The Intersection of AI and Engineering Leadership with Farhan Thawar, Shopify

We have this thing called the trust battery, which is basically when you meet people, it depends on the person. Some people start at like a hundred percent trust, some people start people at 50%. So I meet you for the first time, I'm like, okay, Ray, like you're going to be at 75% trust. And it's hard to build that trust remotely. But what we've noticed is when you're IRL with people and you have lunch and you talk to them, whatever you can build, that trust battery goes up. And then as you go remote and maybe I have not that many interactions with you, maybe that my trust battery starts depleting a little bit because I don't have a lot of interactions with you. And then if we get together again next quarter, it builds up again and then I can feel like, okay, I could work with you and I don't have to do a lot of the pleasantries and like, how's it going and what's happening before I get to the real meat. And so we feel like once every four months or so is a good time to get together with your team. We have Shopify summit. All 10,000 Shopify come together in Toronto. And again, we want to keep that trust battery fully charged.

Welcome to another episode of our Traction podcast. This is the Next Wave. The Next Wave is a different take on our podcast where we're bringing back some of our most favorite guests that have either graced our stage or joined our podcast 10 years ago, five years ago, and we're going to revisit their topics to see do they still hold up in Today's world of AI traction is brought to you by LaunchAcademy CA, one of Western Canada's main tech hubs and an international accelerator program that provides you access to the Canadian startup Visa as well as boast AI your go to partner for R and D tax credits. This week's guest is a fan favorite of Tracktion and a good friend of ours. Farhan Thawer is VP and Head of Engineering at Shopify by the acquisition of Helpful.com, where he was a co founder and CTO. Previously he was a CTO mobile at Pivotal and VP of engineering at Pitbull Labs by the acquisition of Extreme Labs. Busy man. Getting a lot of acquisitions in there. Farhan is an avid investor and an advisor at startups in Toronto and San Francisco, including being a mentor at Y Combinator at First Round Capital. Previously, Farhan held senior technical positions at Achievers Microsoft and Trilogy. Farhan completed his MBA in Financial engineering at Rotman and computer science and electrical engineering at Waterloo. Farhan, welcome back to Traction.

Thanks. Nice to be back. It's cool. I'll see you folks after a couple years.

Yeah. And things have changed since the last time you're back. You've been on a few times and the premise of Next Wave is we're revisiting past talks and seeing do those beliefs and visions and tips hold true today. But we're going to talk about remote work. So this is your one of your more recent talks, 2021. And the world was upside down during that time. That's the time that will not be mentioned. The COVID times and things have changed and we want to kind of revisit what does remote work look like today and moving forward. So right into the fire we're going to show a clip and go from there.

Remote work was growing as a thing.

In the world, right. So over time I started to see more remote postings. I started getting reached out to more often for remote VP engineering positions. Remote work was a trend that was gaining some steam. It was still pretty early, but it was something that was growing over time.

So that's the second thing that was happening.

And what we realized was we can one be at the mercy of the when we think these things are going.

To open up or we can choose.

Certainty over uncertainty by one attaching ourselves to the trend we already see happening in the world, which is office like the reduction of office centricity, the ability to hire the best people wherever they live and the amazing new tools that are coming to remote work to just claim that we're going to go remote forever. And that caused a bunch of things to happen in our company.

One, if you used to like work.

On an ironing board or have a really crappy in the corner like setup for your laptop and camera, we just encourage people to say we're going permanent. We were talking about this. Get high speed Internet, get ethernet, get a good camera, get a good mic, get a good home setup because we're going to be here for a while. That was the first thing we wanted people to really understand that. And then the second thing we said was today is the worst, the worst type of remote.

Right?

It's not remote. This is pandemic mode. You can't go outside, you can't go to lunch with your friends, you can't.

Travel to other places.

This is not really remote today. It's pandemic. And so when we move to remote, there's going to be an acceleration of how we think about it, the tools, the techniques, and the ways in which to create those relationships.

Cool. By the way, I like this format of, like, show me something I said before and then have me challenge it. I love it.

Yeah, no, because, like, there. There's a lot of things that have changed over the time. Again, we're going back to 10 years with talking about bots and now AI agents. But you've also got cybersecurity, so different. You've got different political situations around the world. You've got different types of trends that are happening in. In terms of media and consumption. And so there's so many different ways to look at this. And remote work seems pretty straightforward. Okay, you're working from home, but no, there are so many different layers of complexity that come into the play. So hindsight, 2020, you. What you saw in 2021, and the vision of where you saw everything going, did it come to fruition the way that you predicted it or envisioned it?

Yeah. So first off, I looked way younger then, so I can see how much time has passed or what the years have done to me. But I would say, actually, I'm very surprised that we kind of. We got that prediction right. Like, going remote actually was one of the amazing things we did at Shopify. We were able to hire amazing people wherever they are. They didn't have to come into the office. You saved on all that, that toil, the commute times, and like, you have to be near. You have to be the office centricity, and you have to be near an office. I think we made that call correctly. I think what we learned, though, at the same time is that there are amazing experiences you can have only irl. So you can only like grabbing. If you and I had lunch over Zoom, it's not the same if we went to a restaurant and had lunch like there, irl, if you. If IRL has less toil and you can do it, afford it, or live in the city, whatever, IRL is better. There's no. No challenge there that if ir. What's better, IRL or Zoom, like, you want to go on a date, have hang out with your friends, go with your family, like IRL is better. That being said, the ability to hire the best talent in the world, lose the commute, work from anywhere, like, all those things actually contribute to a greater sum. And that's why we chose it. We said, actually, while IRL is better, we think all these other advantages outweigh. And then the other thing we realized was that by taking the IRL experiences, we should have. Even though it's not every day nine to five in the office. It's like we have Shopify Summit once a year. We have this thing called Bursts whereby you can use a Shopify port or another location if you have to get the team together to brainstorm prototype work on roadmaping. We leaned into those experiences as well. And so I like to tell people, then they go, oh, you guys are remote, right? I'm like, yes. And they go a hundred percent. I'm like, oh, like 90 or 95. I like to throw those numbers out because I want to share with people that there are times you should get together. I had a leadership burnt, what we call burst with all engineering directors and above and all principal engineers and above in the summer. And we got everyone together and we walked through technical topics and we had an AMA with Toby about the state of software engineering and where Shopify is and how we play in the ecosystem. And we do this thing, like we get together for six week reviews where every six weeks we get together. And so there's these areas where it makes sense to be irl, but the majority of the work and the people can be wherever they need to be for 90% of the time. And we get all the benefits of having the best people in the world being wherever they are. And so I would say I'm like, not shocked, but like surprised that, like, we were right on that. But then we modified it. We said, okay, it's remote, but this is irl. Shopify Summit once a year. This is irl. Leadership Bursts a couple times a year. This is irl. I'm doing like six week reviews and every six weeks we get together or parts of the team get together for that irl. And then for me, you know, I live in Toronto and we have these ports which are come in if you want. I used to come in once a week. Sometimes I come in two or three times a week, like at my choice. And I think that really, again, we have a lot of interns at the interns. I did a talk for the intern group and I was like, great. We were able to do an IRL thing for them. And so I get in the best of both worlds here where I'm like, we got. If you want to come in, don't want to come in, don't come in. I'm like, oh, it looks crappy outside today. I'm like, I'm going to stay home, save time on the commute. I get to work a couple extra hours and I'm winning That way as well.

Yeah, it's a hybrid mentality, but also.

Hybrid mentality, but not. Yeah, but not hybrid in the percentages.

Yeah, there. There's some like Gen Z, Gen A that are working from home. Their parents are there and they just want to get out of the house.

Yeah. And so for them, come in every day if you want.

Yeah. And then there's others that they got to drop the kids off at daycare or whatnot. And it's just more convenient to be.

And then the other thing we did was we said, for the early talent cohorts, I require them to come in three days a week. So we realized that for people early in their career, they not only need the camaraderie of the office and they might be, like I said, you might be living in a one bedroom condo and they're downtown. I don't want to stay in this environment. We require them to come in three days a week. We want them to have a cohort of people who started with them that they can have lunch with. They can feel like they're growing together and they can go tap someone on the shoulder, hey, how do you do this? Like, it's hard to do over, like, zoom or. And so we have cohorts of people in early talent. And so we're our intern program. We require them to come in three days a week. They can build those bonds, that camaraderie. And then of course, they can keep coming in after the internship if they come full time, or they can go remote like everyone else.

Yeah, that's something that I highlighted back when this all started in 2020. It's that younger workforce that's really going to suffer here because you and I are seasoned. We're in a Zoom call, 20, 20 other people. But something might come up and I'll recognize, hey, that's actually an issue that we need to deal. And I'll either direct message you or say, hey, Farhan, we need to figure that point out. And then we're going to take it offline and we're going to have our own Zoom call. So those interns or younger employees are not going to benefit from hearing us hash that out. Whereas if it was in person, we might just jump on the whiteboard and they'll still be in the room witnessing that.

Yeah. I think part of it is you're also mentioning, like, the early town. I think part of it is to be prescriptive. So by me requiring them to come in three days a week, it's forcing them to build those Bonds. Because I'm making it a requirement. Like, the opposite would be where I just say, hey, by the way, we have these offices. Come in if you want. And by the way, I've come into the office and I've definitely been like, Toby was here and I was here. And we were struggling to find people to have lunch with because no one was here.

Yeah.

But now with a little bit more structure around saying, hey, you're the early talent folks, Come and have lunch with maybe one of the greatest entrepreneurs of our generation who happens to be in the office, and your boss. Let's. Let's force the collisions. And that has really changed, I think, some of the environment that they're growing up in at Shopify.

Yeah. One of the techniques that one of my friends has done is they do free lunch on Tuesdays and Thursdays, but you have to order your lunch on Monday or Wednesday.

Okay.

And so now the managers can see, okay, who's actually coming to the office because they placed their lunch order. But so you have to be prescriptive. You have to curate this. So this is now a new task that's under the manager's job description, is to be able to corral your team and find ways to make them most efficient.

Yeah. And it's not just. It's not just early talent. I'll show you a funny thing we just did this week where we told again that same. I have this group of the directors and above and principal engineers and above managers and ICs in engineering. There's about a hundred total. And we said, hey, Black Friday, Cyber Monday, which, by the way, is the biggest week for Shopify every year, said, we're all going to be in the office, so if you want, come to Toronto. And then we said the same thing for hack days, by the way. It's hack days. Get together irl, come to Toronto, go to Montreal, go to Ottawa. But we're trying to, like, again, even for the seasoned people, give them advice on, hey, this is a good week to travel. And then even if you don't have meetings with people, if you're hacking for hack days, it's the best time to come because everybody's walking around and hacking on different things and prototyping. But then Black Friday, we're gonna have Hacker Radio and we're gonna talk about, like, the amazing things happening with our merchants, and we'll obviously have an amazing new globe and people can talk about. So there's all kinds of things happening. And again, have those random connections and.

It just has a Deeper impact IRL than it would.

Yes.

On Zoom.

Yeah. And we have this notion. Yeah. Have you heard of the trust battery story at Shopify? Like, we have this thing called the trust battery, which is basically when you meet people, it depends on the person. Some people start at like 100% trust, some people start people at 50%. So I meet you for the first time, I'm like, okay, Ray, like you're going to be at 75% trust. And it's hard to build that trust remotely. But what we've noticed is when you're IRO with people and you have lunch and you talk to them, whatever you can build that trust battery goes up and then as you go remote, and maybe I have not that many interactions with you, maybe that my trust battery starts depleting a little bit because I don't have a lot of interactions with you. And then if we get together again next quarter, it builds up again and then I can feel like, okay, I can work with you and I don't have to do a lot of the pleasantries and like, how's it going and what's happening before I get to the real meat. And so we feel like once every four months or so is a good time to get together with your team. We have Shopify summit. All 10,000 Shopify come together in Toronto. And again, we want to keep that trust battery fully charged. And now these are the ways we do it, even though we're fully remote.

And after three, four years now of going through this, you learn best practices, but has that carried forward through all the different divisions of Shopify? So yeah, engineering, it works. But what about marketing? What about business development?

Oh, no, for sure. So we have again, every team is different in how they want to approach it, but Shopify, the company's remote. And like I mentioned, if you live near a port, and right now we have ports in Toronto, Ottawa, Montreal, New York. Come in if you want. And then if we see like, I definitely run to lots of salespeople in the elevator and like, so there are different disciplines who kind of use the remote to their advantage. Same with me. I happen to be traveling in London. I did took some sales meetings in Paris and that that allowed me to connect with shop of folk in those areas. But it's not as. It's not that we've gone again back to office. It's more about this idea of can we hire the best people wherever they are? And then if I happen to be going to a certain locale, I will ping a group or I'LL tell my admin, say, hey, I'm gonna happen to be this, this is happening to me this week. I'm in San Francisco this week. And so we have a big dinner planned on Thursday night with all the Shopify who happen to live there because we're trying to create those connections for maybe those people who are not coming into a port regularly.

So I want you to take your Shopify hat off and put your investor hat on. Now, you've invested in a number of companies.

Yeah.

What is it like for a company that's not the size of Shopify? They don't have the luxury of ports. Are they leveraging WeWorks or like, how do they take the best learnings from Shopify and apply to their startup?

Yeah, for sure. So I would say I'll go back to the beginning where I said, there's nothing like irl. IRL is amazing. If you got two, three people and you happen to be in the same city, I would do some IRL action because there is nothing like the ideation phase, the brainstorming phase, the prototyping phase that can be done. It can be done for sure, remotely. But you're just adding additional complexities to that. It's like saying, hey, we're going to. I don't know if this is controversial or not. We're not going to use the cloud, we're just going to build and we're going to start buying machines and servers and put like you're adding more complexity. It's just on prem. We're just adding complexity to like your potential, your solution. Meanwhile, if you just said, you know what, we have aws, GCP and Oracle and Azure, let's just use one of those, get Credits from post 5 of those folks and like launch. You remove the complexity, yes, you can do remote, but no, not everyone should adopt it because you don't. You might not have the infrastructure in place. And so if you don't have the infrastructure, I would remove a variable and if one variable could be, how do we do remote? Well, and how do we keep trust high and how do we keep the culture and how do we. What tools should we use for pair programming? Like, you gotta figure, find these things and instead it might make more sense to go irl. Now, of course there's. I talked to a company the other day which had two co founders, one east coast, one west coast. Of course you can make it work. It's just you're adding a piece of complexity to the puzzle. So my advice is remove as many complexities as possible. And if you can, because, hey, by the way, your co founder is on the west coast and you're in the east coast, then be intentional about what are the things you're going to do at, at the scale you're at to keep the trust building high. Maybe once a month one of you is flying to the other person's location and you're working out of their space for a week. Right. Maybe you're investing in some tools, right? There's a tool called Tuple we use here for pair programming. Maybe you're buying that tool because you're actually using the highest quality pairing software. Maybe you invest in good cameras, maybe invest in good microphones, make it actually better for that experience than to just go, oh, we're going to do remote, it'll be fine. Like, I wouldn't do that. I would actually try to ensure that you're building something that you can. That can. Has a foundation that's strong.

I think that's key is building that foundation. If you are going to build a remote company from those, from the beginning, it's just like building a house. If your foundation sucks, it's all going to come coming down on you.

And by the way, IRL is not easy either, Right? Like what time are people coming to the office? What kind of laptop are you using? What cloud software are you going to choose? Like, there are still lots of questions. I'm just saying if you're going to be intentional about that environment, do you have an office space? Do you have a wework? Are you paying for food or are you not paying for food? How easy is it to get for people? There's all kinds of things that go into that also. And I would put as much care into that remote decision as you do to those decisions.

Yeah. And again, culture, you're starting at two people. It's easy. Get to 10.

Yes, exactly.

Get to 50. Okay. A lot more complicated.

Complicated, exactly.

Different boat, right?

Yeah. And it always changes. I mean, even for helpful. We had this funny thing where because we're Toronto and I was saying earlier that we didn't have snow last season, but most years you have snow. And at our helpful office, when we were doing our startup, we always took our shoes off. It was a weird thing because of snow. Whatever. We just took our shoes off at the door and so helpful became this thing. Even though we got up to about 25 people, we just everybody took their shoes off when they came in the office. So we were like, it was just a thing we did. And then I remember our first investors best summer gave us slippers, helpful slippers, because they knew that we were a shoeless office. But that was part of the culture we came in. Everybody takes their shoes off, walk around in slippers, and that was the environment we were in.

Are you judging people by their socks?

Yeah. No, not really. What's funny is. What's funny is I'm standing right now without my shoes on. I actually do have slippers in the thing because when it's winter, I do want to switch out of boots into my slippers and walk around the office. Yeah.

Being comfortable is underrated.

Exactly.

We're going to jump to our next clip.

Now, you mentioned you said who would you hire? So I think the way to think.

About it is how should we think about reducing risk?

That's always how I think about these things. And if the risk is in, and maybe I'll be controversial, if the risk is in, go to market, I would hire somebody and go to market first because we think that's the risk. If the risk is in technology, like can it be built or is the technology possible? I would hire somebody in engineering. If we think the risk is in, is the problem actually one in which we think we understand the scope of it and we're close enough to the customer, then I would hire a product person. If we think the risk is in how the solution works, I would hire a UX person. So it actually depends on what the risk is for me to decide who to bring onto the team first.

I have a completely different answer to this now.

I was gonna say it. So again, times have changed and the type of trajectory, like we go back to foundation, like again, 10 people to 50 people to a hundred people, it's the things change. And what's your take on that clip? Let's go back to it.

Yeah, so yeah, that. That's a cool one to show because I have a different answer now. So two things. Whenever you hire someone, what you are admitting is that you are have failed at performing the duties of that function. Right. So when I go and I say I have to hire somebody to do sales or someone to write code or something, like I'm basically saying the current team, including myself, is not up to the task of doing this function. So you should start by saying we have failed and we should all be sad, then we should say, okay, one solution is if we think it's important, should we. Can we stop something else and someone here should do that thing. If we still think we cannot do it, we should be sad. Again and say we have failed at being efficient and let's go hire somebody. So this is like a penalty. Oh my God, we're going to have to increase costs. It should be a very, it should not feel like a good celebratory thing. Oh my God, I get to hire somebody. It should be like, oh my God, this really sucks. I can't believe I have to hire somebody. And like we really are terrible now. The difference that I have now in hiring is I wouldn't try to hire a specialist. I would likely try to hire like a generalist who can solve a bunch of the problems. So for example, the best people to sell the product in an early stage company are the founders, right? You shouldn't have, you shouldn't have somebody like have to come in off the street and be a GTM specialist and learn about the product and learn about the problem and learn about the like how it was. Learn all those things. It should be the founder selling very similar to if you need someone to talk to customers, that should be the founder. And then if you need someone to write the code, that should be one of the founders. Like all of this should be founders. And if you again realize that you have failed and need to hire somebody else, I would try to hire somebody with that founder mentality. And there is that, that essay going around, right Founder mode, right By Paul Graham. I would more likely now want to hire founders. And by the way, this has really been shaped by watching Toby and how he's thought about building his executive team. What I realized, I think somebody tweeted this, but I couldn't find the tweet which is something like Toby collecting founders like Pokemon cards or something like it was literally like he goes out and finds founders to run various. Like I run all of engineering and he bought the company so that I met my co founder. Daniel is running partnerships. He's like, I just need founders around me because founders can kind of do anything. I don't really have to stick them into a certain box. I can say like one good example was in 2022. Toby and I were pairing like along other people too. But we rebuilt the compensation system right? Like now like two non comp experts like Toby and I shouldn't really be doing that. But at some point in that exploration of compensation, I'm pretty sure we knew more about compensation than most people. And we built I think one of the world's most innovative compensation systems that like we rolled out of Shopify and very surprised nobody else has copied it. So there are this Founder thing is real. I think there is a little bit too much of like a bar to clear. Are you a founder? Whatever. But founder mode, you can be an employee and you can be inside of a big company and you can still operate like a founder. It just means, like, you are typically a generalist. You can dive into all sorts of different problems. You have a systems thinking approach. Hey, that problem I solved over there, I can use a similar solution over there. Use the system to solve it. It means you take extreme ownership. It means you get into the details. That's what founder mode means. It doesn't mean you actually have to be a founder, but. So I would change my answer. Like, I would really be like, I would be reluctant to hire, which I think we did a good job at. Helpful. We didn't. We were not quick to hire. But I would try to hire like more founder types that can work on lots of different types of problems. And I think we did that again. But I would change my answer because I think that most people should think about generalists versus specialists. And there's a great book range by David Epstein, which is one of the books that probably was the hardest for me to read because I always felt I was like a specialist. And I read it and it kept saying how generalists were better and I kept thinking this is wrong and should always go deep versus going broad. And took me to like the end of the book to realize that actually I'm a generalist and that he was right, that like, you actually want people to be more worldly and have all sorts of problems and solutions in their tool belt and then to use those to solve the next generation of problems versus being narrow in a field and saying, okay, I'm a math major and I can solve math problems. And that's.

So you yourself were a generalist. You just put a moniker on yourself, like, I'm a specialist.

Yeah, yeah, it was funny. I always thought this is. Toby convinced me of this as well, where I kept fighting it and he's looking at me like, you worked like you. You went to fix the recruiting problems and then you went to fix the HR compensation problems and became an expert. Like, why do you think that you're not a generalist? I was like, I just thought of myself as an engineer trying to fix systems. And my solve in the past in my startups has always been to have finance, engineering, recruiting, HR reported to me. And I just thought that was like a normal thing to do because I'm like, oh, yeah, I know how to steer these Things. But it turns out I was using that mindset to fix problems in lots of domains.

And so what do you say to the founders that are engineers? They're not salespeople. They're not. They're introverted. They're not extroverted. But as you said, and I agree, like, the founders are the ones that need to sell. Nobody knows your business better than you. Nobody knows why you created this product to solve this problem. Yeah, but they're not the salespeople. They feel intimidated or they don't have the sales skills. And they say, I need to bring in this VP of sales from Shopify because look what he did at Shopify. He's going to do the same thing at my company. Yeah, I know my answer, but I want to hear your answer on this.

I'll say two things. One, and again, I'll go back to what I saw on. I saw this tweet on Twitter, and it totally resonated with me, which was, you can learn extroversion. I'm actually an introvert. I get my energy from being by myself. Now. I'm a very good fake extrovert. I can go out there. I can go to a party. I can be like, the shooter and walk around and meet people and talk to everybody, but I'm, like, faking it the whole way. And actually, if you saw. If you knew me Until I was 16 years old, I was extremely shy and introverted and just sat on my computer and I just made a concerted effort to be like, I don't want to be like this. I want to learn, and I want to learn from people. And the only way to do that was talk to people. And so I changed. I forced myself into this extroversion. So that's number one. And then two, on the skill side, you could do the same thing. I'll give you an example, actually. When we weren't helpful, by the way, my co founder was Daniel DeBeau, a phenomenal salesperson, phenomenal speaker, and, like, persuasive, like, phenomenal. In one of our first big pitches, he made me do the pitch. And I remember exactly what happened. He told me how he wanted me to pitch it. He told me the pitch, and I took it. And I. And I told him I did this. We had these Nest cameras in the office that were like, for security. I grabbed the video clip of Daniel doing the pitch to teach me how to do the pitch. And I learned it word for word. I played it over and over. I think I might have played it 50 times. And I Literally wrote down, this is. Obviously, I could have used, like, drag. This is before, like, LLMs could transcribe it. I literally just wrote down it was part of the. Part of it was the routine of listening and writing it and learning it. And I literally delivered the script word for word, which is not my style. I like to go off the cuff because I. He. I thought it was so good and so well put together that I'm, like, just going to copy it. And so I didn't say, no, you should do it because clearly he's veteran. And by the way, he, like, he should have done it, but he was like, teaching me. I'm going to teach you how to do go to market. And it was phenomenal. It was like a. Like, it was an amazing experience for me, but it also showed me that I could do it. And so I did that pitch. And then the same thing. We always tried to just do the thing that is required at the time. Even though. Here's another example. We say we were pitching to a bank and we had to build a security deck. Like here, like, there's gonna. They're gonna ask you, what's your security story? And we didn't. Obviously can look up security decks online, whatever. And we just sat down and we said, let. Let me tell you the story of our security. And I sat down with the. One of the other co founders and we built a security deck we know anything about. We're not CISOs.

We didn't know a lot of security.

But we built it. And by the way, we passed the security review. And so I'm just saying we tried to again, use the now there's a label founder mentality, founder mode. But we use that to say. I would say now change. My advice is hire generalists that are very smart and intense and have done things, or you can figure out if they've done things very quickly and bring those people, types of people into the company.

Yeah, I think specialists do have a place. Of course, there's people that have really honed their skills, but as a founder team, bringing in a $300,000 salary, VP of Sales because you think, okay, this person's gonna solve all my problems.

No, very rare that that's gonna happen.

Very rare that that's gonna actually work out for you. And if anything, you just killed your Runway because you just paid this guy.

Oh, no, for sure.

Huge salary. And they're in retirement mode because that was their cash out. Right.

Yeah. I would say another thing. I would say always do the job before you hire for the job. Right. Like I, I extreme labs. Like you mentioned at the beginning, I was the first recruiter. And I recruited until I was like, okay, I've done this now long enough. I know what this person can what can what I can get leverage now. So I hired someone and taught them how to recruit. And I said, here's how you're going to recruit. But because I did the job, it made it easier for me to then outsource the job after. So if I want to hire gtm, I got to do GTM first.

Yeah. Filling in redundancies. Right. How do I replace myself? But at the same time, can you replace those other components of your organization?

Exactly.

So we focus more on the employer side from the company side of things in remote. Let's get to the employee side. I've got another clip for you here.

It's useful to have goals in support of some larger theme. So I don't think a good goal is I want to be promoted, but I do think a good goal is how do I increase the impact I have on my team? How do I help bring others up alongside me? How do I ensure our merchants are getting value from what we build, Whether it's they're saving time or they're able.

To grow their business.

Like, all those things are good themes that can have goals attached to them. To say, okay, I'm working on this thing. Merchants spend 5,000 hours in this one area. Can we reduce it to 3,000 hours? That's a good goal, but it's in support of a theme which is let's help our merchants save time.

I love this clip because I can only say that I would double or triple down on that statement. We, we are very much not a goal oriented company in the sense of, hey, we like revenue goals and profit goals. Like, we very rarely. It's actually amazing how rarely we talk about those things because we are here to serve merchants and to increase entrepreneurship. And if we focus on that, everything else will come, right? Like, it's this weird. Not like in a karmic way, but in a, like, we're on the same side of the table as our merchants way. So, for example, some people, this is actually one interesting thing. If you look at a business who uses Shopify and you look at their stack of software, we are usually the cheapest part because we are the part that we make money when they make money. And so if we work on enabling them to jump, make the jump to entrepreneurship or save time or find buyers or sell in a new market or Go from online to offline, like selling a retail location or advertise more effectively. Like all of those things will make them make more money, which is good for them because they want can, you know, reach for entrepreneurship. And then it's good for us because then that helps us funnel whatever money comes from that back into the first thing, which is like increasing entrepreneurship. So I would double or triple underscore that original message, which is don't have very specific goals because it's very easy to basically overfit the metric.

But switch yourself to the other side of the table. If you are interviewing to find a new career path or a new company to join. What's that mentality for the next five years? Like, how should you approach this?

This is the classic play long term games with long term people, right? So for example, I'll give you an example. I remember when I was leaving, we had stole ExtremeLabs to Pivotal and I was starting to think about what do I wanna work on next? And I made. I met this amazing entrepreneur who had just raised money for his startup, I think it was series B maybe, and said, hey, I need to like, I need a number two in the company. And I love these like amalgamous like job descriptions. Cause I'm like, number two means like do anything, fix everything. I love that. And the thing that turned me off on the pitch was he said, oh, I just raised like $10 million and I can pay you a lot and you'll make a lot of money. And I was like, wait, what a weird goal. I can see how it's okay if that's your framework, right? And it seemed like such a short term incentive for me. And I was like, wait a sec. My goal is not like to make. Of course in some place you have to make money. I gotta pay a mortgage and I have kids and like you have to make money. But my actual goal was like, how do I learn from smart people and how do I have an impact in the world and how do I ensure that? I'm like, what are the most interesting problems in the world and am I, am I working on one of them? Those were my goals and I always felt like the rest would come. And again, not karmically, but actually because if I worked on those things, it would make me valuable and that value would somehow accrete part of it to me via like compensation. I think the same thing is true of like, we pitch people who come to Shopify with the same idea, which is like our goal here is to increase entrepreneurship in the world. We work on hard technical problems in engineering and in product and in data and ux. And our goal here is to allow people to reach for entrepreneurship and make them successful. And if they can be successful and we can teach them the skills to become successful, then we'll be successful. And so I don't think the short term like you'll make X by Y actually works for those mission oriented people. Right. This is the, you probably heard the mercenary versus missionary type framing before and I don't want to get too like religious about it, but like it's a real thing and like finding people who are long term. Actually, I'll give you a funny quote. When we were looking at being acquired by Shopify, Harley had this crazy line which he said, what's amazing about Shopify? I said, what? He goes, year six. And I was like, what do you mean by year six? He goes, the sixth year that you are at Shopify is amazing. And maybe he knew, maybe he didn't. I'd never worked anywhere for six years and now right now I'm about to hit my six years at Shopify. So it's just a funny thing where he's like, it just gets better over time because you're continually breaking down the barriers for people to get into entrepreneurship. And then we're teaching our merchants the skills in order to get there. Like we, we really do. Maybe I'll make a little bit of a plug. We really do think our software is like a video game where like, you know, in Mario at the beginning you learn how to jump and then you learn how to get the mushroom. And we want to teach entrepreneurs the skill and when they pass level one. The goal for level one, like the, the reward is you get to play level two and then the reward for level two, right. Like now you're like, okay, I built my product and I started my website. Cool. Maybe now I need to learn how to do advertising and then you learn how to do advertising and then now you've mastered level two. And what's your goal? What's your reward? I get to play level three. And that's literally what we play here. We play this long term game where the reward is you get to keep playing.

Yeah. The way I see it is for the entrepreneur. I'm an entrepreneur. True and true. You as Shopify for E commerce entrepreneurs. You're helping me fall in love with other aspects of my business.

Right, Right.

I'm in love with my product, I'm in love with my customers. That's what I'm doing, and then I've mastered that. And, oh, there's this marketing feature, or there's this distribution feature, or this is a customer support feature. And I can use these tools to love that aspect of my business.

And I think that's exactly right. We want to take away the undifferentiated part of your pro. Your entrepreneurship journey, Right? Like you're building. You should build amazing products, spend your time there. You want to curate an amazing community, spend your time there. Let us take care of the other parts. And so we don't have these weird incentives like, you should have a lot. You should spend a lot of time in our admin. We want you to get out of the admin. We want you to go back to selling, talking to customers and building products. And so our interview process tries to discern and find the people that are here for our mission. We don't need everyone to work here. We actually don't want everyone to work here or apply here. We're like, it even I think if you go through the process, it says, here are the hard things about working at Jobify. Here's why you shouldn't apply. Because we want to make sure we heighten the misalignment. And then when they realize, no, actually I do want to work here because I care about this mission and I care about working with smart people, and I care about working intensely, and I care about working on hard technical problems at a scale that the Internet has never seen before. I want to apply. Those are the kinds of people we're working for. So we try to make it actually difficult. And we've even toyed with the idea of hiding the careers page further away from the homepage so that you can be like, you really want to work here? You got to find, yeah, not in a gamified way, but in, like, you really want to know, you really want to, like, work here. Like, you will find, you got to find a way in.

Hold that thought. I got another clip.

Oh, geez.

Transition. This is all about hiring now.

I tried to not rely on interviews and instead do job trials. And so I did. I hired lots and lots of people with the explicit assumption and explicit notion that, hey, we're going to understand how each other works over the first 90 days. And if you think we're a fit, and we think we are a fit, we'll keep working. But if we don't think you're a fit, we'll give you feedback. And if you don't think we're a.

Fit either, give us feedback.

And then there's a no harm, no foul rollout in the first 90 days. And so that's not how we do it at Shopify, but that's how we did it at extreme. And it was a very easy way to have a conversation with somebody and really test the work versus the interview. Now at Shopify, for example, we look for something completely different is we look for like a big range. We want to make sure people bring a diverse set of deep skills. And so we test for that. So, for example, we have this interview called the Life Story, where we try to figure out not what somebody did, but why they did all the things in their life. And we believe that's a predictor for performance at Shopify because they've got a deep curiosity in a bunch of different areas versus looking for particular keywords on a resume.

There's some things there that I can expand upon because we still do the Life Story at Shopify, we have updated the life story to continually understand what are the things we care about and test for those things. And on the early talent side, we do almost like what I say there, which is, maybe I'll step back one sec. So we're going to hire a thousand interns in 2025. And the idea behind that was one, I want to meet more people who come to work with an LLM and a brain than just the brain, because the AIs are everywhere now. And we want to make sure that you're growing up with having this pair programmer, AI person that you can work with. And then two, the internship is an amazing job trial. It's four months. It's intense. You're going to learn about us, we're going to learn about you. And you might love it and be like, wow, I really want to come back and work at this company and I've crushed it. And then come back and do another internship. Or maybe you want to come back and come back full time. You may realize that you're not aligned. You're like, whoa, this is completely different type of company. It's. I have to have more range than I'm used to. I like to be in this more narrower band. Or it works at an intensity that I'm not used to and I want to do something else. Or I've done spend a lot of time in software and I want to actually move into another discipline. There's all kinds of reasons why it may. You may feel like it's not the place for you, but it's very much like a job trial. And those Thousand people can be an early talent pipeline for us to ensure that we are finding the best and brightest who are aligned with Shopify. And so I think that that ties in nicely with that view of trying to use the work as the byproduct versus the interview. Of course we have interview steps because I think we get like a hundred thousand resumes a year. So we couldn't, we can't try everybody, but we're trying to figure out a way to use the real work data to make hiring decisions versus just like interview data.

Well, this is the Next Wave podcast and so one of the next wave of tech is AI. It's that you can't escape it. So how does this factor in? You said like, you want understand, bring an LLM with them and they understand how to use it. So how does this play into the interview process to find the talent that embraces the next wave of tech as opposed to puts up walls against it?

Yeah. So two. A few different ways. So one is when these AI co pilots first started arriving on the scene. I remember internally, and again, this is normal. Any company, they're like, hey, I just had my first candidate show up with an AI copilot. And I told them they couldn't use it. And I was like oh, like, why'd you say that? They finished the question in three minutes. And I was like, cool. So do you think it's the fact that they had the AI copilot that was the wrong part of this situation? Or was it just that fact that maybe our question was so easy that a human and AI co pilot could finish in three minutes? And maybe we should have a harder question because maybe now they can just build a much more deep solution or build something else that's more complicated because. Or solve more complicated problem because they've got an AI. This is the rethinking we had to do. And so we were, I think, one of the first to allow LLMs in our code in our interview process. And we just went with it and we said, hey, if they finish it quickly, like add requirements to your problem, pivot the question, make changes. Let's try to figure out how people think. That's one. And then the second thing we did was we started using AI to evaluate more people. So for example, I told you we have a hundred thousand people apply. We can't talk to as many people as we'd like to, so we have to use some sort of resume filtering. And then that gets to like an, like a synchronous conversation with somebody that's not ideal because resumes are kind of garbage, right. They don't tell you much. And again, and you saw them in that clip, they tell you what somebody did, they tell you why they did it. We can use AI now to have people submit asynchronous, like coding sample coding challenge. Here's an open source project I worked on. Like, we can and we could use an AI to evaluate whether we think that the type of engineer, in an engineering sense that they are actually lines up with what we're looking for. And so I think that's a good way for us to use AI on the inbound funnel to help us to have more people get an opportunity to break through and have a conversation with somebody at Shopify. So we're using that in two ways. And of course, I'm starting to get. Maybe this is controversial. I'm starting to get skeptical when somebody comes to an interview without an LLM. Right. And actually, maybe I'll give you one anecdote. So I do a lot of like, shadowing of interviews. Meanwhile, meaning like the interviews remote. We'll both be on the call with the candidate, and then I'll tell the candidate, hey, by the way, I'm going to be watching the interviewer and I'm going to turn my mic and my camera off, so feel free to ignore me in the call. And what I realized was I was doing that with the interviewer. But what I started to notice is if a candidate didn't have an LLM, I would be like, oh, cool, I'm gonna just do the question that the interviewer is asking of the candidate at the same time as the candidate. If there was no LLMs on both sides, the candidate usually would beat me because, like, I'm rusty and I don't code all the time. But what I noticed is that even for principal and distinguished engineers, who are pretty technical, if I had an LLM, I would beat them by two eggs. I would finish twice as fast, and I would send my solution to the interviewer for grading as well. Cause I'd be like, hey, you should look at my solution versus the candidate. And it was pretty clear that me plus an LLM was better than the candidate. And so there's something interesting. If a. If a candidate does not come with an LLM. Now it's possible they're not allowed to use one at work or whatever, but it's getting weird. It's like not using spellcheck, right?

Yeah. In an academic sense. Yeah. Okay. You need to show that you can do the work on your own without those tools and resources. But in the real world, in the office, it's the fastest to get the problem solved so we can move on to the next problem. And so if you're not using those tools intentionally, then you're actually hampering the company and hurting the company.

Yeah, but what I would.

There are some scenarios where you can't use those third party tools, but use as many tools as possible to get your job done.

Yeah. What I would say is it's going.

To generate code and you of course.

Still have to understand the code that it generates. And so this is like another skill. Like read. Reading code is actually more of an engineer's job than writing code. So by actually having the LLM produce code and then having the engineer have to look at the code and see if it's correct and evaluate what the.

LLM has written and understand what the.

LLM has written and then go, oh, I have to modify. These are amazing things to test for in an interview because they're exactly like the job.

Yeah.

And so I think we uncovered like a sneaky way to actually look at a bunch of skills that we need anyway. And so I think that it's kind of harder the other way. It would have been like, hey, here's some code. And there some interviews are like this, here's some code, tell me what's wrong with it. But here you've had the candidate do the prompt and then the code is generated by the LLM. So completely nothing to do with us. And now the candidate has made some code and they have to evaluate the code that has been made. It's like a beautiful story for us to then walk through with the candidate, be like, why did it do that? Why do you think? Does it work? Does it not work? How would you test it? What is it trying to do? Would you have written code this way? There's so many avenues to go down. It's like a beautiful melding of the two things.

And now I want to skip past the hiring process and into the actual real work environment. You've alluded to pair programming a few times. We've got a clip on that. And I want to dive into what pair programming looks like today.

I'm a huge fan of pair programming.

I think it is the single most underutilized engineering technique in the world. And it reminds me of the underhanded free throw in the NBA where it's proven to be more accurate and easier to execute, yet few people do it because it looks dumb. It's very similar to pair programming. I think the, the way to phrase it for this person is like, are there tools? Yeah, there's great tools. I saw a few come out last week, I think from. I saw it on Hacker News. I think something called CO Screen and others. We use one called Tuple T U P L E App. We use that one exclusively at Shopify and it is a great way to pair program. I know some people use it for not just programming, but other tasks as well. And we have a cool innovation inside Shopify. There's a slack group called Pairing Sidekicks. And if you join the group, you will automatically once a week get scheduled to pair program with somebody for an hour. And it is awesome because it allows me to just meet different people in the organization and pair with them for an hour every single week. And so I encourage people to do that. We, we have people who pair eight hours a day using Tuple. We have people who pair like one or two hours a week, whatever you want to use. But it is quite good in high fidelity and it's a very small team. I think of four or five people who build that tool.

Cool.

So the underhanded free throw is not sexy. But is pair programming sexy now? Because of all the AI tools that.

Are coming out, pairing with an AI is sexy. And so I think we've jumped the shark on. Hey, people feel like, oh, pairing with a person, like, it's not that cool, whatever. But what's funny is there were times when people would pair. Like, so when I worked at Pivotal Labs or at Extreme, there was this funny thing where they call like, pair with.

If.

If you had an odd number of people, let's say you had seven people in the office that day. You can't pair with seven people, right? So there'd be three pairs and then the last person we would have this thing where like they pair with a banana. You'd put like another chair and have a banana stuffy beside them and they would talk to the banana. And what's funny is actually just the process of talking through what you are working on. It's like when you call somebody like, hey, I have a problem and you start explaining it. You're like, oh, I figured it out because you're explaining it. Yeah, same thing would happen. Pairing with a banana actually works. And so what's amazing now is that everyone gets to pair. No one can. No one has to program alone anymore. You come with a. You come with your LLM of choice and you're now already pair programming with the AI, I still believe so. One that's amazing. I still do believe pairing with another human does add an additional different layer of pair programming because they, they won't be as quick as the AI, for example, to dump out like dozens of lines of code. But the thought process, the thinking through, and now what I've seen people do is they go, hey, I'm going to be pairing with, I'm going to be pairing in cursor with a Claude model for half an hour on this problem. Anybody want to join? And so now you have two people and an AI, and together you're like prompting and pairing and working through a problem. It's such a phenomenal time to be working in software engineering because instead of being afraid of this tool is going to take my job, it should be like, oh my God, I've got this amazing tool and brain to help me cut through the machete, through the jungle, and I can actually get farther down a problem that I could have because maybe I don't know the language that well or I don't know the API or I'm importing a library and I don't know that API. And so, like, there's so many reasons to, like, be so excited about software engineering. And so I'm just as bullish or more bullish on pair programming now than I ever was, especially with pairing with an AI.

So that reinforced learning is still there with the human aspect and having a partner there, you're just now making yourself 10x more productive because you have the AI to cut through all the mundane work or the iterations.

Yeah, well, I mean, I think it's like level one and level two. Level one is like pairing with the AI. And by the way, you can do cool things. Like, I tried this a little bit, it didn't really work, but I'm hoping it works now, which is like, turn on the voice activation. And I saw somebody using like whisper to prompt. And so they weren't even using their keyboard anymore. They were talking to the AI, it generated code, they looked at it, then they prompted again to fix errors. Like they never actually touched the keyboard. So level one is paired with the AI, and level two is grab another human and together pair with the AI. Because now you're going to have even again working with another smart person. You'll get even more feedback and go back and forth around solving that problem. So again, I'm even more bullish on pair programming. I will state again that it is the most underused tool in all of software engineering and still most People do not do it. I think more people doing it with AI, less people do it with another human.

That's. This has been a fascinating conversation so far and I really love diving into this with you. I want to circle back and close off with back to remote working and hybrid situations. How does cybersecurity fold into the mix here? Things are changing rapidly in terms of attacks and breaches and whatnot. So what are your thoughts? Closing thoughts on remote work but making sure you're securing your environments.

Yeah, so actually what's funny is I just saw a paper, I think it was yesterday, released by Google that I think was their Gemini model, found a vulnerability in SQLite that was patched the same day. So many parts of that story, one, the AI found that vulnerability, but then two, they, I don't know if they patched it or not, but it was, it was patched the same day. So like everybody wins. The AI found it. The patch came out the same day before it went to production. The world is better across the board and so I really believe that we can use the AIs to find these vulnerabilities. Now of course bad guys will use it too, but if we can use it to be proactive to find the vulnerabilities, we can fix them. And it's actually great. It fits into the skill set of these AI tools to find these types.

Of things and then we can, it.

Can of course suggest a fix and then we can verify the fix and then also roll it out. But I think that, I don't know if it changes really the stance like you needed cybersecurity infra whether you're remote or not. So I don't know if it changes it that much, but it does require us to embrace it for, for these like forward thinking workflows. Hey, do you think, do you think.

Non tech companies are taking cybersecurity more seriously now because of these remote environments and.

Oh, I see what you mean. Yeah, I think yes, because there's this idea of like bad actors who don't have to come into your office and bad. And then there's also, hey, person A interviews but person B comes to work the next day. Like super weird things like that. Or you give a programming question to person A but person B completes it. So there's all kinds of weirdness like that. Hopefully we can find those things using like better tools. I am a little worried that people overfit on like the monitoring side. Oh my God, companies got to monitor my work more because they're worried about this I'm not as worried about that because like, good work is good work. But yeah, if you're trying to work two jobs, like, I hope the software finds you.

There's a lot of stories around that.

I know.

Farhan, this has been awesome. I want to give you a chance to have your ask, what can our traction audience do for you? Do for Shopify.

Cool. Number one, I hope that anyone who's reaching for entrepreneurship reaches for Shopify. And if it is not the right tool for whatever reason, I'd love to get the feedback. Like I'm pretty accessible on Twitter. I'd love to hear like, why Shopify is not the right tool because we're here to solve problems and the worst thing for us is to be the best at solving a problem that nobody has. And so we do hope that if you have a problem and we hope we think we're the right solution and we're not, tell us that's number one. Then two, if you're early in your career or late in your career and you want to work on interesting problems, check out Shopify Careers because we are always looking for the best and brightest and we are remote. So you can work wherever you want.

Any tips for those people to find the career page?

It right now is still right on the front page, but I would say we're looking for a certain type of person, right? Curious learners with range who are super interested in hard problems and who want to work with people on those problems. And so we have this like Toby has this great line like, we're on a difficult journey solving important problems with friends. That's what we are.

Can't ask for anymore anything more.

I agree.

Farhan, this has been awesome. Thanks for your time.

Cool. Thank you.

This episode is brought to you by Boast AI. Each year the US and Canadian governments give out billions of dollars in R and D tax credits and innovation incentives to fund businesses like yours. But the application process is cumbersome, prone to frustrating audits, and receiving the money can take up to 16 months. Boast AI gets you access to research and development tax credits and innovation funding opportunities without the headache and red tape. Join thousands of North American Companies leveraging boast AI software to maximize cashback. Check out www.boast.AI.

I need some traction. You need some traction.

Thank you for listening and we hope you enjoyed this week's episode of the Traction podcast. If you enjoyed the show, please leave us a five star review and you can find more information and all the resources mentioned in today's episode. At boast AI that's b o a s t a I forward slash blog.