of that. But it turns out the set of stuff that you need humans to do also 10x is. And so they're actually hiring more lawyers than ever before. And the most surprising example of this that I came across was customer support hiring, which you would have thought of all the things. Surely that's where AI is going to have the biggest impact on jobs. Well, it turns out customer support hiring has recently rebounded and is now growing 10% to 15% faster than other sectors, which I think is remarkable. And it's exactly the same phenomenon as everything I just described. Maybe I'll spell that one out just to make it more obvious.

So what do companies do when AI can handle a lot of customer support? So, well, one choice is you eliminate a lot of those roles, but you keep on delivering crappy customer support. Turns out that's not what most companies are choosing. What most companies are choosing is to invest in their customer support and making it amazing. And so you end up taking those people and asking them to do more complicated things, high-level tasks, more difficult cases, and let's try and compete on the quality of support that we're delivering. So that's a lot. But hopefully that explains why I am personally so bullish on the magic of AI as a new technology for changing every aspect of how we work and who we are and what the economy is and what America is and what it means to be a human. I'm not going to pretend it's not going to be without its challenges. And I think that whenever there's any kind of transformative technology like this, there's a lot that we need to work through as a society.

And we want to do our part to be responsible parts of that conversation. But I'm very excited about what's to come. Oh, cool. That is the most dense and articulate description of Devin's paradox that I have heard yet. So you get the gold medal. Nicely done. I think more people need to understand it, so thank you for the soapbox. Very well done. Very well done. You know, unbelievably, we're about out of time. I told you I'd get you out of here efficiently. But you're not going anywhere without answering one last important question for me. We're talking about AI and humanity and beyond email, beyond the inbox. One of the things I'd love to know from you is when Rahul and Dan are back here in a decade, It's 2036, and we're having a version of this conversation. What's one thing that you believe will be commonplace then that most people would disagree with sitting where we are today in 2026?

Okay, so to make sure I understand the question, what do I think will be commonplace in 10 years that most people would disagree with now? Yeah, something contrarian about what you think will be commonplace in 2036. Oh, boy. That's a really tough one. And I think by definition, we are ill-equipped to, not by definition, but like it's well observed that we are ill-equipped to predict on the 10-year timescale as opposed to the one or the two-year timescale. I don't know. I think I just gave you a speech that is basically so super hard to predict, right? Like the, I, okay, here's what I can predict. I think in many ways the world that we will be living in will be unrecognizable compared to now for the better. Maybe I sound like a techno-optimist or I sound almost utopian when I say that, but I really do believe that from advances in drug discovery through biotechnology,

through productivity to every aspect of knowledge work that you can imagine to manufacturing and industrialization, the supply chain, food, drink. It's just all going to get so much better over a 10-year period in a way that we can't imagine. Maybe that's not contrarian. Maybe that's obvious. But that's certainly something I believe is true. Where we sit today, it's still unsettling for all of the humans who thought that they nailed the right career and they're maybe mid-career or even more so early career. And right now, AI feels like a very destabilizing force. And they can hear you and I talk and be optimists about the future. and yet it's scary because it requires an investment in skills that we don't know what they are yet

and industries that we don't know what they look like or they don't have a name yet. What's your message to everyone who says, yeah, but, you know, here I am today and I don't know what I'm going to be doing between now and 10 years. You're asking me to, you know, kind of just take a leap of faith. How do you have those conversations, which I'm sure you have frequently? Yeah, I mean, I would start by saying I have a lot of empathy. I feel that same fear. I'm 42 years old, and I thought I'd assembled a formidable set of skills, some of which will continue to be relevant going forwards, many of which won't be. and I can rail against that all I want, but it just doesn't change the march of progress. What I will say is I think a lot of the fear is people amplifying each other's fear. I'm going to avoid my social media rant and how social media does that

and just try and ground us in numbers. So, for example, the share of U.S. jobs that are software engineering jobs. That's actually up 40% year over year. We went from 1.6% of US jobs that are software engineering to 2.2%. That's huge. And it's very counter-narrative, right? That's counter to AI is taking all the engineering jobs. That's not true. So I think what I would say is, where possible, let's ground ourselves in the data. Let's actually look at what's happening to the overall, let's look at the overall shifts and then see how we fit in. On the flip side of that, though, we, all of us, me, you, everybody listening, everyone in the future, we all need to stay connected to what's happening. And, you know, in some ways, it's the great equalizer. AI doesn't care how old we are or how, I think you use the word, invested we are.

It's going to make everything easier and faster. So it's really down to our personal agency. What do we do with that? And I think the right thing to figure out how to do, well, before I say the right thing to figure out how to do, let me paint an analogy that hopefully makes this sound obvious. Would you hire someone who didn't know how to use Google? Would you hire someone who didn't have the requisite level of proficiency with, and it obviously depends on the job, right, the Word or Google Docs or Excel or whatever. And you don't need all of these things for every role, but I picked a very low bar because it would be unthinkable to hire someone who did not know how to use a web search engine today. But that wasn't always the case. And there was a time where people had exactly the same concern. And, of course, what was the answer? Well, the answer was to get on the Internet and to get on the web

and to figure out how it works and to become digitally native. Exactly the same answer applies today. The answer is to become AI native, and that would be my encouragement to everyone, noting it is scary and I have a ton of empathy. I'm scared as well, but that is the answer. Brilliant. That's helpful. And you have an open invitation. Come back. I think we got to zero of the topics that we had scheduled, so thank you for letting us go off the strip, but please come back and let's continue the conversation. This is really, really impactful. Yeah, I'd be excited to. We've got so many cool product announcements coming that we didn't get to, but maybe that's a good thing because we can have a subsequent conversation. Part two. Yeah, let's do a part two. So for the maybe one listener in this audience that somehow managed to not yet follow you on socials, where can we learn more about you online? Great. I am on Twitter and LinkedIn, and you can follow me at my first name and my last name.

That's R-A-H-U-L-V-O-H-R-A. My DMs are open, so just ping me. I pay more attention to Twitter than I do LinkedIn, so ping me over there if you want to. and thank you for having me. This has been so fun. Absolutely. My pleasure. We will definitely continue this soon and gosh, that is all the time we have for this week on AI and the future of work. As always, I'm your host, Dan Turchin from People Reign and of course, we're back next week with another fascinating guest. Yo, it's the end of this show. Time to power it down. AI exists for humans with a crown. Being human in the house got the vibe so clever. Dan Turchin from People Reign making us better. We learned about AI for humans. We're ready to conquer the future. That's all for now. See you next time.