Bootstrapping to IPO, product-led growth, & scaling SaaS with Atlassian's Scott Farquhar | E1800

31:24Darmesh is obsessed with AI. Are you obsessed with it now too? And what impact is it having in the organization here as we move into this year one of chat GPT, language models, and, you know, an actual platform starting to emerge that can be used by, you know, any business user can just use any number of these language models themselves and you get hugging face throwing up new changes every day it's got to be in your consciousness huh yeah anchor ai in the way i think about it in technology because it's so much of a winner takes all market it usually takes some sort of technology shift to shuffle the playing board and in those technology shifts you find some companies that endure you know across multiple of them and some that stumble and if you look back historically you know back to at least my lifetime uh you had this sort of move to desktop software which obviously you know microsoft and windows and office were you know kind of the big beneficiaries of that and then you had you know the sort of turn of 2000 you know 1999 2000 when you had the internet came along and you'd say you know netscape netscape was uh you know the big window there and microsoft played play catch-up but it it birthed you know companies like google um that wouldn't have existed previously.

32:41And then you sort of roll through to mobile and, you know, that birthed Apple really as a company. And, you know, in Google, you know, sort of with Android did a good job there, but Microsoft wasn't anywhere really to be seen. Then cloud came along, which is not so much a consumer product, but Microsoft, you know, caught up heavily with cloud and suddenly Amazon's, you know, in the race as well. And so you see these technology shifts where a couple of the big players maybe make the shift and a couple of players don't make the shift and i think ai is the next big technology shift that's going to shuffle the playing board of technology and uh when i look at you know i've gone deep with all the different large language models and how i know how they all work um in some ways the industry has been saved by large language models because if you weren't deep in ai you can basically rent a large language model from someone which is very different i think to how most people thought this world would play out And so there are a lot of companies who are really just playing catch up effectively by running these large language models who didn't have investment.

33:43So it really is different than most pundits would probably thought it would. I think the real value is going to come from putting data together with these models because the large language models at the moment, I view them a bit like a Swiss Army knife. They're expensive. They're not particularly good at any individual thing, but they do a lot of different things for you. and that's great everyone's going to use them because they're the first thing available and i can pick up one of them and use it for lots of different use cases uh i think over time there's going to be specialized models um you know if you just want to do language transformation which we do we convert text to a query language we don't need a you know multi-trillion parameter model to do that and we can have much faster and cheaper and uh you know even you know cheaper and faster are pretty much the same here and so we can actually have better user experience and save ourselves money so over time i think you see specialization of particular niches or niches i think is it said in the us yeah either of those are acceptable we'll accept both of those as answers for niche uh or niche you're allowed to say both yeah but that's that's i think that's the that's that's the correct answer is you know you right now the swiss army knife i love the way you put that i can go in there and ask it about travel or to write me a blog post or throw in some code and clean it up great great swiss army knife but if somebody makes a verticalized thing like you know the you know github's copilot or stack overflow is working on one or somebody works on something just for finance or just for travel of course it's going to have a lot of features wrapped around it and a lot of verticalization and then reinforcement learning and it's going to blow away the general model that that's i it's that should be pretty obvious i think it's pretty intuitive to think that's going to happen and those are going to start showing up in the next year probably especially with all the open source uh projects so you're uh in open source a lot of your success is based on open source so do you think the proprietary models like closed ai previously known as open ai is pursuing what do you think is going to win the open ai model where it's closed uh or the actual open source models you know i guess uh facebook's lambda is uh open source and other ones that are coming out uh which one's going to which one's going to win the day?

35:59I don't think that for us, where we sit in the ecosystem, it really matters if one or other people win the day because they're so interchangeable. For me, it's an API call to use a large language model. And so unlike saying, well, how hard a choice was it to choose between Amazon or Azure or Google to host in the cloud? If I'm going to switch between one of those things, and I've invested millions of dollars, you know, right into their APIs and working that way, but the switching costs are very high. The switching costs, you know, for large language models are very low. And we've, at the moment, partnered with OpenAI because they're the best and, you know, we've worked with them even on their contracts and how they store data and making sure they're much more, you know, B2B friendly.

36:44We think that we'll be probably using lots of models over time and the real value is going to come from putting data together and having those, you know, data use cases. And this is where I think the landscape is going to change a lot because if you're a very small point player at the moment, you do one very, very specific thing for customers. I think you're going to be at a relative disadvantage compared to the larger players out there that do multiple things. Because if you do multiple things, you have lots of data across the lifecycle of that customer. And that can mean your experiences can be way, way better than they could be initially.

37:20And so I do think there's going to be a bit of a, the big get bigger in this next phase of the internet. I'm not sure if that's good for everyone. I mean, it's great for Atlassian. I don't know how good it is for society. But I do think that the, you know, the use of data is going to be the big win here. It does seem that you could switch these out really easy. I've been playing with it. And you can very easily, you know, send your prompt engineering through one and then try it in the other. Look at the response teams. I think there's going to be some meta layer here, kind of like CDNs, or, you know, I guess some people with their cloud providers create, there's a term of art for it where you can have your, you know, jobs in the cloud go to different providers, you go to AWS for one thing, go to another, and have that done.

38:05Multicloud, yeah. So you can have like a multi-cloud thing. I could see people having like multiple, you know, chat GPTs and just go to the language model that you think is best for this current job. And Meta's model obviously is Lama. Google's is Lambda. Founders always ask me for pitch deck punch-ups and how to present their startup in a better way. Well, I've got some great news. We worked with a team at Miro, the awesome whiteboarding software to create an amazing pitch deck template for founders. you can see it if you're watching the video right now or you can just go search for it you go to miro.com slash miroverse and search for pitch deck you'll find it immediately and this pitch deck will help you go from zero to vc ready our founder university participants they love using this template it starts them on second or third base and if you're hybrid or fully remote miro is incredibly useful to you it's like an old school in-person whiteboarding session but distributed and asynchronous so you can work on your time schedule.

39:02Miro lets you brainstorm ideas and collaborate on projects from anywhere in the world, whether you're in the Adirondacks or you're in Cabo or you're skiing in Lake Tahoe. When you think about Miro, think zero to one, but faster. And Miro is so much more than a simple digital whiteboard. Your team can collaborate on planning, research, design, and feedback cycles. Now remember, faster inputs equals faster outcomes, and velocity is how startup wins. We look for product velocity in all of our startups. So to access our new Miroverse template and thousands of others, sign up today for a free Miro account at Miro, M-I-R-O dot com slash startups.

39:38Again, M-I-R-O dot com slash startups, Miro dot com slash startups. Yeah, it does seem like they're already getting commodified in some way. And then whoever has the data is going to win the day. You have all this data on, you know, I mean, that's another thing with a long tail of 260 ,000 customers. you've got a lot of data and you can provide a lot of speed efficiency if people are trying to manage tasks or clear out bugs or whatever they're doing in the software that you provide you're going to be finishing their sentences huh and being their co-pilot any of those products released yet have you put any in the wild or you're in the laboratory right now yeah a couple things one is i think that from a b2b sense like they'll be interchangeable um we all know that consumer behavior is something that you know is harder to change and so you know there might be a better a search engine but if i'm used to using google every single day am i going to try a vertical search engine for that one search to do something probably less likely so i think there is you know and i think open ai has probably got the lightest consumer uh side of things so i think it'll be you know from a b2b sense yes there'll be a lot of providers i think it's still open how the consumer side will play out and if i go back to alasian strengths uh you know we have knowledge about how teams work and uh you know how strictly around engineering teams and development teams and so So we have information around what code gets written, what customer problems happen, what bug reports happened.

40:59And I keep challenging my teams to not just improve the way that our current customers do something, but really sit about the job to be done. And so take an example like customer support. No customer wants a faster customer support experience. They want to have not had to reach out to customer support in the first place. and uh you know if you take it to the example of a app you know on your phone it's like well why don't we send all the log files and all the errors from that app on your phone you know every single night like uh you know back to the developer and you know previously if there were minor errors in products you just couldn't find the needle in the haystack and or fixing them would be too expensive but these days with you know the with ai you can now see oh actually this little bug here i can even see the code it creates that bug and to fix that bug is a trivial you know change and so i can you know change the code and maybe it's a human gets reviewed it maybe it's a different large language model that has a whole different training set reviews it so you've got two different sets of eyes you know reviewing that code and suddenly your code gets more robust and you know sent out based on uh you know real life data that comes out from the field so instead of your customers having to file a bug report they actually just never had to solve the bug in the first place And I think that's what we're going to see over, you know, in the short term, we're seeing great party tricks like, hey, turn this list of three things into a list of 15.

42:25That's great. It looks good. It's kind of cool. But the real value, I think, comes from totally reimagining, you know, customer experiences. How far away do you think those will be? We've had this parlor trick, like, incredible. It's helping me write the blog post. I make it shorter. I make it funnier. And, hey, it got 60 % done, 70 % done. I polish it. Wow, this is crazy. I don't need to have a PR firm write a press release. The AI wrote it for me and I just polished it. So for a young startup, why hire some PR firm if I can just, the CEO or the person running product can just become a bionic.

43:00So when do we see what you're talking about, which is, hey, we're going to intercept bugs. We're going to have multiple language models, review it. And it's sort of like precogs and minority report predicting what's going to go wrong and solving it in real time. That's trippy stuff. Is that five years out, 10 years out, two years out? I don't think that's five years out. I think that we're working on things like that at the moment. And to take a more knowledge work example for, you know, the listeners who use that don't write code, you know, if you go away for a long weekend, you come back on a Tuesday, and you're like, hey, I want to catch up.

43:33What did I miss? At the moment, that is a painful process because you go, well, what do I need to read? What's important? What got actioned already that doesn't need my response? And because of the data we have on Teams and teamwork, we can say, well, Jason, actually, your closest peers over on Monday all collaborated on this document. By the way, here's, at the end of the day, here's the change from X to Y. And I don't need to do that as a diff with red lines. I can actually explain the changes in human readable form for you. And I might be able to explain it in one sentence or two sentences.

44:05And if you want to get the three-paragraph version, you just click. And if you want to see the actual red line changes, we can do that. And so, the ability for a knowledge worker to catch up on work, uh is huge and if i look at at least my my time and i'm sure your time is like reading through or even just understanding what to read through is like a huge burden for most knowledge workers and that's the stuff we're working on at the moment uh that's absolutely brilliant and i'm having this experience we started recording every meeting we have with founders for funding then we zoom gives you the transcript so you kind of get that for free then we put it into notion notion's got ai built in and we said summarize it and so now i get these little summaries hey the investment team met with this person they talked to the founder about this the founder said this they're talking about a term sheet there's a liquidation preference i mean it's scary how accurate it is and and it's saving me having to do that catch-up it's like having jarvis in iron man or something you walk up to the desktop and it starts explaining to you hey here's what you missed you know thor's in the other side of the galaxy solving this problem hulk's over here causing these problems what do you want to do iron man and that is invaluable that that doesn't exist as a current product and uh i don't know why slack doesn't do that right now slack's ai is just that is is kind of mia uh it would be incredible to open up my slack and just have it tell me here's what's going on here's the changes that you missed and so it's that's a that's a great vision um i think for time savings what do you think is going to happen because you're running a good yeah no respond well i think the thing with slack is that the if you think about the benefit like at the moment you're stringing multiple tools together you're saying okay let me take let me do a zoom meeting they get a zoom transcript and throw it into a you know notion or confluence and get confluence to summarize it and then i'll organize that in some way where i get to see the summaries in a certain you know why in a certain time and i've got to work out my workflows and for most people that they can do that but it's very complicated to make that happen and same thing with Slack is that, you know, Slack has short form, you know, here's 10 words, here's 50 words, like it's not paragraphs that need to get summarized.

46:13And for you to really understand a Slack conversation, you need to understand the context. Who is this person? What is their job? What have they been doing? What are they, what are they written in Confluence like, you know, today? What have they been doing in their coding? And I think you actually need to understand more data points across that ecosystem. And because of where we are at Elasian, because we can see the code and the specs that they write and the jobs that they had to get done in Jira. We have that data. We can provide all that. So eventually, we can summarize it in a way that's much more turnkey than you having to do it a bit at a time.

46:45And so that's why I think someone like Slack might be able to do it. But if you had to pick out of the two, you'd say Microsoft Teams probably has a better starting point because they see more of the workflow that a knowledge worker does. So that's why I think, again, that data gravity makes a big difference. and we have a huge advantage there. And we didn't even talk about it suggesting what your next move is. So you come back to work and it's like, hey, here's the three things that happened. By the way, you know, you have these three, this topic of this customer churning or threatening to churn.

47:17Here's ways in which we've saved customers before. Whatever, you know, issue that you've suddenly been faced with, the AI could start giving you ideas of how to address that incoming issue. How efficient have you gotten internally? I know you guys did a small riff, maybe 5 % of the company or something during the 2022 period, I guess, in the down market. I think that's when it occurred. A lot of people did that. People did bigger ones, obviously, 10%, 20%. And they didn't see, they saw things get more efficient. Obviously, if you cut the bottom couple of people, it's going to do that. It's just a performance metric there on any team.

47:52but today with ai in the enterprise do you see yourself having to add a ton of people or just making the decision hey how do we point ai at this problem what what is your default as the leader of the company or the co-leader of the company yeah let me um i'm bringing another trend here that's related because we're putting the two together we're the largest company that's committed to remote work in the world we have about 11 000 employees and no one is required to come to an office any day. You can work from home. You can work from the office. You can work from a cafe. You can work from a trailer, you know, traveling around the United States if you wanted to.

48:31And, you know, for a lot of times, you know, people's objections to that is that, well, how am I going to learn from that person sitting next to me at the desk? And, you know, you talked about the sales call example that like, hey, we know how to save a customer or not. And, you know, one way to learn in a sales call is to, yeah, listen, you know, you know, them sit next to the person at a desk when they do the sales call but that's a very much a sort of whack-a-mole just happened to be you know at the right place at the right time to hear someone slightly better if i look at using a company like gong which effectively records sales calls and then you know does the transcription and then looks at competitive analysis and you can tag hey like this was the best saving of a sale customer from this particular competitor or this is the best pitch in this particular scenario or this vertical i'm still into the healthcare space like here's the you know the best person that pitched the healthcare space i think we can end up with a world where you know training and learning from other people is actually mediated by computers in ways that are way way way more effective than they ever could be by just happening to sit next to the next to the person and so we as a company have committed to remote work to build out a lot of those capabilities and we're kind of the canary in the coal mine about how we're building out those experiences, whether it's whiteboards in our Confluence product, you know, and how do you make a digital whiteboard experience better than kind of running into people, or it's, you know, the summarization of data, like, so that you can catch up on people and what's happening that's not, you know, bumping into the water cooler.

50:03So we sort of put remote and AI together because we think that the combination of those two is really changing how knowledge workers interact. And so, you know, to back to your original question around, or how are we doing this internally? We've got, you know, AI projects across the entire business and we think, I think sales and customer touch will be heavily disrupted because I think there's a lot of busy work in many sales teams' job in terms of communicating with customers, but even just researching and understanding like what a customer is doing with our products and we can surface that in incredible ways.

50:38And so take an example of a, you know, salesperson wants to upsell someone, you know, to an enterprise version of the product. You know, they can see, well, you know, we have all this data about what our products get used for and how they get used. Of course, you know, not looking at the customer's, you know, private data, but just like, hey, which features get used. But previously, all that stuff would have been too hard to look at on a feature-by-feature basis. But we can, using AI, summarize that down for people and say, well, actually, we think the enterprise features that will be most appealing are X and Y and Z.

51:11And you can do it in real time, right? You could be doing that in real time. That's data that some team would work on, a team of data scientists for three or four weeks for some off-site meeting every other year. People would say, this is amazing. Then it would quickly be outdated and it would be like institutional knowledge that, you know, just doesn't exist anymore. Now it's going to be real time, right? So these assistants are going to be telling you in real time while you're on the phone call, yeah, you know, this type of customer has gotten these features that work best. This customer is using two out of the three.

51:41So maybe this third one, they don't even know that maybe they need training on that one. Maybe they don't even know that product exists. And we need to sell that into them. Do you worry about I mean, it's kind of a, you know, a very important question in some ways and a silly one in others. And I'm wondering where you sit on the sort of spectrum. Do you worry about this technology, which is moving faster than I think you would agree anything we've seen in our lifetime, maybe the spread of broadband, the spread of smartphones were also very fast. But this is faster, I think. this is going to have a tremendous displacement effect on certain jobs so do you worry about that or do you think yeah you know it's it's overblown it's a couple things here one is that you know the initial technologies i did a lot of research around how electricity uh ran through and kind of changed the world in the 1800s and it's interesting if you ever gone to new york you see go to soho you see these multi-floor uh you know warehouses basically that have often been turned into loft apartments.

52:41But you walk down and you think, okay, this is the manufacturing district and it looks nothing like how we do manufacturing these days. And it turns out that the way it worked was back then you had a steam engine that was in the middle of these factories, often on the second or third floor, and effectively the steam engine would drive belts and pulleys to effectively have all these machines. And you would bring your products up to the floors because it was really almost a sphere because you want to be as close to that steam engine as possible because the belts and pulleys would you know effectively lose momentum and you know slack and that you know over time so you had to be as close to the center as possible and when electricity came in what they did is you know those steam engines used to explode and kill people and other things the the factories would replace that steam engine with an electric engine in the middle of the same factory in the same floor with the same darts and pulleys and now it was great people didn't die but the jobs didn't really change you know there's less someone throwing coal into a steam engine but apart from that if you were a worker on the floor you didn't notice the difference and so that was phase one and that was basically you know sustaining what you currently did better and then phase two came along where it's like well hang on now we have this thing called electricity we don't need a central like uh one central boiler effectively powering everything we can have multiple smaller tools and that's when you sort of saw the henry ford production line it's like well actually let's move the product between the different tools as opposed to you know the reverse and that's sort of the second stage is where you start retooling how how things get made and eventually then the third phase is of course you know electricity being embedded in all the products like and uh i guess you've probably seen that now with electric cars but you know sort of you had various stages of that you know along the way and uh so that's what i think about ai is you go okay phase one is going to be existing tools existing products existing processes slowly augmented we've still got you know the belts and pulleys to the same engine but like it's a better way of doing things and that's what we're seeing right now and then over time you say okay well actually have to reimagine what that looks like.

54:53And that reimagining part, that does change the jobs that are available. But all the experience we've had in the past is that the jobs that are available exceed the new ones, exceed the previous ones. But there is a period of turmoil in between. That period of turmoil can last a decade as things get jumbled up. When I look at the industries that be affected, I say, well, what's demand constrained and what's supply constrained and in software which what we sell to you know if we could have more software developers out there we would like they would get sucked up in the market immediately i feel that's a supply constrained environment i think the open questions is is in sales is sales supply constrained or demand constrained if my my sales people were twice as effective would i hire more of them would i have less of them and i think that is open in certain industries as to like when we can make it more effective do we need more of them or less of them things like accountants and back office things are probably clearly in the twice as effective i just need half as many other areas of the business it's not as clear you know it's so funny i think as technologists of a certain age we're constantly trying to now that we have our you we both have like three decades of this and we watch the dot-com bus we watch the great recession and we've watched multiple paradigms shift in our own lifetimes from mainframe computing mini computing desktop client server cloud we've watched this so many times mobile that we even go back further and look for additional context the one i use you did electricity i did and i'll pull it up here because i think it's hilarious while we were talking i pulled up my chat gpt and i was like hey chat pt tell me about the history of ice shipping to homes before electricity because i had realized like this was the big craze and when you were talking about you know uh soho where i grew up in new york and um i was obsessed with that area in the warehouse i wound up living in a warehouse building on the west side of manhattan on 26th and the west side highway 13 14 foot ceilings and it was made to have you know like steam engines and all kinds of stuff in it there was a 95 year period which here's the description of it from 1805 to basically 1900 where there was a massive amount of investment and entrepreneurship for a century around harvesting ice and bringing it to india the caribbean all over the place and coming up with new ways to systematically harvest it and ship it and maintain it and there were a ton of people who would just drive around with horse and buggy in new york to bring ice to your refrigerator you got it dropped off every day and then boom overnight electricity and then boom the refrigeration unit and this is what but we didn't we didn't have a permanent unemployed class after this happened nor did we have a permanent unemployed class after um you know phones went away and phone operators connecting calls went away we found new uses for human ingenuity that's what will happen here um but the thing that i do find very interesting is every time i'm doing a new job rec now i look at the job and i'm like what are they going to do every day what are the actual tasks what are the goals what could we automate and i'm finding about 20 to 30 percent of each job could be automated away so and watching a portfolio of hundreds of companies i'm seeing the same thing happen you know three or four person startups don't add the fourth or fifth position they just you know they had the fourth but they don't have the fifth and sixth so the capital efficiency of these you know the the most um uh what's the word you know dexterous the most scrappy companies.

58:31They're looking to AI first and then solving their problem with AI and then going on. I have a friend, Brad Gerson, he wanted to make a video for a new project he's working on. He was working into script and then he took it and then he put it into, you know, Claude to make the script. And then he made this like generative AI video that looks like something, I guess he used 11 labs or something to make a marketing video. Now, this is something that a marketing agency would have spent, I don't know, low tens of thousands or an individual contributor might have spent three or four thousand dollars building but when i talked to him about it i had encouraged him to just do it himself write the script himself and then he took it to the next step the the big learning is well if you take some people out of the process it becomes more efficient to your point about when uh hip chat went from 10 to 25 it's actually more efficient if you can do the tools yourself all these creative things you know you you weren't allowed to as a business executive explore your creative chops you had to go to a video editor a designer a logo person and now you got the founders of companies i watch them i'm like oh it's a beautiful logo how did you make it they're like ai i'm like you made your logo with ai they're like yeah and then i remember five or ten years ago maybe ten years ago people were using like you know they would use fiverr or something they'd make their logo for 500 bucks 10 years before that when you were you know doing atlasson in the you know 2001 to 2010 period what did a logo cost you'd pay a design firm three five ten grand to do a logo study What did you pay for your first logo?

59:59Do you remember? I think we actually, if you look at the history of our logos, I think it's pretty clear that we made them ourselves, but that would have been made a lot better, I think, if we'd had all the AI back then. And I think it's the creative class. I think that when people ask me, what advice do I give to kids about where to spend your efforts and what should you learn? And clearly, the accumulation of knowledge is not something that is going to win anymore. Like the person that knows the most in the room, you know, is going to be beaten by Wikipedia every day of the week. But, of course, the accumulation of knowledge builds you the skill set of, you know, learning and like learning quickly and learning how to, you know, think differently.

1:00:39And so, I do think there's a benefit of still acquiring knowledge, not for knowledge's sake, but for getting great at, you know, at learning. And so, then it comes down to, well, it's can you bring multiple different disciplines together, which is sort of creativity. Like, can you create some new idea or something new to the world? And that's the interesting part. And so, I do think there should be a renaissance in, you know, the way we teach kids to be less about learning and more about what are the ideas that haven't been thought up yet. Again, I don't know if there's any traditional schools in the US.

1:01:10There aren't as many in Australia that teach that way. Yeah. We definitely need to rethink education because if you can just sit there with one of these guides, like you said, you know, pull up the Wikipedia page, just kind of create equalizer, wrote learning, you really need creativity, drive, grit, resiliency, teamwork, leadership, you know, all of those skills, which, you know, it's very hard to teach those things become, I think, really defining. listen you give me a bunch of time i just want to ask you one or two more questions which is how do you stay motivated in decade three of your startup a lot a lot of folks you get jeff bezos is on a boat somewhere he's like done i think he's going to come back i think he's going to get bored uh you saw bob eiger retired he told he uh he told uh cmbc he's like you know i was on a boat for like six months i lost my mind i had to come back you think about retiring you think about other projects you think about philanthropy i know you've done some great philanthropy work investing you you backed a lot of the like i think you lp'd a lot of the australian funds from what i understand did a bunch of angel investing and stuff like that you ever think of life beyond atlassian and how do you stay motivated in the third decade of the same company yeah i do want to plug the philanthropy side of things for answer your question please yeah open it up i uh i started about a decade ago actually almost two decades ago we committed ourselves to the sort of one one one model which salesforce did as well which is we give Atlassian 1 % of our equity, 1 % of our profit, 1 % of our product, 1 % of our employee time away.

1:02:37And we did that for a long time. And it's been great for us. Employees love it. They join for it. It's been a huge boon for us. And about a decade ago, I looked around and realized we were still one of the few companies doing this. And so I started a foundation called Pledge 1 % and the aim was to convince every company to do that same model. And in that decade since, we've got about 17 ,000 companies now have pledged to give 1 % of product employee time, product or profit away. And so we're really starting this corporate philanthropy movement. So I just encourage many of your listeners are startups.

1:03:16I just encourage you to check that out because no matter what stage you're at, it could just be the first day of your business. it could be you've been running for 10 years um you know pledge is a very easy thing to do because it doesn't take time until you've actually got something to give back and so uh quite proud about that um on your original question around well it's a big number for atlassian right like one percent of atlassian is a lot of equity and one percent of 10 000 people's time or so at 20 that's 20 hours each that's a lot of hours you're talking about 200 000 hours We've given away about 200 and something thousand hours.

1:03:52We've given away 100 and something thousand licenses, either free or discounted to communities or non-profits. And, you know, we've given away, I don't know, $50 or$100 million to charities over that time. Obviously, it's still a corpus of money to give away. So it'd be huge for us, but I think it's eclipsed by what Pledge 1 % can do if you... Amazing. There's many companies out there from, you know, Twilio to PagerDuty that have, you know, taken the pledge. So if you're a startup out there, check it out, pledge1%.org. Go to pledge1%. On the next side of things, what's next? I still get inspired by a mission, which is to unleash the potential of every team.

1:04:31And the sort of Archimedes said, give me a lever long enough and a fulcrum on which to move it and I'll move the world. And, you know, the idea is leverage. How do you, like, improve things with a lot of leverage? and if you look at our products they get used by everyone from spacex to the american red cross to the australian antarctic division and if we can make every team you know 50 100 more effective if we you know that is going to change the world more than anything else i could do individually like i could go you know start a space company or start you know any sort of like you know huge entrepreneurial venture out there that's made a difference but the compound effect of helping 250 50 ,000 companies and tens of millions of people to be more productive and enjoy their jobs more and get more done for the world.

1:05:19That's way bigger impact that I could have in almost any other domain. And so that keeps me excited at the mission level. And then as an entrepreneur, you know, we're 11 ,000 people. Two decades ago, we were two people. So every single year, there's a different challenge and a different journey to go on. And that is hugely intellectually stimulating uh for me i love i love learning i love trying new things i love uh i don't like failing as much but like that's part of the process uh and uh yeah failing every day but it's really interesting to think about what it's like to manage 11 000 people i mean you meet you must be must meet people you're walking around on the weekend with your kids you must meet people who work for you you don't you never met it's gotta be surreal you do yeah when people wearing a Lassian shirt, you know, is around the streets or, you know, getting a selfie with your staff is a new one for me relatively recently in the last couple of years.

1:06:14And again, it's just a realization that, you know, across 11 ,000 people, you can't have a one-to-one relationship with all of them, but you can inspire them. In fact, one of my maxims is that the most important thing about leadership is to set a vision. People work for a complete asshole who has a great vision. they won't work for the nicest person in the world that has no direction and so my number one thing for leadership is vision and so i'm hoping across 11 000 people we can still set a pretty compelling vision even if i can't have that one-to-one relationship yeah fantastic listen scott thanks for taking so much time for the this week in startups audience amazing journey keep at it and i'll be back in australia i think next year we're going to do the launch festival again so hopefully i'll see you down there in either city or melbourne one of those cities that would be great yeah congrats on everything and we'll see you next time on this week in startups bye bye