Season 2, Episode 9: GitHub

[0] This is what I, I'm definitely my brain is not working right.

[1] Um, it has been a long day.

[2] Welcome to season two, episode nine of Acquired, the podcast about technology acquisitions and IPOs.

[3] I'm Ben Gilbert.

[4] I'm David Rosenthal.

[5] And we are your hosts.

[6] Today, we are coming at you on Tuesday, June 5th, the day following the big announcement that Microsoft is buying GitHub.

[7] Ooh, I thought you were going to say the day after Monday, June 4th.

[8] It is also after Monday, June 4th.

[9] And, you know, this show is clearly patterning after acquired.

[10] It is a San Francisco company and a Seattle company that are joining forces.

[11] And I don't know where I'm going with that.

[12] But we sit here in Wave Capital in San Francisco near the headquarters of GitHub.

[13] to celebrate this momentum occasion.

[14] Indeed, indeed.

[15] And this is our second in -person episode in a row.

[16] We actually recorded what we thought was episode nine last week together in Seattle.

[17] We did.

[18] And it turns out now that that's episode 10.

[19] So listeners, you'll have to wait and find out what that could be.

[20] What Seattle company might we have covered?

[21] Tune in next time to find out.

[22] All right, getting into this show, if you're new to, the show.

[23] You can check out our Slack at Acquired .fm.

[24] You can also check us out on Apple podcasts, and we appreciate any love or reviews that you could show there.

[25] Love coming in the form of reviews.

[26] Okay, listeners, now is a great time to thank one of our big partners here at Acquired, ServiceNow.

[27] Yes, Service Now is the AI platform for business transformation, helping automate processes, improve service delivery, and increase efficiency.

[28] 85 % of the Fortune 500 runs on them, and they have quickly joined the Microsoft's at the NVIDIA's as one of the most important enterprise technology vendors in the world.

[29] And just like them, ServiceNow has AI baked in everywhere in their platform.

[30] They're also a major partner of both Microsoft and NVIDIA.

[31] I was at NVIDIA's GTC earlier this year and Jensen brought up ServiceNow and their partnership many times throughout the keynote.

[32] So why is Service Now so important to both NVIDIA and Microsoft companies we've explored deeply in the last year on the show?

[33] Well, AI in the real world is only as good as the bedrock platform it's built into.

[34] So whether you're looking for AI to supercharge developers and IT, empower and streamline customer service, or enable HR to deliver better employee experiences, service now is the platform that can make it possible.

[35] Interestingly, employees can not only get answers to their questions, but they're offered actions that they can take immediately.

[36] For example, smarter self -service for changing 401K contributions, directly through AI -powered chat, or developers building apps faster with AI -powered code generation, or service agents that can use AI to notify you of a product that needs replacement before people even chat with you.

[37] With ServiceNow's platform, your business can put AI to work today.

[38] It's pretty incredible that ServiceNow built AI directly into their platform, so all the integration work to prepare for it that otherwise would have taken you years is already done.

[39] So if you want to learn more about the ServiceNow platform and how it can turbocharge the time to deploy AI for your business, go over to servicenow .com slash acquired, and when you get in touch, just tell them Ben and David sent you.

[40] Thanks, ServiceNow.

[41] All right.

[42] David, are we ready?

[43] Let's get into it.

[44] So we didn't have a lot of time to research this one, given that it's hot off the press.

[45] Yep.

[46] But fortunately, I actually had a little bit of a shortcut, which was back when I was.

[47] was a first year, a young whippersnapper, first year in business school in one of my classes, I believe it was like high -tech business strategy or something like that.

[48] Sounds official.

[49] It sounds very official.

[50] Me and myself and two of my classmates did one of our, did a final project for a class on GitHub.

[51] So I just so happened to have a 17 -page document that we had all put together all about the history and strategy of GitHub.

[52] If only every episode's prep was that easy.

[53] Unfortunately, it's from back in 2013, but as we will see, not that much has changed with GitHub since 2013.

[54] Okay.

[55] So with that fun, oh, and other fun story, one of my professors at the time, not actually in this class, but one of my professors at GSB was Peter Levine, the general partner at Andreessen Horowitz, who led GitHub's Series A. And it was a. It was a. on the board.

[56] Unprecedented freaking $100 million first capital in Series A. That's a small series A these days.

[57] At the time, it was unprecedented.

[58] Also, not really a series A, as we will get into.

[59] But to start, before we get into the history of GitHub, the company, there are a couple things that we want to cover that are pretty important to understanding the context here.

[60] And Ben and I were laughing before the show, probably half our audience, the more technically minded folks in our audience, engineers and the product managers are probably going to know all this already.

[61] Probably we'll know it a greater detail than us and we'll write in and tell us what we got wrong.

[62] Yes, as usual with a technical acquisition, for those that are more technical than us, we apologize for all the things we will surely get wrong.

[63] For those that are less technical, bear with us.

[64] And I hope that we can make something that's a little in the weeds and a little in the details, digestible, and something you listen to by choice.

[65] Yeah.

[66] But whether it's fun or not, this is really important to understand and why Microsoft paid $7 .5 billion for this company.

[67] Okay, so basically there are kind of two concepts.

[68] Two, not concepts, but areas that we want to talk about as background.

[69] One is the whole idea of how managing software projects work.

[70] And that's this area of software itself called software configuration management or SCM for as for as initials or as an acronym.

[71] The other area that we want to make sure we talk about is kind of the evolution and dynamics of the open source software ecosystem.

[72] And we will get into both of those.

[73] Okay.

[74] So first, software configuration management.

[75] when you're building and managing a software project, a software project could be a small thing that you're building yourself or with a few friends, could be an open source project like Linux, or it could be your company's code base for a private company.

[76] It's really hard once more than one person starts working on the project to kind of manage all the changes and contributions and different pieces of development that all the different engineers on the project are making.

[77] And I would say even just as one person, as a person that used to write PHP scripts and then name them V2 and switched and dot backup, it is excellent to be able to use version control and all elements of software configuration management, even for your own private projects.

[78] Even for all our investment banker and former investment banker friends, you've certainly lived through merger model V25, final, final.

[79] underscore final underscore do not delete underscore v you know DR this is basically making all of that easy and seamless for way more people than ever work on an Excel model in cases so there's basically kind of three main ways to do this in the very very old school ways it was really easy everybody just worked on the same machine so there was one file on one machine and that was the code base.

[80] And if you were an engineer working on the code base, you went on that machine or you telexed into that machine and worked on it.

[81] Seems simple.

[82] Why change?

[83] Well, the computing world evolved a little bit.

[84] And what was developed was the client server model.

[85] And this was known as the centralized model.

[86] And so in this case, there would be what was called a shared repository.

[87] And that's where all the code lived and all developers, anyone who's working on the code could work on it.

[88] but there was one canonical central repository.

[89] And so, David, you know, let's say I'm just doing a little weird thing on the side here.

[90] I don't really want anyone else to know about it or be exposed to it or have to pull down all the crazy code that I'm doing for my side thing.

[91] You're saying that's not possible in this client server centralized model.

[92] No, Ben.

[93] Why would you want to do that?

[94] Hmm.

[95] Feels like I'm going to spend a lot of time syncing my code base with the centralized server.

[96] or like something existed like an open source project and you wanted to clone it and use some of it and make your own version of it.

[97] Would that be easy to do, Ben?

[98] Well, it doesn't strike me that once I clone that, it would be easy for me to make a few changes and maybe over the course of several months and upload that back up.

[99] That feels no. No. That would be tough.

[100] So as you can see, this kind of all worked well in the pre -internet days but once the internet comes around and distributed software development becomes more of a thing, especially in the open source community, this becomes a lot harder.

[101] So fortunately, in the late 90s, just as the internet and distributed software development, and in particular the open source movement is becoming a thing, a new model emerges for managing the code base for a project, and that is the distributed model.

[102] And that was pioneered by a bunch of folks, but there was a company called BitKeeper, a company in a product called BitKeeper, developed in the late 90s.

[103] And it was actually used for the development of Linux, for managing all the development, ongoing development and maintenance of the Linux software kernel.

[104] Even though it itself was paid software, it was used in this open source, massive open source project for Linux.

[105] So BitKeeper was the first decentralized, was it the precursor to get?

[106] Did they develop Git?

[107] No. So, Big Keeper, I don't know if it was the first, but it was one of the first or early, you know, sort of stable products that could be used to do this.

[108] And it was a private for -profit company.

[109] And their software was used by lots of other companies.

[110] But because Open Source was sort of this crazy nonprofit thing that was emerging and a new way to develop software, pioneered, of course, by Linux and Linus Torvalds, and they, Big Keeper actually ended up licensing a free, version of their product to, I don't know if it was to the Linux Foundation or to the Apache Foundation or just what, but made it available for open source developers to use that.

[111] And that became a way to like massively scale the number of developers all around the world that can work on these projects.

[112] So that continues for a couple of years and all sort of well and good.

[113] But then in 2005, BitKeeper, the company actually pulls the free license that they had been providing to the open source community.

[114] They, yeah, exactly, you get what you pay for.

[115] And, and so this creates a huge kerfuffle in the open source community.

[116] But of course, as we will get into, people have lots of opinions and lots of different opinions.

[117] And this leads to two separate projects that are pioneered and undertaken really in an amazingly short period of time to.

[118] replace Bitkeeper.

[119] And those two projects are, first of as Mercurial, which I believe has less market share now today, but it's still around, is actually used by Facebook.

[120] So all of Facebook's codebase runs on the Mercurial source control system.

[121] And the second, which leads directly to where we are today, is a product called Git.

[122] Now, Git is interesting because it is actually developed and implemented.

[123] in like a week, apparently, by the open source man himself, Linus Torvalds, creator of Linux.

[124] Well, if somebody was going to do it in a week, it would be him.

[125] It would be him.

[126] And then quickly, he, I don't know that he loses interest, but he hands off the project to other folks to keep them implementing and developing.

[127] So Git naturally, because it was developed by Linus, becomes first the source control management system that Linux uses, the Linux kernel uses, then it starts quickly gaining adoption in the rest of the open source community.

[128] Hadoop moves over to it, the Hadoop project, many others.

[129] I think Mercurial still lives on, and like we said, Facebook uses it, but GIP becomes very popular in a very short period of time.

[130] And so is there any sort of graphical user interface to access it at this time?

[131] Is there any sort of, you know, not centralized, but pseudo -centralized place like are people using source forge with git loaded onto it yes well so git and all this stuff as we said in the beginning is just software uh in and itself is software that manages the version it doesn't actually host the software the software uh the code which itself is just files has to actually live somewhere uh still on a server or on a distributed set of servers um and so ben just like you said, there's still no way to really do this in a nice graphical user interface way.

[132] There are what's called forges, which are essentially servers that host code out there, and they're used in the open source community.

[133] They really should have kept that name.

[134] They really should have.

[135] They really should get forge.

[136] And so these are companies like SourceForge.

[137] I think a lot of people will know.

[138] Google actually launches its own version of this called Google Code.

[139] Yep.

[140] And they're the most popular places for the code to actually live.

[141] But the problem is there's not much more to them beyond a place to live and like a very standard Wikipedia -style webpage about the projects.

[142] There's no concept of being able to interact with other people who are working on the projects.

[143] It's just hosting the code and then you're using Git or Mercurial to manage all of the changes made to the code.

[144] So right around the same time, we're now into like 2006, 2007, 2008, we're in the heyday of on the completely other side of the tech world from all this infrastructure, you know, developer, technical, friendly, technically heavy work, you have the consumer web popping up.

[145] You have Facebook.

[146] You have Twitter.

[147] You have LinkedIn.

[148] And this whole idea of social networks and the combination of what is possible when you take software and people and communities and you all marry it.

[149] people start thinking about, well, what if there's more of a community type site instead of these forges where developers can interact?

[150] Yeah, and that really takes the concept of, you know, people arguing on mailing lists and completely decoupled forums that were, you know, totally different than where the code itself was, um, was maintained and, and move it into one place together.

[151] And one of the big, you know, the trope of the time, and it's still a trope a little bit, but, like, if you were contributing to an open source project, there was likely a flame war.

[152] Like, there was, it was political.

[153] People were, you know, grabbing on for dear life to have control over their little corner of the fiefdom, and then the mailing lists would have flamers that went out.

[154] And of course, like, it's not purely because the software made it all better.

[155] The industry has matured quite a bit.

[156] But it is, you know, we've moved on from that time, at least with a lot of the major open source projects in a lot of ways.

[157] yep um so people are thinking like maybe there's some sort of community where all this can come together off of mailing list maybe some sort of hub you might think uh and that gets into the founding of GitHub so i'm going to i'm going to switch over here now to the canonical founding history in the words of co -founder tom preston warner himself from his own blog uh quote and tom tom was a developer working for a company called PowerSet, acquired by Microsoft also, in San Francisco at the time.

[158] Two -time Microsoft Acquiree.

[159] Two -time Microsoft Acquiree.

[160] In his own words, I was sitting alone in a booth at Zeke's Sports Bar and Grill on 3rd Street in San Francisco.

[161] This is Ensoa.

[162] I wouldn't normally hang out at a sports bar, let alone.

[163] Love it.

[164] This is just like this episode is going to be filled with with a...

[165] Just instantly throwing shade on anybody who's like not a proper nerd.

[166] So many tropes, let alone a sports bar in Soma.

[167] But back then Thursday was I can has Ruby Night.

[168] Wait, like I can has cheeseburger?

[169] Like I can has cheeseburger.

[170] No way.

[171] Ruby, of course, referring to Ruby on Rails, the web development language that was starting to take off and was the development language behind a lot of these social, you know, 2 .0 websites.

[172] I think Twitter was built in Ruby, right?

[173] Yeah, Twitter was originally a Rails app that, like, massively, massively didn't scale.

[174] And that was why it says the front end would always fall over.

[175] And I think Alex Payne was brought in, a total side tangent of it, AL3X on Twitter, Alex Payne was brought in to kind of redo that whole thing in Scala and decouple the front end from the back end.

[176] Yeah, totally.

[177] Oh, what a mess.

[178] Facebook originally, of course, being built, not even in Ruby, but PHP.

[179] And actually, I think GitHub is still largely Ruby.

[180] GitHub itself.

[181] Oh, interesting.

[182] Yeah, that would make sense.

[183] That would make sense.

[184] So it's ICAN has Ruby Knight, of course, has been alluded to a reference to the ICAN has cheeseburger memes.

[185] Law cats.

[186] From the Seattle company, Ikenhas Cheeseburger at the time.

[187] And this is a developer meetup for Ruby on Rails developers in San Francisco.

[188] So Tom is at the bar, and he's taking a little break from the meetup.

[189] He goes, he orders a beer, and he sits down in a booth.

[190] And apparently right at that time, his not even friend, acquaintance, Chris Wonstrath, walks in.

[191] And Chris is also a Bay Area Ruby developer at the time.

[192] He has his own startup that he's working on.

[193] And Tom, for some reason, even as he's writing this, he doesn't know why he doesn't know why does it.

[194] He just kind of gestures Chris over to the booth and he's like, hey, man, check this out.

[195] I've got an idea I'm working on.

[196] And the moral here is always be looking for people gesturing to you that could have startup ideas.

[197] In sports bars and so much.

[198] And so Tom says, you know, about a week earlier, he had started on this project that he was calling grit at the time to basically access these distributed Git repositories and do so via.

[199] And do so via via a bunch of code that he'd written in Ruby.

[200] And he has this idea that he can take this code that he's written for accessing these repositories and turn it into a website that can act as this kind of idea of social coding that people have been talking about.

[201] And apparently, Chris, they barely even know each other.

[202] He's like, I'm in, let's do this.

[203] And this is like the...

[204] Literally, hold my beer.

[205] This is like literally the dream of engineers.

[206] with like this thing happens 10 ,000 times more often than it works.

[207] Yes.

[208] Like software developers can make anything and what ideas people come up with are generally problems they're experiencing themselves.

[209] And so it is like the total dream of I have an idea that is a tool or a platform for other software developers or to fix this problem in software development.

[210] And like these businesses can be very hard.

[211] There's very few that have made it to be 500 million billion, billion dollar.

[212] plus companies.

[213] And this is like, you know, this is the realization of the, the total home run right here.

[214] And, and of course, there's no better, you know, customer feedback for your, as a developer, for your idea for a developer tool than other developers who you respect.

[215] There you go.

[216] There you go.

[217] So it was all, it was all magically ordained.

[218] So they start working on it on nights and weekends.

[219] They both have full -time jobs.

[220] And Tom is working for PowerSet, which was a search technology company.

[221] and Chris has his own startup with his co -founder, PJ Hyatt, that they're working on.

[222] They get an MVP kind of hacked together, and Chris and PJ start to decide that they're going to start using this interface on top of Git within their own startup.

[223] And pretty quickly it works great.

[224] And everybody's working there is like, yeah, this is awesome.

[225] You know, the whole be your own first customer thing.

[226] Exactly, exactly.

[227] So they work on it for a few more months.

[228] January 2008, a couple more months go by, they launch it and they open up, they tell all their friends about, hey, now there's this way.

[229] If you're using Git, that we have this cool web interface on top of it and you can basically interact with the project and understand, see a wiki of everything is going on, talk to other people who are working on the project, manage it all from there.

[230] It starts spreading like wildfire.

[231] It's one of these things, it's a lot like sort of Uber when it launched or Twitter when it launched, where as soon as you turn the lights on, there's perfect product market fit.

[232] Perfect.

[233] It was so, uh, the product market fit was so perfect that, and so many people started using it that they actually started becoming really concerned about their own hosting costs because their GitHub is now the forge, the hosting, uh, service for all this code and all these people are now using it for their companies, for open source projects, for everything.

[234] And they're like, oh, crap.

[235] I understand why it's great.

[236] It always seemed ironic to me that we had this big breakthrough in.

[237] decentralized version control and then GitHub centralized it again.

[238] Yeah, totally.

[239] I mean, like, it's not fully centralized again.

[240] So there's lots of like, you know, reasons why it's good, but it's still, it's like, this is the like engineering counterpart to the business axiom that there are two ways to, you know, make money of bundling and unbundling.

[241] It's like centralizing and decentralizing.

[242] Yeah.

[243] It is funny.

[244] I thought about that a lot with, um, um, so cloud computing is this, uh, client server model where if you look, back in the 80s, if you look back at like dummy terminals and mainframes, like nothing lived locally, everything lived in the mainframe.

[245] Then CPUs got really powerful and like we got good at storing stuff and hard drives and they got really big.

[246] And then you could buy cheap computers.

[247] So we didn't do anything remotely anymore.

[248] So everybody's got there.

[249] Everybody's processing locally, everybody's storing locally.

[250] And now like we're in this place where if your computer crashes, you can likely just go get another, you know, MacBook Pro, bring it home and within an hour be back up to speed because everything's in some cloud service somewhere.

[251] And it's funny how there's this wax and wane of centralized decentralized, you know, client server back to, you know, powerful PC back to client server.

[252] Like the client server.

[253] It's, it's the never -ending TikTok of the technology cycles.

[254] It is.

[255] So, they're worried about hosting all of these centralized, forges, or this forge that they have now.

[256] and they're like, oh, crap, we got to build a business model into this ASAP, because they had launched it in anybody who had a Git -based project could use it.

[257] Could be open source, could be a private company.

[258] They're like, what could we do?

[259] Well, what if we just keep it free for open source?

[260] Because remember when BitKeeper, which had been this private company, was doing distributed version control for open source, and then they pulled the license and had to make, try to make open source projects pay for BitKeeper, everybody defected, and then that led to Git and Mercurial.

[261] So obviously they can't do that.

[262] But what if we charge companies who are using this for their own private code bases?

[263] We'll just charge them and we'll make it free for open source.

[264] It's really brilliant.

[265] Like it is the thing that's keeping us growing, the ability for public repos that have no permission set on them, that people are doing open source.

[266] projects being free to access and thus continue to fuel growth.

[267] However, monetizing people that want to do things that are counter to growth by locking it down, it's like the perfect place to draw the freemium line.

[268] Totally.

[269] I think that from a lot of the companies that we've examined on this show, I think it's really helped me inform what makes sense in a freemium model versus what makes sense in a free trial model.

[270] And in the freemium model, it's really all about, I mean, a massive component to, are you going to build a successful business is, are you able to draw the line in the place such that you're both fueling growth and monetizing exactly in that place when people are finding it so valuable that they'd be willing to pay?

[271] And when people want to do private repos, like that's exactly that place.

[272] It's exactly that place.

[273] And not only that, they've just completely nailed the timing back to our hopefully not too dense and not too poor description of the open source community early yeah more more more exactly i agree please don't judge us harshly it's been a long day the uh back to the open source community this is right at the time when open source projects and development is just exploding like it was like people used to think open source they used to think Linux and operating systems and value destructive yes and value destructive interestingly uh and And particularly to foreshadow the Microsoft acquisition, like Microsoft was the one carrying the banner on open source is value destructive.

[274] It's a threat to our business model.

[275] It's a threat to selling licenses, selling software licenses.

[276] It's a threat to everybody who's trying to create a software development ecosystem and platform.

[277] And I mean, people had done okay, like Red Hat figured out a business model around open source, but it was unclear if open source could ever actually be a successful business.

[278] Yep.

[279] And you go from that.

[280] Again, it was all centered around operating systems to now you've got programming languages are open source, Ruby's open source, you've got Hadoop, which is open source, you've got data tools, you've got visualization tools, you know, I don't think like React was around at that point, but like now it is like all of these mobile development frameworks.

[281] Now you have 100 ,000 JavaScript's frameworks.

[282] Yeah, exactly.

[283] So you can pick four of them and use them in some strange concert that proliferation of open source tools, needless to say, has its own.

[284] And that had a couple, there were a couple secondary effects from that.

[285] One is that simply there's so many more projects, so many more people working on it.

[286] That just leads to lots of people now using GitHub, period.

[287] But also, it becomes your sort of your resume, your badge of honor.

[288] As a software developer, people want to know not where have you worked or what have you worked on there.

[289] They want to know what open source projects have you worked on?

[290] And can I see your code?

[291] Can I see which contributions you have made that have actually become part of the accepted by the community as core part of the project?

[292] What's your impact?

[293] And that's way more traceable and open in a way, just like LinkedIn and Facebook in the business world and the social world, than it ever used to be.

[294] Yeah.

[295] And of course, people don't always have time to be contributing to open source projects when they have day jobs.

[296] However, plenty of people start little side projects and weekend things where they're just the only one contributing to a repo.

[297] They put it up as a utility for other people to use.

[298] It's public.

[299] It's searchable.

[300] And that can be a resume piece where you show off the quality of your code and what you're, what small problem you were passionate about solving.

[301] And that's a thing that I think is looked for much more by prospective hiring managers than open source contributions for a lot of employees.

[302] Yeah.

[303] Well, certainly both are.

[304] And it depends what you're hiring for.

[305] Yeah.

[306] But totally changed the way that software.

[307] development recruiting was done completely uh none of that visibility existed at all i mean it was like literally it was like linked in uh which you know we'll get to with microsoft and the acquisitions in a little bit but um it was like linked in your resume before then you know it was people didn't want to share their resumes and it was weird like it was this super private thing it was like your bank account number um but after linked in of course you wanted to share your resume and it became the basis by which all recruiting happened uh very similar dynamic in the software world.

[308] Yep.

[309] So they realize, they come up with this brilliant premium model.

[310] They realize they need to start making money to keep the lights on.

[311] They quickly hack on the site.

[312] April 2008, they launch paid private repositories.

[313] So again, if you want to make it open to the public, which by nature is anything open source or Ben, like you're saying, your own side projects that you're just putting out there for anyone to see, your code, it's free.

[314] But the minute you want to make it private, you have to pay GitHub.

[315] you have to pay GitHub a certain amount per month to cover that.

[316] And once again, it works like wildfire.

[317] Shortly thereafter that, we had alluded to PowerSet where Tom was working.

[318] So Chris and PJ quit there, folded their earlier startup, went full time on GitHub as co -founders.

[319] Tom, the third co -founder, who originally had the idea, he's still working at PowerSet.

[320] July 2008, PowerSate gets acquired by Microsoft for a hundred hundred, hundred million dollars and uh microsoft offers tom a three hundred thousand dollar retention bonus if he signs on to stay for i believe three years and um tempting tempting but this premium thing is really kind of taken off of get up so tom makes the correct decision does not go work for microsoft uh leaves power set microsoft and goes full time on github uh and so the three of them are full -time Thomas CEO, but it's completely flat.

[321] They're all equal and working on the company.

[322] And so just like in the Elassian episode, which we also covered and is also very reminiscent of developer tools and tools for teams and collaborations, they decide that they're going to bootstrap and they're making enough money.

[323] They don't need to raise venture capital.

[324] And the system that they come up with is they devise, I wasn't able to find exactly what it is, but an elaborate formula based on how much money they made on the site that month, how much they pay themselves.

[325] And like, they have to hit certain goals and they're basically cliffs and then they increase their salaries by a certain amount.

[326] They still haven't hired anybody.

[327] It's just the three of them.

[328] And that goes on.

[329] They also decide, though, that they, since they have started the company, they need a logo.

[330] Where should they look for a logo uh stock photos this is this might be there are a bunch of good stories in here i think this might be the best story of the episode so they had heard i'm actually it's unclear to me if they're friends with the founders of twitter um they must be it was actually still a much smaller place uh software development and the startup ecosystem in san francisco at that time and and if they were both um sort of active contributors in the ruby community probably probably which is how they to each other.

[331] And so it turns out that the Twitter logo, the bird, not the bird you see today, but the original version of the bird, was actually an I -stock photo by the British graphic designer Simon Oxley, who is just very prolific and produced tons of stuff, tons of stock art on eye stock photo.

[332] The Twitter founders saw that on there and like, that's the bird we want.

[333] and they paid $15 for the original bird, and that became the Twitter logo.

[334] Yeah, but David, that one couldn't be appropriately described by circles.

[335] Did you ever see that?

[336] No. So the quick aside, the newest, and really nice, really simple, really beautiful mark for Twitter is an, you know, aspirational sort of upward -turned bird.

[337] And the day they announced it, made a bunch of hay about it.

[338] They were illustrating how it was sort of perfect in all these ways, and it's because you could sort of use a bunch of circles, use the edges of circles to create the exact shape of the bird.

[339] So the bottom curve of it and all the feathers.

[340] And then it became this big thing that like people would just go and create any logo by using circles because given an infinite amount of circles and circles of any size, you can kind of like make any shape you want.

[341] So it's kind of this like people at first were like, oh my God, it's all made from circles.

[342] And then everyone sort of looked at each other like, wait, wait, wait.

[343] It was also like really disorienting because Because, you know, the current Twitter bird logo faces right, but the original one faces less.

[344] Disorienting.

[345] And I don't think there's a wordmark anymore.

[346] So it's like difficult to write Twitter properly.

[347] Like if you go look at Twitter, it doesn't say Twitter in any logo text anywhere on the site.

[348] This is not the Twitter episode.

[349] Total aside.

[350] We will revisit that in a Twitter episode someday.

[351] But is helpful background, the GitHub founders, they're like, great.

[352] work for Twitter.

[353] We're going to go on Ice Stock photo.

[354] We're going to look at this dude, Simon's work.

[355] He's got a ton of stuff on there.

[356] We'll find something great for us.

[357] Animal, like, birds, sure, what should we find?

[358] What does he have?

[359] They're scrolling through.

[360] Oh, how about an unholy marriage of a cat and an octopus?

[361] Perfect.

[362] Nailed it.

[363] Done.

[364] And I think it was not called octocat yet.

[365] No. It also only had five tentacles.

[366] It was very confusing.

[367] very very confusing um no it was not yet called octoket i i did look up but i forget what it was what it was originally called by octopus yes i think that was right yes i think it was octopus or something like that and um so they bought it uh i don't know if this is correct but i what i read was for three dollars i think that they bought it yeah um they really got a deal versus the 15 dollars that twitter paid yeah it's like just about half the cost of one month of uh registration on their site yeah totally uh you know it was uh it was a bootstrap startup uh so they buy the octopus rename it the octocat and that is how the octocat became the logo and then very quickly official mascot uh adopted worldwide by githbers and github fans um around the world and with that um really on the back of the Octocat.

[368] Growth continues to come very quickly.

[369] So about a year, we're in early 2009 now.

[370] That's a causation correlation.

[371] Absolutely.

[372] I think it was no doubt causation.

[373] Early 2009, they announced that they now have over 50 ,000 public repositories.

[374] So these are projects, including open source and personal projects, that are open to the public that have been created on GitHub.

[375] They never talk about how many.

[376] repositories they have.

[377] July 2009, they announced they have over 100 ,000 registered users.

[378] So 100 ,000 developers, which is a lot like the, there are not that many developers in Americas.

[379] They already have, well, around the world, they're more, but they have pretty high penetration pretty quickly.

[380] They have 90 ,000 unique public repositories.

[381] By the next year, they have a million public repositories, 211, 2 million public repositories.

[382] And when you look at, these numbers are a little hard to comprehend here on the radio, but when you look at charts, I mean, either by repo or by user growth, it is just at any scale that you look at it, just completely geometric.

[383] Or actually, I don't have a geometric, exponential, logarithic, whatever.

[384] It's up as it goes to the right.

[385] It is getting upper as it goes righter.

[386] Yes, exactly.

[387] By the middle of 2011, they have now, GitHub is now surpassed.

[388] three years into the company.

[389] It is now surpassed SourceForge and Google Code as the number one place where projects were open source and any personal public projects are hosted on the internet.

[390] So that was a major milestone for the company.

[391] Then the next year, so they've been bootstrapping the whole time, profitable the whole way, and making money and very nice salaries by, we would assume, by the formula that the three founders devised.

[392] Um, July 2012, Andresen Horowitz, uh, comes in and invests a hundred million dollars in the company, the largest check that they've ever written.

[393] And by all calculations that anybody can even think to do going back as much history as possible, the largest quote unquote series A ever raised in history.

[394] Got to be.

[395] It's got to be.

[396] Now, it wasn't really a series A because like, the company existed.

[397] They had 100 employees.

[398] Uh, it had just been bootstrap for a long time.

[399] Um, but it was their first.

[400] first round of funding.

[401] It was at a $750 million valuation.

[402] Incredibly, incredibly impressive.

[403] Yeah.

[404] And from what, from what David and I understand from some primary source research, a very, very competitive round where Andreessen Horowitz just decided, yeah, we're going to do this.

[405] And nobody is going to, nobody else is going to do this.

[406] It's just going to be us.

[407] And it was a great company, obviously, very hot, a great round to do it.

[408] And obviously, Andresen did great in the investment, but also not super clear that it warranted that much.

[409] I mean, they had grown so much.

[410] As we alluded to, I wish I had in my old term paper from GSB, I had the number of software developers in America.

[411] They were like approaching 100 % penetration.

[412] So it was unclear like how much more GitHub could grow.

[413] Well, David, if your entire fun thesis is that software is eating the world, you could imagine there's probably going to be a lot more software developers.

[414] One would imagine, and one would be correct.

[415] So January 2013, six months after they raise the Series A, they announced that they have 3 million users and 5 million repositories.

[416] And again, you're wondering about how much headroom is there in the market.

[417] By the end of 2013, they have 10 million repositories, so doubling the number of users, or doubling the number of, of projects hosted on the site publicly, not to mention the private ones, and things are going great.

[418] Unfortunately, 2014 was not the best year for the company.

[419] And let me, I actually, so I pulled up your term paper and looked at it.

[420] At the end of 2012, GitHub had almost 2 .08 million members compared with the total professional software developer, the software developer population of the United States of only 1 .3 million.

[421] Right.

[422] So they had over 100 % of the United States software.

[423] developer population.

[424] So presumably it was probably, what, like, you know, half a million from the U .S. and a ton internationally or something.

[425] But crazy.

[426] Crazy.

[427] How much bigger can it get?

[428] So as we were alluding to, 2014 was not the best year for the company.

[429] Tom Preston Warner, who we had talked about who originally had the idea for GitHub and was the first CEO in a sort of bizarre series of events that it's still unclear.

[430] exactly what happened.

[431] He and his wife, who did not work at the company, were accused of harassment at the company, led to an internal investigation, and ultimately in Tom leaving the company.

[432] So since 2014, he is, he's still a founder, but has not been part of the company.

[433] Yeah.

[434] And, and, and, and, you know, David and I don't profess to know anything about what goes on at the company or, and, you know, by all accounts, it's, it's still kind of unclear.

[435] what happened there, but you can kind of read between the lines on, you know, super young founding team, company growing incredibly fast, you know, a lot of very unorthodox sort of management principles.

[436] They adopted the valve management style of a completely flat organization.

[437] Nobody reported to anyone else.

[438] Anyone could work on whatever they wanted.

[439] Yeah.

[440] You could imagine it would be a culture that that wasn't sort of the tightest run ship.

[441] And, you know, that, that, I would say we, the public never really knows the true story of what went down there.

[442] But the internal investigation did reveal, I believe it revealed, according to the internal investigation, not specifically harassment, but lots of signs of lapse in judgment.

[443] Judgment, yeah.

[444] Anyway, as we've said multiple times on the show, we're not super equipped to prosecute one way or another.

[445] and this is a particularly tough to parse case.

[446] But anyway, it results in Tom leaving the company and Chris Wonstruff, the co -founder who walked into the bar, becoming the CEO.

[447] And he would remain the CEO until yesterday.

[448] But we will continue on that.

[449] The company still keeps growing.

[450] In 2015, Google just gives up and Google shuts down Google code, says, just go to GitHub.

[451] It reminds me of Google Video.

[452] except they didn't acquire in this case.

[453] Yes, they did not.

[454] Microsoft did.

[455] Could you imagine if Microsoft had acquired YouTube?

[456] That would be a what would it happen otherwise.

[457] Huh.

[458] Yeah, Bing video wouldn't be flourishing.

[459] So it would not.

[460] It just doesn't have the same ring to it.

[461] We love Microsoft, but we really do.

[462] July 2015, things are back on track, going really well.

[463] the company raises another $250 million led by Sequoia at a $2 billion valuation, and they announced that they have 10 million users.

[464] And it's important to note this is not, we have covered times on this show before where a company is cash flow positive or profitable or sort of the could be profitable at any moment scenario, and they're raising extra cash just to double.

[465] down on growth, or just to have cash reserves in the bank for competition or to go to an IPO or for whatever reason.

[466] This is not the case.

[467] This is, we are not running a profitable business.

[468] We're generating nice revenues and the adoption is crazy and the stickiness is huge, but we are, you know, we need this capital in order to continue to operate the business.

[469] Well, and, you know, I think it's also a fair question to ask.

[470] Certainly the, any market size estimate of the number of software developers in the U .S. or in the world in 2011, 2012 would have massively underestimated the growth.

[471] Yeah.

[472] That said, I have a GitHub account.

[473] I have contributed to do a few projects.

[474] I would in no way consider myself, as is obvious by this episode, a software developer.

[475] So there is very much a question of, like, how much of...

[476] I'm sending you a poll request tonight.

[477] You just wait.

[478] Oh, no, I'm scared.

[479] No, I'm real scared.

[480] So there is some question in there.

[481] But yes, as Ben, as you referred to, you know, over the next few years, again, the company continues to grow very nicely.

[482] But last, end of last summer, August 2017, they make a curious announcement.

[483] They announce that they have passed the $200 million annual revenue revenue.

[484] run rate, which is, you know, impressive.

[485] But relative to being valued two years before at $2 billion, it is lower than you would expect.

[486] And the punch of that announcement comes from the fact that 110 million of that was from their business line.

[487] Basically, hey, we sell to individuals and we sell to businesses, kind of like Dropbox in the episode that we covered there.

[488] And our, um, our, um, our, um, our, you know, business division or enterprise division has sort of outpaced the, uh, the consumer adoption.

[489] So before when we were talking about, hey, I just want to make this private repo, um, versus, okay, my company has adopted GitHub as the way we develop software.

[490] And so I think it was, uh, in part to tout, hey, we actually have a real sort of, uh, enterprise B2B business here.

[491] And that's all ARR.

[492] So, you know, it's, it's not just one off revenue.

[493] Yep, recurring subscription revenue, which is great.

[494] The other part of the announcement, though, is that Chris is going to be stepping down as CEO, and they're going to be running a CEO search in the company to bring in a new, you know, experienced, not that Chris has an experience at this point, but a new more experienced CEO to run and lead the company.

[495] An action jack, if you will.

[496] Indeed.

[497] Indeed.

[498] And then that leads to.

[499] to yesterday, well, really to over the weekend when the rumors started swirling, but then the announcement yesterday, which was that they had found the CEO, Nat Friedman, who had been the founder of Zamoran, a developer platform, mobile developer platform, right?

[500] Cross -platform mobile development that Microsoft had acquired a few years ago.

[501] He's going to be leaving the Zamoran division within Microsoft to run GitHub, and concurrently Microsoft is buying the company.

[502] Oh, yeah, and there's another piece.

[503] of the announcement, too.

[504] Yeah.

[505] Oh, yeah.

[506] Yeah.

[507] We're getting this guy as CEO.

[508] Also, he's currently employed.

[509] Therefore, we're going to get owned by them.

[510] So a curious case of a CEO search.

[511] But an incredible outcome.

[512] Seven and a half billion dollars is a lot of money.

[513] And David and I joke, but it's certainly not how it went down.

[514] It absolutely went down from the, the GitHub investors and the GitHub board and the GitHub management, you know, entertaining the sale of the company, hiring investment bankers, I believe Morgan Stanley in this case, to help make it happen, to get the best price, to looking around for counterparties.

[515] And then when the transaction, you know, in thinking about how the transaction was going to get done, I'm sure Microsoft thought, gosh, you know, the only way that this is going to succeed is if we really have the right person here at the helm and then a sort of process with the former Zamoran CEO to make that happen.

[516] But really interestingly leading with the leadership announcement in the press release because they sort of anticipated people are going to freak out about this.

[517] And they have to have a beloved open source community technology leader at the helm of this thing in order to give us a shot.

[518] Yep.

[519] Well, and this gets to, we alluded to this earlier in the episode, but GitHub is really the outcome of the rise of the open source community and everything that's done to transform software development and huge swaths of the enterprise over the last decade plus, but the historical enemy of that, mortal enemy, sworn, sworn enemy, has been Microsoft.

[520] And it's just such a funny testament to how much the world has changed now that that's not Microsoft and now Microsoft is buying GitHub yeah so story time from me being at Microsoft in one of my first few days as an intern I was specifically warned not to look at anything actually first it was not to look at any patents then it was do not look at any source code that has the GPLV3 license it's sort of the most strict crazy license where you can get in the most trouble if you are a company like Microsoft.

[521] And the threat, and I, you know, DPL3, DPL license being the, uh, good new public license.

[522] This is the, the, the, the core of legally what it means to be open source software.

[523] And there's other ones.

[524] So it's the GPLV2.

[525] There's MIT license.

[526] But, and I caveat, I'm not a lawyer.

[527] I haven't quite looked into this.

[528] But the, the thing always going around whenever you talk to people about this at Microsoft was the fear that if, uh, we used a bit of code that was GPLV3 licensed in our software, then we could be forced to open source the whole thing.

[529] Wow.

[530] Like all of Windows, all of office.

[531] And it was sort of unclear exactly what all of would be, but LCA legal and compliance affairs or something like that made that very clear.

[532] And so, and I think is this part of the whole philosophy?

[533] Because there's so much also like in open source and all of technology of the 60s.

[534] culture, you know, element embedded in.

[535] Is this part of the whole copy left idea?

[536] So there's copyright.

[537] I'm not sure.

[538] And then there's the idea of copy left, which was part of the open source software movement, which is not only is what you're making copy, you know, the opposite of copyrighted, it's that you have the obligation to contribute back, to contribute back to keep open and anything else you use in conjunction with it to open as well.

[539] Yeah.

[540] And so even one step further, I remember at one point I had GitHub open during my internship and got in a little trouble.

[541] And it was, it's kind of crazy.

[542] I mean, it was, it was, reprimanded by Microsoft for having GitHub open.

[543] Only eight years ago, eight years ago, something like that, I got in trouble Microsoft for having GitHub open.

[544] And like it seems absolutely insane to say, but, and, you know, it even seems insane like two years later to say because that was sort of the end of an era.

[545] Toward my end of my end at Microsoft, there was a group starting MS Open Tech that was a partnership between Microsoft is like a sort of Microsoft half spin out.

[546] And then slowly the dramatic adoption of open and source software in the company.

[547] So being able to do Linux on Azure and, um, uh, and to the point where Microsoft, of course, uh, we, we really, in the best interests of everybody, including ourselves, glossed over a lot of the history of version control software and software configuration management.

[548] But Microsoft has a massive historical product that they've sold a lot of money in the space called Team Foundation Server.

[549] And that was based on an old centralized version of source control management.

[550] But to the point where in 2013, they actually added Git support to TFS.

[551] Yep.

[552] And even today, they have another thing that is currently operating that, you know, many people outside and inside the company are sort of wondering about called Visual Studio Team Services.

[553] They shut down CodePlex about a year ago or maybe six months ago.

[554] And, you know, the company has this incredible developer division or a group of people that produce software to enable developers throughout the world.

[555] And a big part of that is things that looked like GitHub in different ways over the years.

[556] And so historically, they've competed with it.

[557] They're currently sort of competing with it.

[558] Um, there's definitely a, uh, um, you know, it's not like Microsoft hasn't thought about this problem.

[559] Yeah.

[560] Well, it'll be what will be interesting.

[561] Well, flash forward here to grading.

[562] I don't even know how we're going to put a grade on this, but, um, given it happened yesterday.

[563] But, um, will Microsoft keep those or will they, you know, go all in on GitHub?

[564] Like what's, what happens to the quote unquote Google video version of everything, you know, that's, as opposed to the, you know, that's, YouTube.

[565] Yeah, so the stated plan is keep it, and it serves a different purpose, and it's for different customers, blah, blah, blah.

[566] You know, we, David and I reached out to sort of several different birdies in preparation of this episode, and one, one source in the company said that they felt like it's, it's kind of only a matter of time until some executive gets upset and looks and says, how come we have two of these things?

[567] And then you have another Skype link merger.

[568] This is a different Microsoft.

[569] I don't think that will happen.

[570] I think this will look a lot more like LinkedIn, but, you know, LinkedIn was a business that was doing really great on its own and didn't compete with any internal Microsoft products.

[571] So there's, that one's harder to screw up.

[572] Yeah, yeah.

[573] All right, listeners, our sponsor is one of our favorite companies, Vanta, and we have something very new from them to share.

[574] Of course, you know Vanta enables companies to generate more revenue by getting their compliance certifications.

[575] That's SOC2, ISO 2701.

[576] But the thing that we want to share now is Vanta has grown to become the best security compliance platform as you hit hypergrowth and scale into a larger enterprise.

[577] It's kind of wild.

[578] When we first started working with Vanta and met Christina, my gosh, they had like a couple hundred customers, maybe.

[579] Now they've got 5 ,000, some of the largest companies out there.

[580] It's awesome.

[581] Yeah.

[582] And they offer a tremendous amount of customization now for more complex security needs.

[583] So if you're a larger company and in the past, you showed Vanta to your compliance department, you might have heard something like, oh, well, we've already got a compliance process in place and we can't integrate this new thing.

[584] But now, even if you already have a SOC2, Vanta makes maintaining your compliance even more efficient and robust.

[585] They launched vendor risk management.

[586] This allows your company to quickly understand the security posture of the vendors that you're choosing in a standardized way that cuts down on security review times.

[587] This is great.

[588] And then on the customization front, they now also enable custom frameworks built around your controls and policies.

[589] Of course, that's in addition to the fact that with Vanta, you don't just become compliant once, you stay compliant with real -time data pulled from all of your systems, now all of your partner's systems, and you get a trust report page to prove it to your customers.

[590] If you click the link in the show notes here or go to vanta .com slash acquired, you can get a free trial.

[591] And if you decide you love it, you will also get $1 ,000 off when you become a paying customer, make sure you go to vanta .com slash acquired.

[592] Well, should we move into acquisition category?

[593] We should.

[594] We should.

[595] I invented a new one for this because...

[596] Oh, innovation.

[597] I can't play by the rules.

[598] Are you going to open source your invention?

[599] It depends.

[600] Is our structure following a copy -left license?

[601] Absolutely.

[602] I'm going to contribute back.

[603] to our shared Google Doc.

[604] So I initially, so for new listeners of the show, we have acquisition category, which is people, technology, product, business line, asset.

[605] We recently, and somehow only recently, added consolidation or other, which is the part that gives David my license to do whatever we want because we're the hosts.

[606] I initially said product and not business line because if you look back at that stat of $200 million in revenue, they I don't believe we're profitable if you take another stare at that from the previous year the run rate in 2016 was that they did 140 million dollars of revenue and generate an 88 million dollar loss I don't think they had reversed that in the in the next year probably not the two years either but they could be closer um as is if if it was a 200 million dollar revenue business it would be a 38x revenue multiple purchase for a probably unprofitable business so I shied away from saying from saying business line.

[607] But as I thought about it more, I think product is right because they're going to they're going to sell this product through their sales channels to additional enterprise customers because Satina Della said that one of the stated reasons is to accelerate the adoption of GitHub and the enterprise and they believe that they can do that.

[608] I'm sure they can.

[609] They have the best or one of the best sales forces in the world.

[610] But I think there's another thing here that's sort of ecosystem.

[611] And that's that Microsoft was missing out on this ecosystem because, and particularly the developer ecosystem, because Windows isn't important anymore.

[612] And that's a ridiculous hyperbole statement.

[613] Windows is very important.

[614] But without the growing consumer base with the sort of enterprise shift to be more open to other things, they've sort of lost the ability to coerce developers by just saying everybody's on Windows, I'm pretty sure you're going to develop for Windows.

[615] So it's a way to sort of get the ecosystem of developers to sort of continue to be a friend of Microsoft.

[616] Yeah, so this is, didn't, so Ben Thompson's take on this was similar to this, right?

[617] Yeah, he described it as having the leverage to be able to, yep.

[618] Yep, Ben Thompson, of course, of strategy, which were a great admirers of here on Acquired.

[619] yeah it's okay interesting I can buy that it's like the equivalent of like strategic reasons I'm doing that air quotes here yeah yeah yeah okay so I had product I went through the same line of thinking about like business line like LinkedIn business line I don't remember what we said LinkedIn was but I'm I don't know how we wouldn't have called it a business line read the press release and it's like, you know, Nat Freeman, founder of Zamoran, will report to Microsoft's cloud and AI division chief, Scott Guthrie.

[620] It's within a business line.

[621] Like, yeah, of course.

[622] Like, this makes sense as a product.

[623] But I see what you're saying.

[624] Like, yes, it's a product, but they didn't buy this for the product.

[625] They bought it for the strategic place that GitHub has in the developer ecosystem.

[626] And particularly as the ecosystem of software developers and what a software developer is continues to grow and expand, you know, in the coming years.

[627] Yeah.

[628] And another way to put it is they bought a customer base.

[629] Mm -hmm.

[630] And obviously they're going to sell this product to additional customer base, but this customer base is super sticky to this product, and there's tremendous network effects within this product.

[631] And so, you know, I think they bought a customer base, and they bought that customer base for years to come.

[632] it almost recalls another famous saying within Microsoft or from one particular man who's no longer at Microsoft who we also did an episode on developers developers developers developers developers developers developers developers uh you've honed that well i've never done it before oh wow you're a natural oh it only took us how many episodes to to do that live on the air I think someday in the distant future when Sati Nadella is ready to retire I think we've got a good candidate for next Microsoft CEO here on a choir A return to the honey -pouring developers lifestyle We can have a bomber on at some point We'll redo the Clippers and have him on Yes We definitely have to have Bomber on For the Clippers Not for anything else for what's worth I'm a huge fan of post -Microsoft Balmer I think he's just killing it USAFax is an awesome project he's teaching it Stanford GSB I don't know if he still is but he was for a while I can't knock that okay what would have happened otherwise I think the specific question it's clear I mean we yeah we joked about it a little bit on you know in the history of facts but like when an announcement comes out like we just crossed 200 million in revenue and we're starting a CEO search like you're definitely looking you're gonna be entertaining acquisition offers right so I think the question is who else could was in the mix for buying this year.

[633] And every listener or lots of listeners, I'm sure you're also thinking Google.

[634] Like Google has to have reports are that they didn't formally but informally sort of entertain.

[635] Does this make sense for us?

[636] And Google was one of the main competitors of GitHub, of course, with Google Code.

[637] And they just waved the white flag, you know, three years ago and said, you won.

[638] We're done.

[639] Yeah, I think.

[640] So going back to the point from Strathex, Henry, Ben Thompson had a really interesting point that is that if someone like Google had bought GitHub, there would be enormous temptation to leverage it to funnel developers at your product.

[641] And the great thing about GitHub is that it's Switzerland.

[642] Like, it works for all these things.

[643] It gets to be a horizontal.

[644] And I think Ben is a little bit insulting Microsoft and saying, like, they don't have a platform to funnel these developers to.

[645] And then by product, you mean GCE, Google Cloud Engine?

[646] Google Cloud Engine, Android, probably not Chrome or Chromebook developers.

[647] But, you know, there's more sort of thriving developer ecosystems such that if they bought it, there's an open question of do they prioritize their own platforms on that and use this horizontal thing to feed the vertical?

[648] Certainly at least Google Cloud Engine.

[649] I mean, they're locked in a war with AWS trying everything they can to get market share.

[650] Like, if you've got all the developers in the world, you know, pretty much more than 100 % of the developers in the world, if you include me, on GitHub as like their primary user interface for their work, you could probably think of some bundling strategy to...

[651] Yeah, and I was wondering, is Microsoft going to do that, too, with Azure?

[652] Like, is there going to be some, like, ease of deployment to Azure?

[653] You can deploy to anywhere, but, like, we made this extra easy button.

[654] I don't know.

[655] Yeah, that's interesting.

[656] This isn't quite what would have happened otherwise, but I asked another super sharp CTO friend for his thoughts.

[657] And he mentioned that it's great for consumers or for developers.

[658] Microsoft's amazing at security and scale.

[659] But they could try to change Git upstream to add in enterprise features so that it's Git and then therefore...

[660] The open source project of...

[661] Yeah, exactly, maybe more suitable for the customers that Microsoft has in their sort of sales channel, but GitHub isn't a perfect fit for right now.

[662] And after digging into it with him a little bit, Git uses the GPLV2 license, which is copy left.

[663] And so if Microsoft decided to change anything in Git specifically for GitHub without going to sort of Git as a broad community and getting it, getting it approved by Git everywhere, regardless of GitHub, then they would have to open source whatever changes they made to their fork of Git used in GitHub to improve the enterprise functionality.

[664] So I think what Microsoft may do is try and get that support added to GitHub, and then it would come downstream to them, I'm sorry, to get, come downstream to them for GitHub.

[665] And then I don't know how this works, but maybe have sort of more leverage and lobbying power in doing that to make it more of a product that they can they can push through their channel but I don't know what does Linus Torvald's thing is my question yeah no kidding no kidding hopefully that guy goes on a podcast acquired follow -up so so another really interesting point um and this was on the motley fool podcast very interesting that this was an all stock acquisition very interesting so it's not like Microsoft couldn't afford the cash no Microsoft has $43 billion in net cash right now.

[666] And so there's a really strange thing that's going on that Microsoft has authorization to do $30 billion in buybacks.

[667] And they're in the middle of a buyback program, the same way that Apple is, which is basically saying, hey, we think our stock is really cheap right now.

[668] Like, we're going to pick it up on the cheap.

[669] We got some cash.

[670] We like holding our stock.

[671] It's show of good faith for the company for the future.

[672] If you think that.

[673] If you don't hold the stuff, it said they buy the stock and retire it, so they reduce the number of shares outstanding.

[674] so everybody's value goes up.

[675] They, uh, anti -dilution.

[676] So, but then they go and they buy a company in all stock, which if you have the option to buy in cash and you think your stock is going to appreciate and you think it's undervalued and you're doing this buyback, it's the opposite.

[677] So is it like somehow hedging against their buyback strategy?

[678] Is the signal that they actually think their stock is overvalued and so they want to buy and it's a good question.

[679] The other thing that the, the host of that show, postulated, which I thought was interesting, is maybe they're going to do something else with the cash, and that maybe there's another acquisition coming that is going to be a, require more cash.

[680] I don't know if we can read that much into it, but I thought that was interesting.

[681] It seems like Microsoft has so much cash.

[682] Yeah.

[683] Interesting to speculate why they did this in stock.

[684] Yeah.

[685] Which we should also say the founders of GitHub are now the largest shareholders in Microsoft other than Bill Gates.

[686] Yeah, we were talking about this earlier.

[687] So this must mean that Bomber has liquidated a lot of most of his stock at this point.

[688] I guess so.

[689] So here's the quote from Bloomberg.

[690] The each of the three co -founders would have 0 .16 % of Microsoft from this because the acquisition was about 1 % of their market cap.

[691] And so that's about 10 times more than Satya and Della.

[692] 14 times more than President Brad Smith.

[693] And among insiders, only co -founder Bill Gates, 1 .34 % would exceed theirs.

[694] So I guess the moral of the story is...

[695] More than Satina Della.

[696] Each of the GitHub co -founders now owns more Microsoft than Satina Dela.

[697] Yes, each of them.

[698] Wow.

[699] Yep.

[700] And so the moral here, I mean, there's a bunch of morals, but late capital raises.

[701] after bootstrapping.

[702] So super low dilution.

[703] They did that 100 on 750 raise.

[704] And then that 2 .5 on 2.

[705] I mean, you're just taking very little dilution there.

[706] So late capital raises after bootstrapping plus few capital raises plus an all stock acquisition equals you're one of the biggest shareholders of one of the biggest publicly traded companies in the world.

[707] It's funny.

[708] I actually have the opposite takeaway from that, which is that it's very nice to have a liquid public stock, where if you are an employee or a founder of a company, you can actually get liquidity unlike the world in which we live in San Francisco in 2018.

[709] So you're saying these guys could probably buy some, like a house?

[710] Yeah, a house.

[711] Imagine that.

[712] Yeah, they could probably buy a shack here in San Francisco.

[713] Wow.

[714] Cool.

[715] Okay, so any other points on that?

[716] So let's say they don't get acquired.

[717] Can this company go public?

[718] Like, do they end up doing like a DARA situation and bringing in a great staple CEO and trying to?

[719] It's a little bit like, I, in preparation for this episode, I went back, I re -listened to our episode on Atlassian, which we haven't talked about as much, but it's been a long day.

[720] They compete.

[721] We're going to leave it at that.

[722] In the Atlassian episode, we were grading their IPO.

[723] And I was like, this is an A. It's great.

[724] It was completely uncontroversial.

[725] This is exactly what you want.

[726] Like the narrative was super clear.

[727] And you were like, yeah, but I like the cowboy, like the James Bond style.

[728] You know, you're driving off, you know, towards a cliff at full speed.

[729] You pulled the e -break and turn around.

[730] I think that's what would have had to have happened with GitHub here to ultimately go public.

[731] I mean, they announced that the CEO was leaving.

[732] They've had a history of, you know, management turnover.

[733] There clearly was a lot of company building work left to be done here.

[734] at the company to get it to a point where it could go public.

[735] Not that it couldn't have been done.

[736] It certainly could have been.

[737] But this feels like a great outcome.

[738] Yeah.

[739] So interestingly, maybe if the assertion there is that there's a lot of company building still to be done, maybe there is going to be a lot more integration with Core Microsoft than LinkedIn had.

[740] Like, LinkedIn was a standalone public company, all the executive team in place.

[741] Jeff Weiner's still there, you know, doing great.

[742] Like, yeah, very different situation.

[743] Hmm.

[744] Well, I think let's do tech themes and then I want to hear predictions on what they will do with it.

[745] And I've got a couple also.

[746] All right.

[747] Tech themes.

[748] Do you want to go first on tech themes?

[749] Yeah, I think we've mostly talked about them.

[750] Microsoft's monumental reversal of philosophy on open source was a big one.

[751] point of integration is a big one, and this is getting back to that, that strategic comment from earlier, sort of about ecosystem, linchpin, bundling, point of integration.

[752] Like Microsoft used to have so much leverage in saying, you are a developer, you want to reach customers, Windows is the way, and then they would monetize by selling Windows licenses, and they would monetize by other platform features of Windows and selling Office on Windows and things like that.

[753] Without that enduring linchpin, you know, you have to take other measures to get people.

[754] If developers are a core element of your ecosystem, you have to take great measures to get those developers since you don't have the leverage of the platform that they need to develop for anymore.

[755] Now, the question still remains, how do you monetize that developer?

[756] Why is it important to get that developer if they're not going to build software that will go to 2 billion people that you will make a ton of money off the license of each of those two billion people that are using it because seven bucks a month or even the $2 ,500 a month for 10 -seat enterprise license of GitHub is not nearly as good of a business as selling Windows.

[757] So if you're not making enough money off the developer directly, how do you end up monetizing the software they create?

[758] So that's my sort of big question on the episode.

[759] But, yeah, that's a total key theme for me is using leverage of having users in order to get developers on your platform to be sticky and come and develop.

[760] And when you don't have that, doing things like this.

[761] Yep.

[762] I think we've hit everything.

[763] The only thing I just want to highlight again that I actually hadn't quite thought about until you brought it up earlier in the episode.

[764] So I'll bring it up again is the power of freemium models.

[765] if and only if you'd get the line you know exactly right and the incentives align between growth and usage and a customization of your target market to your platform with a very very easy to understand and aligned incentive way to monetize around that and GitHub just nailed it out of the gate might be I mean having a hard time thinking off the top of my head certainly you know top three if not best freemium model ever devised because if you think about it like ones that are time bounded are weird ones that are usage bounded are super weird because it's like oh it's free to use until like dropbox until you hit a certain amount of space well that incentivizes you to just like delete stuff when you get close to your cap um like this it's just really really really hard to nail this, but like, I've never seen anything that I can think of that nails it so well.

[766] Yeah, my general rule of thumb that we sort of use at PSL, when we're coming up with what's the right sort of pricing model for things in terms of thinking about pricing and packaging is if it requires network effects, it should be freemium.

[767] If it's a utility for you personally, it can be time -bound trials.

[768] And GitHub sort of took the best of both worlds.

[769] So to your point of being one of the best of all time.

[770] It's like up there with Spotify and in placing the line exactly in the right place to convert people.

[771] But even Spotify, I would say, like, it's still.

[772] The network effect isn't really important in Spotify.

[773] Because the little social sidebar thing got them off the ground, but it's not that important anymore.

[774] That was a distribution hack.

[775] Right, right.

[776] So this notion of like, by me using the free version of Spotify, does it make Spotify better for everyone else?

[777] No, not really.

[778] By me using GitHub for free, doesn't make GitHub better for everyone else.

[779] Absolutely.

[780] Yeah, absolutely.

[781] Cool.

[782] That's what I got.

[783] How do you want to approach this?

[784] Oh, man. Okay, well, let's first talk about what we think they're going to do with it.

[785] I'm, I'm, there's two things that I think they could do, and I'm so hand -wavy.

[786] Like, I'll say there's three.

[787] I've mentioned the one about pushing GitHub into the enterprise, so maybe, by opening up the Microsoft sales channel and recommending this to all the Microsoft customers that are developing on any part of the Microsoft stack, which, of course, may compete with visual VST or whatever that is.

[788] Whatever it's called.

[789] Yeah.

[790] You know, you could start turning this into a profitable business line.

[791] It doesn't feel like that's going to make a ton of money.

[792] probably not enough to earn back seven and a half bees anytime soon yeah um i think they will use it to leverage people onto azure i don't know how and i don't know how they'll strike the right balance of not being too heavy handed with that their version of what google would have done with google cloud engine yeah okay um and my left field one is since github It serves many other purposes, but it's sort of LinkedIn for developers doing some kind of LinkedIn integration or some kind of hybrid product could maybe see that.

[793] It's pretty left field, though, and I have a feeling they're not going to risk the continued success of the LinkedIn acquisition without integration by starting to integrate them with this acquisition so i don't know how they'll play together yet but i the the clear takeaway for me is it's it's it's the clear takeaway is it's not clear yet um how they'll ever make back this this purchase price and uh um maybe azure interesting i yeah i i don't have a ton of thoughts other than the thoughts i had were all in that last third bucket that you were talking about on the recruiting side because it really has.

[794] I mean, GitHub is, is the main, you know, resume for software developers now.

[795] And it's kind of like Microsoft is like a VC that has an investment thesis around like, modern, you know, what is the modern day resume in particular industries?

[796] It does cause, here's this guy actually an interesting thing.

[797] This might be what they, nobody said this, but like, it costs $30 ,000 to recruit an engineer.

[798] Well, okay, so this is where I was going to go with this.

[799] It's the most expensive position to recruit.

[800] Yep.

[801] And there's the least data on LinkedIn about the most valuable position to recruit.

[802] Absolutely.

[803] That's where I was going with this.

[804] So it's actually two things.

[805] Yeah, like that's, there's a big market there.

[806] Like, however big the market is for GitHub's services, which is big, for sure.

[807] The market for recruiting engineers is at least as big.

[808] you know what's the so what's the enterprise version of GitHub you're paying effectively $250 per seat per year right but it costs $30 ,000 to recruit an engineer who's going to stay at your company for somewhere between one and X years you know certainly not enough way more in the upfront recruitment fees than however much you're going to pay for a GitHub seat for yeah it's kind of interesting over that say they're there for three years it's $750 versus $30 ,000.

[809] Yeah.

[810] That's a big delta.

[811] And maybe there's also a piece of like, you know, who just like any other company in technology is locked in talent wars for recruiting engineers, Microsoft.

[812] Wow.

[813] Did they spend $7 .5 billion just to peek under the hood and see what private...

[814] We've been working hard here to acquire.

[815] These delusional thoughts are flying around.

[816] What is the, what's the, what's the, presumably the license doesn't let you do this and maybe there's even a GDPR thing around this, but like, imagine you're a rock star developer at Apple and Microsoft now has, I mean, actually they have access to these code bases, which I'm sure is completely off limits.

[817] But like, what signal could they mine about you?

[818] Or even like, you don't even have to go that far.

[819] Just like, it's funny we've been talking about for, for lots of reasons.

[820] Uh, but.

[821] Uh, but.

[822] Uh, Ben and me and, you know, as a way of, like, perceptions of Microsoft.

[823] And certainly, you know, vastly improved since Satya and Navella became CEO.

[824] However, still in the Valley, like, and I know, like, having lived in both worlds in Seattle and the Valley, like, I view Microsoft very differently than I think a lot of people in the Valley do.

[825] Like, to a lot of people here, Microsoft is still Microsoft.

[826] But Microsoft now owns LinkedIn, now owns GitHub.

[827] They almost certainly are going to brand LinkedIn, GitHub, just so that every single page view of an engineer everywhere, seeing them every day, gets a flash of the Microsoft logo as a soft recruit.

[828] Yeah, I mean.

[829] They have to.

[830] I mean, GitHub beloved by engineers.

[831] Either they have to, or that would be like Disney slapping the Walt Disney Castle at the beginning of Star Wars.

[832] Either they have to, or it's completely off limits, and I can't decide which one.

[833] Yeah, I don't know either.

[834] So, David, we're dragging out this episode length because neither of us want to answer the hard question.

[835] What would you grade this acquisition?

[836] I've been trying to avoid this for like an hour plus at this point.

[837] I don't feel good about it.

[838] I'll just, I'll throw it.

[839] Okay, yeah, whatever.

[840] So I think the question facing us is like, what do we think, like, you know, fast forward five years, how are we going to look back on this?

[841] So the reason I don't feel good about it is because it's so much cash for a money losing thing that doesn't actually have that much revenue.

[842] So then it's clearly not a business line that never will be a business line.

[843] line to Microsoft.

[844] It's a strategic thing.

[845] And the point I was going to make when I was looking at it more as a business line was at what price would I like this deal?

[846] Is it a billion dollars?

[847] Is it two billion dollars?

[848] But if it's a strategic thing, and I don't quite understand exactly how it fits in yet, if it's worth two, surely it's worth seven and a half.

[849] There's this great quote.

[850] We do need to do a Twitter episode someday, but there's a legend and lure that Facebook tried to buy Twitter for, I think, I think something like $500 million or a billion dollars.

[851] It's publicly reported, whatever it was at some point, early in the life cycle.

[852] And I believe it was Ev Williams after the meeting where they talked to prices.

[853] Like, you know, if we're worth a billion dollars, we're worth 10.

[854] that's that's the and they were the equivalent and they were uh more than that as a matter of fact um okay yeah for me i oh so you what's your grade i i hate this it's super high variance but i'm going to go c plus okay um and hope sotia makes me look like an idiot in a couple years great great i actually think the most likely path here um um I hope this isn't a cop -out, but I don't think so.

[855] Is it just lands somewhere in between, you know?

[856] I mean, like, Microsoft's a huge company, lots of divisions.

[857] Like, if this were LinkedIn, like, clearly it's going to be standalone, like, you know, a little more.

[858] But, but, like, there's some high degree of difficulty here.

[859] But Microsoft's, like, you know, Souti gets it, like, all this stuff.

[860] But it's just going to be hard to really pull this off.

[861] So I think it probably ends up somewhere in the muddy middle.

[862] I'm going to give it a B. A muddy middle.

[863] B. Yeah.

[864] Yeah.

[865] All right.

[866] Fair.

[867] Carvouts.

[868] Carvouts.

[869] Mine is from Chris Dixon, great blog post called The Idea Maze.

[870] I'm sure many of you have already read it.

[871] It's been floating around the web and I'm a little late to the party here.

[872] But it is a really great read.

[873] It's short.

[874] And it's a lot about basically teasing out the, there's a phrase that floats around.

[875] It's not about the idea.

[876] It's about execution and ideas are worthless and blah, blah, blah.

[877] And he really really.

[878] teases out, like, ideas aren't worthless.

[879] We just refer to the idea as, like, your original two -minute pitch.

[880] And a company is actually a series of ideas.

[881] And execution is really a bunch of ideas.

[882] And so to the extent that ideas matter, or to the extent the execution matters, then there's a bunch of ideas along the way that matter.

[883] And it's a really great blog post about sort of the things you realize while you're executing, reasons to get out into market, ways that you can sort of learn faster to make better decisions and to basically be an informed, continuous idea person as you operate your business.

[884] And so I highly recommend that.

[885] Mine related to that blog post that went around the internet, the interwebs about a week ago, Eugene Way, former Amazon employee and part of the original team at Hulu, wrote a great blog post called Invisible Asmatotes about his time at Amazon and they're in the relatively early days, I think just after going public days, when they started thinking about the business in terms of what are the ceilings to growth and market size that we're going to bump him to?

[886] How do we address those in the product and then how do we start adding more?

[887] really cool, interesting, unique take on thinking about strategy and tech that I hadn't thought about before.

[888] Highly recommend, if you haven't read it already, which I assume many of you already have, well worth doing.

[889] Cool.

[890] Our sponsor for this episode is a brand new one for us.

[891] Statsig.

[892] So many of you reached out to them after hearing their CEO, Vijay, on ACQ2, that we are partnering with them as a sponsor of Acquired.

[893] Yeah.

[894] For those of you who haven't listened to...

[895] Vijay's story is amazing.

[896] Before founding Statsig, VJ spent 10 years at Facebook where he led the development of their mobile app ad product, which, as you all know, went on to become a huge part of their business.

[897] He also had a front row seat to all of the incredible product engineering tools that let Facebook continuously experiment and roll out product features to billions of users around the world.

[898] Yep.

[899] So now Statsig is the modern version of that promise.

[900] and available to all companies building great products.

[901] Statsig is a feature management and experimentation platform that helps product teams ship faster, automate A -B testing, and see the impact every feature is having on the core business metrics.

[902] The tool gives visualizations backed by a powerful stats engine unlocking real -time product observability.

[903] So what does that actually mean?

[904] It lets you tie a new feature that you just shipped to a core metric in your business and then instantly know if it made a job, difference or not in how your customers use your product.

[905] It's super cool.

[906] Statsig lets you make actual data -driven decisions about product changes, test them with different user groups around the world, and get statistically accurate reporting on the impact.

[907] Customers include Notion, Brex, OpenAI, FlipCart, Figma, Microsoft, and Cruise Automation.

[908] There are like so many more that we could name.

[909] I mean, I'm looking at the list, Plex and Versel, friends of the show at Rec Room, Vanta.

[910] They literally have hundreds of customers.

[911] now.

[912] Also, Statssig is a great platform for rolling out and testing AI product features.

[913] So for anyone who's used Notion's awesome generative AI features and watched how fast that product has evolved, all of that was managed with Statsig.

[914] Yep.

[915] If you're experimenting with new AI features for your product and you want to know if it's really making a difference for your KPI's Statsig is awesome for that.

[916] They can now ingest data from data warehouses.

[917] So it works with your company's data wherever it's stored so you can quickly get started no matter how your feature flagging is set up today.

[918] You don't even have to migrate from any current solution you might have.

[919] We're pumped to be working with them.

[920] You can click the link in the show notes or go on over to stat sig .com to get started.

[921] And when you do, just tell them that you heard about them from Ben and David here on Acquired.

[922] Well, that's it, kids.

[923] If you aren't subscribed and you want to hear more, you can subscribe from your favorite podcast client.

[924] If you you feel so inclined.

[925] We would love a comment on Apple Podcasts.

[926] Feel free to share with your friends.

[927] Tweet.

[928] I hear that's doing well.

[929] Twitter's back in the S &P 500.

[930] Yeah.

[931] Acquired is blowing up on Twitter, by the way.

[932] We got like 50 new followers last weekend.

[933] You could be that 51st.

[934] All right.

[935] Thanks, everyone.

[936] Have a great night.