Best of: Forging a New Path with David Heinemeier Hansson

Episode 106 • 9th May 2024 • Innovation and the Digital Enterprise • DragonSpears

Patrick:

Hello fellow innovators. This is Patrick Emmons. Today we're resharing the insights from our previous guest on the show, David Heinemeier Hansson. David is an incredibly accomplished technology executive, co-owner and CTO of 37 Signals, makers of products such as Basecamp and Hey. Last year we had David on the show and his episode made some waves. So today we're sharing it again for those who did not hear it yet, and for those who want to hear his insights again. Here we go folks.

Patrick:

Hello fellow innovators, this is Patrick Emmons.

Shelli:

This is Shelli Nelson.

Patrick:

Welcome to the Innovation and Digital Enterprise Podcast, where we interview successful visionaries and leaders, giving you an insight into how they drive and support innovation within their organizations. Today we're thrilled to welcome a true Renaissance man in the realm of technology and beyond. He's not just a tech luminary, but a multifaceted innovator whose influence spans from coding libraries to endurance car racing. Meet David Heinemeier Hansson, the co-owner and CTO at 37signals, and the brilliant mind behind Ruby on Rails, a coding framework that has revolutionized the way web applications are built. But that's not all, David is also an accomplished author who has co-authored an impressive range of books that delve into agile development and work culture.

r, having participated in the: 2012

David:

Thanks so much for having me.

Shelli:

David, if you don't mind, can you share with our listeners a little bit more about your role at 37signals?

David:

Chicago all the way back from: 2001

Patrick:

I think everybody I know in software engineering follows what you and Jason say. I'd say almost on a weekly, somebody brings up an article that you're involved in. It just happened last week with one of the CTOs that I work with, and it's amazing how long and how big of an impact you've had in the software engineering space here, not just in Chicago but globally. Obviously, I think most of our listeners are going to be familiar with Basecamp and some of your other products, but I'd love to know more about HEY.

David:

nt product, we launched it in: 2020

It's basically builders and people who support the products. We thought, you know what? The traditional thing to do now is to hire a bunch of people, maybe even raise money. We have four products, they're all showing traction, you should be firing on all those cylinders at once. Well, we chose the other path, which was to say, do you know what? We'd like the size of the company, what do we have to do with our product portfolio to fit this size of company? If we only want to be 40 people or thereabouts, what do we have to cut? We ended up going all in on Basecamp. We called it Becoming Basecamp, so much so we even renamed the company from 37signals to Basecamp in 2014 and just doubled down on Basecamp.

hink, when did that come out?: 2004

Patrick:

Sounds right.

David:

At the beginning I thought like, wow, Gmail is just such a revolutionary product. I mean, I remember just going, what a gigabyte of free space, that's incredible. Right? Like, when everyone else will be offering you 20 megabytes, or oh, you got to constantly delete things, make sure your attachments are not too big, Google comes to the scene and they immediately just conquer the email space. I think today in the US, Gmail is responsible, something like 80% of all email that flows back and forth, which is just truly incredible. But it also means that it ended up being a very stagnant product, because that's just what happens to any kind of product that has that kind of success for the long term. I think Google has a billion and a half users on Gmail now, if they even just change their comma somewhere, they're going to hear from 50 million people who really dislike that decision.

It's very hard to innovate once you've established such a position as they have. I just thought, you know what? I've been emailing since '93, '94, I've been emailing for a long time. I've sent literally probably hundreds of thousands of emails. I have a lot of opinions about how to make better email software, so when we were looking around for doing something else, we were thinking, do you know what? Email is something we should tackle. Which in many ways is just crazy. Who in their right mind would want to compete against, not just Google, but essentially perhaps Google's most well-liked product that is given away for free? I mean, that just sounds like a suicide mission to anyone pitching investors.

You know what? We should go up against Google with a paid product. That's what we did. We thought we had enough good new ideas that we could make a paid product, that we could charge customers who really cared about sending email. This was the thing that motivated me so much about HEY was, a lot of people would say they hate email, and then I would dive into why they hate email and it would usually be around because they got a bunch of emails that sucked up their attention that they didn't want. We started talking about what's email actually good at? Well, it's good for long form writing, it's good for long form thinking. Everyone still liked those things, it was just buried under so much stuff, and so many emails you didn't want to see, everything mixed together.

possible on Gmail since about: 2004

Patrick:

Tell me more about the exactly... Forgive me, I haven't had an opportunity to try the product, but I definitely am going to go check it out as soon as we're done, because I do feel that email is not going anywhere. Right?

David:

Exactly.

Patrick:

It's critical for what it's supposed to do, not things you wouldn't put into Slack or Teams or things like that. What is the real advantage? What is it that HEY does, and how does it do what it does?

David:

Let's start with that first annoyance that I had. I kept receiving emails, and usually multiple emails from people I wasn't interested in hearing from. In my position as CTO and a co-owner, you get pitched a lot, by salespeople, by vendors, and some of them are kind and courteous and a lot of them are not. If you're trying to be kind back and like, oh yeah, we are not interested in this product right now, often that turned into this whole conversation that would just suck me in and we'd go back and forth and all of a sudden I've traded four emails with this person who've sucked up my attention for something I didn't ask for. I didn't ask to be part of this conversation, I prefer not to even be in it, but the way email works is that it just puts the next thing someone sends right at the top of the list, and it's their choice. Someone who gets a hold of your email address can simply invade your attention span at their choosing.

one just sitting somewhere at: 11:45

If you receive an email from someone you haven't heard from before, it goes into this quarantine space, it doesn't pop at the top of your inbox and it sits there until you add your leisure, go through the list and go like, do you know what? All right, I'll hear from Alex, that actually sounds interesting. Boom. He's going to go into my inbox, I can now reply to him, we can have a conversation. But do you know Ben here? No. We literally just have two buttons like [inaudible 00:11:40]. It just goes in like this. You push the thumbs down button, you will never hear from Ben again. Ben can send you seven emails, they're never going to land in your inbox. That's one example of sort of taking power back over email and restoring it to a place you actually want to be, because the things that land in your inbox are things you're interested in reading, people you're interested in hearing from.

The problem with email is that that often gets drowned out by all the people you're not interested in hearing from. As a media screener, it was one of those huge aha moments. We switched quite early in the development from Gmail to using HEY, and I just saw the amount of time that I spend on email just drop. I used to be this obsessive inbox zero person. I would pride myself on just getting to inbox zero as quickly as possible, which meant I was just puncturing the attention of my day by constantly checking emails and dealing with people like Ben who I didn't ask to hear from, who I didn't get any value out of, and it was disrupting my day, it was making it hard to focus, it was making it hard to concentrate, and I just immediately got that sense of bliss when we put this screener in place.

You know what? This is just a whole realm of work that I no longer have to deal with. In fact, the first sensation I got was one of regret. Why did I do this for 20 years? How did I waste so much time for so long working like this, giving people access to my attention like this? That is just absolutely crazy. That's one of the examples. The other example that I really love as someone who receives a lot of email and actually enjoys hearing from listeners, hearing from readers, hearing from customers, is to deal with things in buckets instead of one by one. I think a lot of the problem with the inbox zero style is, I'd look at the inbox, I'd see something from, for example, a customer who is writing me something expecting to get a reply and I'd go like, I actually want to reply to this customer, but do I need to do it right now, this moment?

If you're running inbox zero in Gmail, you do. Now, people come up with their own schemes and they have folders and they build their own little systems. Well, we took a lot of those systems and we just made them better. This feature I use in HEY for this is called Reply Later, so whenever I get something in, like an email from a customer, I'll simply just push L, it goes into Reply Later, which is this special mode of HEY, where I can respond to a whole stack of email in one sitting. This stack, I'll often just tackle maybe once a week, once every two weeks, and I'll knock out a hundred or 150 replies in maybe 20 minutes.

It's all sitting there right now. Now, I could have dealt with that one by one as it was coming in. It could have constantly sapped a little bit of my attention during the day, through that whole week, or I could just deal with it all at once in one dedicated setting, and I thought that was another one of those revolutionary things for me. I get to get back to people. It doesn't have to be immediately, a lot of emails does not require an immediate reply. They're happy to get a reply, but I can do it on my timescale and I could protect my attention. Just those two features alone, and we have literally dozens and dozens of other features that are like that, tweaking the ways you deal with email, how you find things, how you reply to things in a more efficient way that just makes it more pleasant to be an email user. A.

I think especially if you use email for business and you receive and send a lot of email, just so much more efficient, so much time getting back, I had a testimonial from one customer who actually runs this time tracking software where they see exactly where they're spending their time so they could measure precisely what the drop was. I think this person went from spending something like six or seven hours a week in Gmail to spending something like an hour and a half in HEY.

That kind of quantification is just really a testament to the fact that email can change, email can be better. The protocol itself, as you said, is not going anywhere. Email is the lowest common denominator, it's what everyone has, it's how you reach anyone anywhere outside of your organization and it's really wonderful for that. There's been so many attempts at killing email over the years. Every new productivity tools would come out and say like, whoa, we're going to kill email. I think to some extent that has happened internally inside of organizations. Organizations use chat or they use Basecamp or they use something else, but between organizations, email still reigns supreme for a very good reason.

Patrick:

The killing email, the way I look at email is more of a, you've got the micro messaging like Slacks and the Teams of the world and that's fine for immediate communication, but from a long-term, how many people use their email as pretty much their to-do list of this is the things I have to get to today, or at least it's one of the areas that people use as their to-do list? To your point, you said, hey, people get control over my attention. I don't know if you've seen the pattern recently with a lot of this unsolicited inbound? They'll actually even say, hey, I just wanted to nudge this back to the top of your inbox-

David:

They give the game away.

Patrick:

They are fully aware. Right. I mean, fully aware of like, I know you're ignoring me, but I'm willing to be this pesty.

David:

Yeah, let me bring up one other feature. This pesty, hey, I'm just nudging it to the top of your email, that's bad enough. There's a version that's even worse, that I've seen repeatedly. That's the kind of people who put the little spy pixels into email so they can actually see whether you've opened your email, when you opened it, how many times, how long you spent looking at that email and where you were. They could do the IP lookup. Oh, so I see you are in Chicago. Hey, you've actually already looked at my email three times, what's up? Why are you not giving me a response? I just thought, that is so incredibly creepy and we want to do something with that in HEY. In HEY, we have this name and shame feature when it comes to spy pixels. When someone sends me an email, if we detect one of the 50 or 60 different patterns of spy pixels that our system knows about, it'll say at the top of the screen, hey, Jay is trying to spy on whether you open this email or not.

And then you at least know. You know all the information that would've been sent to Jay, unless you used HEY, we strip out the pixels so they never see whether you open the email or not. But it's just so insidious, and to me, actually kind of revolting that that has become accepted, that that's accepted, that you can spy on whether someone opens your unsolicited email. Would we stand for that if someone was sitting in a tree across the street from you with the little binoculars spying on whether you opened your mailbox and see like, oh, you opened my mail, I can see that, right? You'd go like, no, this is super creepy. The thing is, most people don't even know about spy pixels. They don't even know that this is a capacity that salespeople and marketers were the two categories who abused this the most, in my opinion, are using these tools against them, so we wanted HEY to also be a bit of a shield to keep kind of the bad tactics of email out of your inbox.

Shelli:

Wow.

Patrick:

Totally, and yeah, they can tell if you're on your phone, they can tell if it's actually a desktop.

David:

It's so invasive. It really is.

Patrick:

One of the things that you touched on was, taking off HEY and the pandemic was kind of settling in. How did that impact your organization? Obviously, I think we talked a little bit earlier that you were still here in Chicago at that point in time.

David:

ed working with Jason back in: 2001

They were just as meetings heavy, they were just as likely to be on a synchronous time schedule, which just doesn't work very well in the remote world, which is why we saw articles about Zoom fatigue about five minutes into the pandemic when everyone realized, oh, you know what? Sitting on Zoom for four or five hours every day is actually totally miserable in a way that even meetings in person are not. All of those we ruled, but the pandemic still had a big impact because I think even after you embrace remote work, you have to realize that humans are social animals. If you confine them to a closed space with few other individuals, perhaps no other individuals that they interact with on a daily basis, I mean, some of them might literally go crazy and the rest of us might just start feeling really bad.

I think it was just a traumatic time for a lot of people, not equally so. This was one of the interesting things I thought during the pandemic. I have three small kids and our family here, I got a lot of human interaction and stimulation during the day, sometimes too much human action and stimulation, and then on the other end of it, we had people we have who were single, who live alone or whatever, who were just totally isolated from months at the time, so it really hit differently, but what we realized was the lack of that human connection between coworkers. We meet up as a company twice a year, we bring everyone together, even though we're remote, into a city. It used to be Chicago for many years when we had the office there, now we do it around the world. We are actually going to Barcelona for our company meetup in three weeks, I think it is.

Patrick:

Awesome.

Shelli:

Nice.

David:

you guys write a book back in: 2013

Patrick:

I totally agree. You mentioned some of those leaps, right? Well, you mentioned a couple of them. What do you think is the biggest lesson that you had already learned, that maybe some people are still struggling with?

David:

Yes. The number one thing that I think you have to do to become a successful remote organization is to switch your mode from being primarily synchronous about scheduling meetings, where we have to be at a certain time at a certain place with a certain number of participants, to being primarily asynchronous, where instead of using meetings for, say, distributing information, we write things up in Basecamp, we let people read it on their own time, and then we use meetings and synchronous collaboration sparingly. We use it to resolve conflict. We use it to do full on blue sky thinking very rarely. It's mostly about conflict, it's mostly about when things are difficult or we don't agree. Usually our process is, let's say there's a proposal for a new feature in Basecamp. Someone will write it up and post it to Basecamp, and if we all agree, good, we're done, let's do it.

We're going to work on that next cycle. If we don't agree, and after going back and forth in written language a couple of times, we still don't agree, we always say, do you know what? Let's upgrade to some high bandwidth here, and then we schedule a call and we get on the call and we resolve those things in real-time. I think real-time is really great for that. Being in-person is even better, but I think it's not that much better. For me, I think we get 95% of that comfort resolution capacity out of something like Zoom. We just try to use it at a pinch. We don't fill up our calendars with a lot of standing meetings. Standing meetings are often about updating people on the status of a project or generally conveying information, and you can do that so much better, in my opinion, in written asynchronous form that allow people to put their schedule together as it fits for them.

I have this pattern where I sit down at my desk in the morning, I usually spend about an hour just catching up on things, but then I spend a long block, maybe as much as three or four hours fully dedicated on just trying to crack those hard problems that I'm interested or is important for the business without constantly seeing my schedule interrupted. I've seen a lot of, especially executive schedules, and they look like a game of Tetris. They look like all these blocks are falling down, they're all different colors and a lot of them are very neatly, tightly packed, such that there's only these small slices of free time as we call it, which is complete misnomer.

Free time is not free time. That's the work time, that's what you're supposed to make progress on the things that are actually assigned to you. I try to see it the other way around that these impositions on my calendar, I group it together first of all, such that I try to do them all on one day. For example, I have another interview today and I did some other calls this morning and then tomorrow I have nothing, literally an entirely free day. Again, it's not free. An entire day dedicated to just solving the business problems that I'm really here to do, that's going to move the business forward, that I'm going to feel good about at the end of the week. I think one of the problems with a highly synchronous work style is that people can end up feeling very busy without accomplishing much.

That is one of those insidious sensations, when you sit on Friday afternoon and go like, wow, what a week? Because it was really busy, but then you think, what did I actually get done? I mean, I talked to a lot of people, but where's the tangible stuff? I think this is particularly something for builders. I'm a programmer and Jason does design, I love writing, so I have this sense that for me to feel good at the end of the week, I need some tangible representation of that work.

Not that having conversations do not work, or making decisions do not work, or planning can't be work, those things absolutely all can, but the lion's share for me has to be more physical progress that I can... For us, for customers, where are we delivering value? When we ship new things, when new software goes out into the world, they don't see the deliberations, however much value they occasionally can have. They see a shipping, and that's really what I think these organizations who've been in a particularly in-person mode for so long, been used to having everyone around, such that scheduling a meeting isn't a big deal because Jay and Jane are just around the corner and we can just book a conference room and we'll jump into that.

That just doesn't work in a remote world, and making that switch is quite difficult. I think what we're seeing right now is a lot of executives who couldn't make that switch, and that's why we're seeing this return to office drive, that there's a lot of executives like, I don't feel productive, me personally, nevermind the fact that they might have a ton of workers who go like, I do not want to get back into the car, I do not want to get back into my commute, can you please just let me work from home? A lot of times people get called into the office and then they sit on Zoom anyway, which really is just adding insult to injury, to call someone into an office, subject them to commute, and then subjecting them to Zoom for hours after that.

Patrick:

One of the questions I have, because I agree, and I'm a big fan of, I follow a very similar time system, and something you mentioned before with free time, dedicated, actual free time, right? Time away from work, turning off the phone, turning off email, all of those things. That's its own thing and should be protected. That's the way I work it backwards. That idea of controlling your week, making sure that you've got what we call as focus time, right? You've got your free time as you use it. I use the term focus time like, this is when you should be focused on the most important stuff, and then you have more buffer time where it's like, here's the maintenance, the hygiene, the not the critical but necessary, and trying to minimize the buffer time all the time.

My question though is, you're an entrepreneur, I'm an entrepreneur, Shelli's an entrepreneur. Is this something that everybody is capable of? I know when I talk to people like you and I know that they've got a lot of control over their lives. Is this something that other people can do? Is this something that you think that they've got the freedom and capacity? Sometimes I see very busy people and they really do value being busy. For some reason it drives value to their self-image. That's one of the things I'm curious about, I don't feel that way, so I can't really speak to it, but is that something that you've seen, or what are your thoughts on that subject?

David:

Oh, I see a ton, and I think a lot of this lies deep in culture. The American work culture is just one that cherishes busyness, that looks to its champions to be the one who works the hardest, the ones who can put in the most hours. I've seen time and again in technology and entrepreneurial circles, people brag about working 80 hours or a hundred hours as sort of tokens of showing their dedication either to their customers, to themselves or to their investors, and I'm just like, you know what? First of all, for me, that would just be exhausting to work like that. I wouldn't be able to go the distance that I've gone working like that. I think perhaps this is also part of it, that there are realms inside of entrepreneurship who look at things on as short-term sprints, even though they're not actually that short.

'd like to work at [inaudible: 00:30:47

You can miss a larger lens that I would actually argue helps you at work. I think I'm a better business owner because I have other things going on outside of work, because I have hobbies, because I have family, because I spend time away from the computer. Oftentimes, I get my best breakthroughs when I'm not sitting in front of the computer. If I was sitting in front of a computer for 10, 12 hours every day all the time, I think I'd just miss out on a lot of key insight and key breakthroughs. Some of it isn't the culture. I come from Denmark originally where this culture is quite foreign. People take almost pride in going home exactly on the dot, which sometimes is also not entirely flattering. Right? I can certainly relate to the fact that I'm dealing with this important interesting problem, I just want to finish it so I put in a few extra hours. I also looked at sort of the moving average of my time at work as being, you know what? 40 hours, that's plenty.

We can get all the things we need to do done in 40 hours, because the thing is, there's an unlimited amount of work if you run a business. You may work a hundred hours. There's easily 200 hours or 300 hours or 500 hours you possibly could, so you have to cut it anyway, and the way I find most problems at work, there are power law. There's about two, three things every week, perhaps, maybe sometimes even just one thing, some weeks even zero things that are really important. If I get those done, that's like 80% of the value of me being present, then there's all this other stuff, all these other conversations, all these other checkups, all these other things that I possibly could do, that can easily get me to 80 or a hundred hours, but the value for me, the lion's share of the value is from zero to 40. When you go from 40 to 80, you're not getting another half of the value.

In fact, you're diminishing, in my opinion, a lot of those 40 hours. This is one of the reasons I'm such a huge proponent of sleep. I love sleeping. I need eight and a half hours preferably every single night. I take zero pride in getting six hours. That's another one of these tokens of workaholics who are like, oh, I only need four and a half hours. Okay, that's possible. You are in literally the 0.1% of humans where that is physically true. More likely you're probably just delusional and you are impairing your cognitive abilities by walking around like a, I want to say brain-dead zombie, because that is basically what it is at times, but it's also a little bit like being drunk.

You don't actually always know how much you're impairing your own creativity and capacity by being under slept or overworked or burned out, and you don't really see those things until you get to a state where those things are not true, and since we run our company in this way where 40 hours is enough, that's my default state, so I see the dips very clearly. I don't always get eight and a half hours, but the days where I only get six, I go like, jeez, is it even worth showing up for work? It's certainly not worth opening a difficult problem that I'm just going to try to crack because I'm going to spend three times as long trying to figure out what is up and down, and I'm probably going to do a bad job at it.

Patrick:

Well, and I think, one, there's a really ironic, I used to tell all my engineers when they're like, oh, we're going to work through the night, I'm like, I'd rather you just go get drunk. Honestly, I'd rather you go to the bar and just get hammered, and they're like, why? I'm like, because at least if you do that, you'll go to bed at one or two, you'll sleep it off and you'll be about 50% functional by noon. If you work through the night from midnight to eight, you're going to be zero productivity, and probably to your point, negative, you're going to do more damage than good and then you're going to sleep from eight to about three, that's not healthy, you're still knocked out for that whole day and you'd probably be 50% by the day after. I'd rather you just go get drunk. That was my standing order. It's like, if you are saying you're going to work through the night, just tell me, I'll give you a credit card, you can go hit the bar.

David:

I think this is one of those things where there are moments in the 22 years I've been at 37signals, we've had maybe three, four true emergencies, where, do you know what? Whatever time it takes, we're going to fix it. The servers are down. I'm not going to tell people when the servers are down, okay, let's just go home and sleep it up. We may rotate, but someone will be working on this problem 24/7 until it is resolved. It is exceptionally rare. It should be exceptionally rare. If it is not exceptionally rare, something is really broken. I see this thing with the 24-hour thing every year, I do the 24 hours of Le Mans, this motor racing in France-

Patrick:

Yeah, that's crazy.

David:

Where we're actually up for about 36 hours with only a very little bit of sleep after that. It takes me a week to recover from that before I feel like I'm back to peak capacity, so you're really borrowing from a loan shark and that loan shark is charging like 300% interest every day. Okay, if there literally is no other options that perhaps the business is even at risk unless you put it all in, fine, do that, but don't take it as something heroic, don't take it as something to be proud of.

Patrick:

Right, don't celebrate it.

David:

If you're in that situation, it's because you fucked up.

Patrick:

100%. Totally. The worst is, to your point, David, you're dealing with client facing software, you've got performance issues, you've got SLAs, you've got a lot of people looking. How many people do it because of some arbitrary date that was picked on a project plan, and the only reason why you're doing it is nobody wants to make the phone call, but the truth of the matter is, you could just make the phone call, okay? You can either tell them now or you can tell them afterwards.

David:

90% of emergencies are made up.

Patrick:

Totally.

David:

You've just decided this is an emergency. We could just decide it's not an emergency and it's no problem at all.

Patrick:

Right. This is my question. Is anybody going to die? No. Then it's not an emergency, right? I say this to my kids, they knock on the door and it's like, hey, are you on fire? No. Go away. Right? I'm busy, I'm doing things.

David:

Yes.

Patrick:

You don't get to pick what my attention is. That's awesome stuff. I'm pivoting a little bit. I know I read some of your articles recently and I think you've got some amazing thoughts, quite a few that I share around, the value and the purpose and intent and the placement of investment in cloud technology. Do you mind sharing a little bit about your current thinking around that, and what do you think is going to be critical for people to be successful in the next couple of years?

David:

Absolutely. We started well over 20 years ago when cloud wasn't a thing, it was not invented as a term, no one were renting their computers from these hyperscalers as they are today.

Patrick:

We didn't even have virtualized machines back then.

David:

n. In sort of the mid to late: 2000

We went into that and we bought it. The premise at the time was, it was going to be faster, easier, and cheaper. That was the pitch, right? This is why everyone should be going cloud. This is why you're foolish if you're staying on your own computer because it's going to be easier, faster, and cheaper. Well, it is faster to some extent. If you need a hundred servers online in the next five minutes, nothing beats cloud. Cloud is truly amazing. The amount of computing firepower that you can enlist in about, not even five minutes, 30 seconds if you are instrumenting it, it's truly marveling, just incredible.

How often is that a thing that's really critical for your business that you need a hundred servers in the next five minutes? For our business, almost never. Literally, almost never. I can think of one instance where we really had big benefit of being able to enlist a ton of servers at the same time. This was at the launch of HEY. We had estimated how many users we were going to sign up for the service and we signed up just a hundred X, because we had this incredible launch. Right?

Patrick:

That's awesome.

David:

cheaper. Our cloud bills for: 2022

Cloud punch is a 3.2 million, might not be large for, I don't know, JP Morgan or something. They spend hundreds of millions on this stuff. For us, 3.2 million, this is recurring cost, we have nothing to show for it. When that period is over, we've just set fire to 3.2 million, it is gone, there's no residual value, there's no nothing. When we did this math and the math was so shocking, we also started doing the math on, was it easier? Were we able to run our services in the cloud with a smaller team? Could the same size team do more? The answer was just no. We've tried really hard for many years, five, six, seven years to make this happen. We have really smart people, so if we could not make it easier to be more productive, something was just off in the fundamental calculation.

Again, not for everyone. There are companies who have these huge spikes where suddenly you need a hundred X servers. Nothing beats the cloud. There are very small companies who don't even have full-time people managing their stuff, the cloud also makes sense, but there's a very fat middle where established companies who can imagine writing off equipment over two, three, five years are spending gobs of money on renting stuff from the cloud that they do not need to rent, that does not make things easier for them, and that's funding the whole Amazon empire.

If you look at the profit and loss statement for Amazon, basically all the profit is coming from the cloud services division and it's coming at ludicrous margins between 30 and 40%. When you think, do you know what? If you ask Dell or HP, what's the margin on a computer? It's not 30 or 40%. It is in ludicrous margins that they're able to reap in, in part because there's heavy lock-in effects, in parts because there are some moats of scale, but it also gives you a clue of where that money is coming from. It's coming from me and you. If you're running a service, you are the one paying for that margin, and I just thought, you know what? We have better things to do with our money than send so much of it into Amazon's coffers.

Patrick:

It's something I've thought recently as well, is like, everybody is looking to get rid of their data centers, but there is a certain amount of compute that you could if you were forecasting for the future. To your point of, there's areas where you're so small, it doesn't matter. Carrying a full-time IT resource, that's a substantial investment and so you can't afford that. Again, the spikes are different, but at the same point in time, there is a certain amount of, well, how much computer are we going to need next year? If we were able to source 70 to 80% of it at a 50% markdown, this makes sense. We still leave ourselves available for spikes or new projects as they're ramping up, or the things that we don't, the unknowns or maybe things we're going to sunset, we put out there with the idea that it's going to make it easier to turn it off, but there is a certain amount of, you could do the same thing about a power standpoint of, I could forecast how much compute at a minimal I'm going to need.

I'm curious to see, I'm with you to see how people move forward with this. Do you think part of the cloud adoption came down to just the tooling from a DevOps developer empowerment? To your point of, the real barrier to your own infrastructure is simply infrastructure, getting created, resources, that would take months, weeks at a minimum, month or two. There was always the, it's the same thing as getting space, when you're leasing space, you buy more than you need to hopefully fill it, to hopefully, that's not enough in the future. It's kind of a Goldilocks situation of like, it's never just right. Is that Goldilocks? Is that the one with the forage?

David:

Yeah. Too hot, too cold, just right.

Patrick:

Right? Yeah. It's like, there's always like, oh, I got to get the beefiest box I can, so you over-invest and then it's like, oh, I didn't really need it anyway. To your point, if it's stable and you're on year two or three of a digital product that you've been hosting, you can come up with some math. To your point of, how do you migrate it out? Are you migrating your products outside of the cloud at this point?

David:

Yes. We actually just finished. At the beginning of this year we had seven applications we were running in the cloud, and as of about a month ago, we were done bringing all those seven applications home, as we like to say, to the two data centers that we use. It was actually shocking how much easier that was than what I thought. We literally spent years getting into the cloud, rewriting applications to be containerized, and all this other setup that you have to do when you were used to servers of pets, like these nicely treated things that were manually set up and switching to a model where you treat servers as cattle where it's all the same commodities. When we had gone through that, I thought it was literally going to take us years to get out, but the fact is that most of the technology that makes cloud a pleasant experience today is open source technology, and it's the same technology you would run on your own servers.

We run VM software, something called KVM on our servers, and then we developed a small tool to help us with deployment called Kamal, and then we run on top of docker. A lot of these technologies are the same kind of technologies that the cloud will sell you, and there's not necessarily that much labor in it, as in, the cloud doesn't have a ton of secret sauce. In fact, that's one of the controversies of the hyperscalers is that they mostly just take open source projects, string a few little pieces of scripts and perhaps some UI in front of it, and then they end up reaping the majority of the value from that, not the creators of it.

rvers, and yet here we are in: 2023

We just figured out with modern insights and technology, when the reality is, do you know what? At the end of it, it's all Unix. At least for most of these things underneath. All the same processes, everything is still there. There are still people alive, they didn't die off in the past five to seven years, who know how to set these things up. People can absolutely learn it as well. I think just also having a bit of confidence that the technical underpinnings are not rocket science. In fact, if anything, it's just gotten easier and easier because servers have gotten more and more powerful. That was one of those other bits that switched for me. I thought that once the hyperscalers get to this incredible scale where they're buying hundreds of thousands of servers, they will save so much money in operational efficiencies that they can pass on those servings to us, says that there's no way we could do it cheaper ourselves and that will strike through when you look at Moore's law.

of stood still for a while in: 2010

There's just something when I look at those two charts and go like, hey, here's the cost of computing power, here's the cost of cloud computing. You know what? Those charts are going in opposite directions, that is not right, that is not how an efficient market would operate. I think some of it is also just coming because there's just such an incredible pricing and lock-in power from some of these vendors once you've moved to the cloud. At least the appearance is going to be very difficult to get out. In some cases it is difficult to get out if you start going down the rabbit hole of serverless and many of these managed services that are proprietary, that don't easily come out, they don't run on commodity open source systems. I was surprised how quickly we got out, and this is why I've been so vocal about it, both because the savings are so big and also because I just don't think it's good for the internet.

I owe my entire career, my entire business to the open, free internet. The internet is the marvelous platform where you don't have to ask anyone for permission to exist. You can literally show up, plug in an ethernet cable, have an IP address, point at the main to that IP address, boom, you're in business. That is a wonderful, unique, rare moment and sort of opportunity in human history.

We should be so foolish to give that up easily, and we're giving some of it up by essentially handing the majority of all the computers that run the internet to five big companies, which you see that time and again when Amazon's famous US-EAST-1, which is the original region that they have and the one they have the most computer power, when that goes down, seemingly like a third of the internet goes offline. That's not the system DARPA designed. DARPA designed a system that was literally resilient to nuclear warheads taking out entire cities and we could still communicate, so I think we're regressing on some of fundamentals of the internet, both in terms of the openness, the freedom, the marketplace, and the resilience, and I don't think that's good.

Patrick:

I got to say this has been an awesome interview, covered a lot of ground here. I hate to stop it, but at some point I got to ask you, would you be willing to come back on again maybe later this year?

David:

Yeah, absolutely, let's do another session another time.

Patrick:

Awesome. We've got a couple other topics I think are really important to talk about and I'd love to get your take and share your thoughts with our listeners because really great stuff. There's something about innovation in entrepreneurs that requires the contrarian thinking of like, okay, so just because everybody says so, it generally means it's not true.

David:

It certainly means there's opportunities to go the other direction. If you see everyone walking one path, you go like, ah, I wonder what's going on over there, actually.

Patrick:

That's right. It's time to zag. If everybody's zigging, it's time. Thank you so much for joining us today. Congratulations on all your success. We wish you nothing but the best in the future. We're looking forward to have you back on when you have more time, when you're not crushing races and starting businesses, and building apps, and writing books in between times, your free time.

Shelli:

Free time.

Patrick:

I'll get some of that free time you've got. Thank you so much for being on today.

Shelli:

Thanks, David.

David:

Yeah, thanks for having me on. This was great.

Patrick:

We also wanted to thank our listeners. We really appreciate everyone taking the time to join us.

Shelli:

If you'd like to receive new episodes as they're published, you can subscribe by visiting our website at dragonspears.com/podcast, or find us on iTunes, Spotify or wherever you get your podcasts.

Patrick:

This episode was sponsored by DragonSpears and produced by Dante32.

Share Episode

Shownotes

Transcripts

Follow

Chapters

Video

More from YouTube