Ep. #865 - An Overview of AWS Programs

00:00.00
Matt Watson: And we’re back for another episode of the Startup Hustle. This is your host Matt Watson, and today we are going to nerd out. We’re going to talk to Paul Duffy from Amazon Web Services about all things cloud. And how you can leverage cloud computing to help your startup. I’m very excited about that; I’m ready to nerd out today. [A] reminder that today’s episode of Startup Hustle is brought to you by FullScale.io. Helping you build a software team quickly and affordably. Paul, welcome to the show. How are you doing?
00:31.18
Paul Duffy: I’m good. Great to be joining you.
00:35.94
Matt Watson: I guess, before we get started, [we would] love to learn a little more about you and your background. I was looking at Linkedin, looks like you’ve looked at Amazon Web Services for a long time. Tell us a little bit about yourself.
00:48.60
Paul Duffy: Yes, I’ve been on and off at Amazon Web Services since 2011. I say on and off [because] I was drawn to work in the cloud largely because of what I saw it could do for startups. Having been a founder of a startup myself. So I’ve worked for AWS, say joined in 2011. I’ve left twice to go to work at startups because doing that was so high and have come back to AWS both times because the tenable platform is such an interesting thing. And it’s such an enabling thing for startups. I’ve spent most of my time in AWS doing these solution architecture types of roles. What that means for my team now is it’s a set of technical folks who deeply understand cloud platforms, have a bunch of experience and empathy with startups, and their job is basically to help. Startups are building their product to do that whilst spending as little money and as little time as possible. So it’s a pretty cool job and I really enjoy it, especially because it’s the most exciting set of customers that you could possibly imagine working with. And real tech as well.
02:02.56
Matt Watson: You’ll probably love this. I’ve used Amazon Web Services. I’ve been a software developer for 20 years. I’ve used Amazon Web Services. But I’ve actually used Microsoft Azure way more for the last ten years since Microsoft Azure first launched. But I am very familiar with AWS as well. Like yourself, I’m a huge proponent of all things cloud and the benefits that it has are absolutely enormous. I absolutely love that AWS provides a lot of benefits to startups. Companies are trying to move to the cloud too. Is that something you can talk about as well?
02:46.43
Paul Duffy: My focus is maniacal day today with startups only. Today, you look at the range of different companies that use the cloud. And you find everybody from stereotypical two entrepreneurs in a garage or a coffee shop all the way to large enterprise customers. When I started my career, my first job was a Saturday job in a local computer shop building computers; physically building servers.
03:24.26
Paul Duffy: When I was building my startup, and this was probably about 2007-2008, that was still the time where we had to spend a whole bunch of money on servers and a server rack and a generator. And all of this kind of stuff.
03:38.27
Matt Watson: Everything except your actual product and service.
03:43.70
Paul Duffy: It still pains me to this day that, unfortunately, statistically, happens with many startups. Unfortunately, we failed and we didn’t work out and it took us so much money just to get to the point where we realized that we were going to fail. Now, I don’t know whether our idea was good enough that cloud technology would have changed the outcome in that way. But the thing for these kinds of young companies, who by definition, it’s not like they have data center legacy stuff. They literally start. It could be the two of us in a cafe, in a room with an idea, and then we want to get going, and the cloud was just this kind of natural thing.
04:21.10
Paul Duffy: Startups, in the first place, where instead of having to buy all this stuff, you sign up for an account with one of the programs we offer today. [Which] activates [when] you sign up to Activate. You get credit up to 100,000 credits depending on the source of investment to get started on the cloud with no commitment. You might have to pivot your business model a bunch of times if you’re trying to find product market fit. And you haven’t had to go through this horribly painful phase of doing the physical building, sticking service in a rack, and all that kind of stuff. Looking across the last ten years or so is when I first joined. And I remember, you might think I’m saying this as an employee, but it’s not. This is before I bought the trade. I remember my first experience with the client where it was like a science fiction thing almost. I just come from this kind of data [that] has sent the environment and it’s like, wow I just launched a server. And I’m SS-ing into it and then I can destroy it a bit later. Now that was magical. How that was magical. But at that time, there was still a relatively limited amount of services available to folks and people. So if you’d have to build something on top of that, the work was definitely easier but you would still have had to do the work as technical people. Now, I think, it is such a good time to be kind of technical folks in startups because with more than 200 services [and] 200 different building blocks if you want to build something. Doing natural language processing in the old days, you had to have built all the language models and all that stuff. So a lot of customers just wouldn’t necessarily have bothered doing it because the slope was steep or they would have had to have hired the best kind of . . .
06:01.20
Matt Watson: Yeah.
06:14.37
Paul Duffy: Language processing engineers that many of them couldn’t afford. The same kind of things, like containers, became a thing. There was a time period where you would have had to have done all the orchestration. Again, for you and me, if we were those mythical founders, it’s like, if we want to hire two engineers, and we all know how hard it is to hire engineers, we have to hire those two engineers. What do we want them working on? Do we want them working on our product? Or do we want them working on what we call AWS?
06:40.53
Matt Watson: Yeah, figuring out. How do we host redis and set up high availability of that and scalability of that and backing it up and like now screw all that, right? Like I just log in to AWS and just enable that as a platform [and] as a service-type thing. I start using it an hour [and] a few minutes later, right? A great thing about being a developer in this day and age is cloud platforms. With these platforms as a service-type feature, as a developer, I’m like a kid in a candy store. It’s like queueing and caching all these different database technologies.
06:59.49
Paul Duffy: Right.
07:16.48
Matt Watson: Like you said, all the AI machine learning, all that kind of stuff. All these things are available to me. The hard part becomes knowing which ones to use and how to use them. And that’s a big part of where you can provide value to people right.
07:20.61
Paul Duffy: Right. You mentioned even more surfaces there. And I think the candy store analogy is one I love because it’s so true. Almost, it’s like the selection in the candy bar ten years ago was a few different things and now it’s like such a broad range. Making the decisions, I think this is a really good point and this is one of the things that my team spends all of their days doing. So as humans, as we learn, you know my first staff, I was sure of your first experience doing this. I made a lot of mistakes. I continue to make mistakes. The only thing I hope is I don’t repeat the same ones. And the first time you’re doing a startup, but you’ve not done it before, you don’t have the benefit of a bunch of experience yet. You haven’t seen those patterns. Yeah, and Andy Jassy, the CEO of Amazon now [and] the former CEO of AWS, he used to have this phrase that I . . .
08:14.25
Matt Watson: You don’t know what you don’t know.
08:25.83
Paul Duffy: Slightly stolen and modified. But he would say there’s no compression algorithm for experience when he would talk about running a cloud platform. Because that’s a hard thing to do. We’ve learned a lot working with customers as we build that where I think . . . can help in some ways. They can be that compression algorithm for experience with the staff.
08:42.88
Matt Watson: Exactly.
08:45.40
Paul Duffy: Every day, if you’re on your one startup, you are working on that technology stack. You are seeing the problems in your business. You are running as fast as you possibly can to get more customers, to go over technical milestones, and to get to more funding. My team gets the chance to spend time helping all of these startups on different architectural challenges day-to-day. And because they’ve seen these patterns, they’ve seen what scale looks like. So as we go from a consumer’s startup from zero to 100,000 or to a million users, they can go to the next customer. And talk about those kinds of common infrastructure patterns. It’s like, “Hey, this is how you should think about database scaling. Or it’s okay to do this thing.” And I wouldn’t quite describe it as it would be rude to describe this therapy. But, in a way, from a technical perspective, there is that kind of thing when you’re in these early stages of startup, you’re always having to make imperfect decisions and tradeoffs. If you build your platform, if you’re an incubator or accelerator . . . and you’re working towards demo day. If you get to demo day and you have the most scalable, decomposed microservices-oriented architecture that’s ready to go for a million customers, and you don’t actually have a product. It’s probably not going to be a very fun demo day for anybody involved. Certainly not from you, from the point of getting investment. So one of the things our team is trying to do is say, “Hey, we’ll help you make the pragmatic decision. Now we’re going to help you make the decisions that will help you long term.” Of course, we want to make sure security is priority zero. We want to make sure what you’ve deployed is secure. But we also want to help you understand that you don’t need to build today for the millionth user before you’ve got anybody on your platform. We’ll help you make sure you don’t go through one-way doors that’s going to leave you in a really difficult situation later on. But we can help because we’ve seen that thing before. Because we’ve been working with startups for a long time since we launched in 2006.
10:51.52
Matt Watson: And the challenge is it’s really hard when you’re first starting out building a product. Because you make some really important design decisions at the very beginning. And using the cloud is one of them. But whether or not we are going to use containers?
11:07.91
Matt Watson: Is it serverless? Are we just installing it on a VM? But some of the harder decisions probably come down to the database. What database are we using? How are we sharing the database? How do we make the database scale? Some of those little decisions that you build around later on are a big deal that really bite you in the butt, and I’ve been part of that. I’ve had two companies that were SAAS companies myself that I founded that grew rapidly. That performance and scale were paramount. Like my last company, we were processing billions of data points a day. We were spending over a million dollars on hosting in the cloud, so I understand the importance of performance and scale and architecture. To your point, it’s so hard, as a startup, to understand the decisions that you make. And you don’t want to over-engineer it, which a lot of people do, right? They over-engineer it. So it’s always a difficult balancing act. I love how you can’t compress the knowledge and experience part of it and having experts. You and your team, who can help guide startups, I think, are extremely valuable.
12:20.41
Paul Duffy: Yeah, and I think it’s funny. You know, you talk about the decisions you’ve made, and I’ve seen this; I’ve made these decisions. It’s almost like that when I was a [first-time founder] that one of my mentors said to me, “You’re gonna fail and you’ll get him off.” He didn’t say absolutely gonna fail, but he was like, you’re going to make mistakes. I was talking about your arrogance of youth, in a way. But I’ve got this partner where I’ve got these things and we’ll do it. And what he was right about is, yes, there were things that would just happen that couldn’t really be predicted. Now, with more life experience, I’ll probably avoid making some of those, and a lot of this becomes really high judgment. But there are . . . I love analogies. One of the analogies of a high-row startup is like those old video games in the 80s and 90s where you’d see the platform just falling away from under the little sprite in the game. It’s kind of like that with a lot of these companies building things where the concrete is never dry. Like it’s always dry and there’s always going to be some refactoring that’s going to be around the corner now. Obviously, no one wants to paint themselves into a corner, but there are all these kinds of traits. It’s just the world of tradeoffs where you’ve got to get to that next milestone. And how do you get that balance just right between scalability that’s not going to cost you another twenty weeks of build time. For example, knowing that you will probably have to do refactoring. And the other thing is like no business school class can teach anybody this. That is the thing that has been particularly . . . some of the business-to-consumer startups where someone comes up with an idea and this is probably why I’m not being an investor. Someone came up with an idea that if they told me they have that idea in a coffee shop, I probably would think it was crazy. Six months later, it’s like the top app in the App Store and it’s a complete craze and . . .
14:10.23
Matt Watson: Crazy how that works. I’m gonna rent out my couch to random dudes to come sleep on. Like who the hell would think . . . it’s a thing like people . . . it’s a thing, couchsurfing. It’s a thing, like it’s crazy. The thing that people pitch though.
14:15.47
Paul Duffy: Yeah.
14:20.91
Paul Duffy: It’s always like hindsight now. All these things are so obvious, but one of the things I love about startup customers is that it’s educational for you, as a human. You just learn about these different ideas.
14:37.33
Matt Watson: Yeah.
14:39.55
Paul Duffy: One startup based in Canada that was just fascinating to watch recently was one you may have or may not have heard of called Wombo. Wombo’s app, you should try it. You take a picture on your camera and it lip syncs using some AI stuff that’s built on a bunch of different pop songs. If someone told me that idea without seeing the app and experiencing it, I would raise my eyebrows. But trying it and seeing it in the dark times of a pandemic, with a lot of people being isolated, the kind of joy that it brought to a lot of people was just fascinating. It would be much easier for people to sit around and say that’s a crazy idea. How is that ever going to take off? You don’t need to worry about scale because no one’s going to use your platform. And then, overnight, it becomes this viral sensation that no business plan is ever going to predict.
15:27.73
Matt Watson: Yeah, yeah.
15:36.58
Paul Duffy: You have to be used to being like the concrete thing. You’re always going to have to be stepping and constantly moving forward versus a 20- or 30-year-old business. For example, we kind of know what every year looks like, and it’s pretty consistent, and there’s some growth. But it’s not that overnight or craziness.
15:57.15
Matt Watson: : That’s the key advantage to using platforms like AWS, right? I’m getting a lot of traffic. How do I scale this thing up? Potentially, I login to AWS and I say, “Okay, I need to go from one server to ten. I just pushed a button, and I got ten.” As a developer, I’m like, “Oh we need to do more with caching. And like ABUS can immediately turn on caching. We can write a few lines of code and start caching some stuff and handle more traffic. Or I can switch the type of database I’m using, or scale up the database I’m using.” There’s all these things at your disposal that you can do. Back in the old days, for me, there were so many servers. I got an order—one from Dell. And then, I got to do this major outage to install a new sand, and do this big VMware upgrade. It’s a giant nightmare and nobody wants to deal with that. So the cloud made much of this so much easier, like I can’t imagine ever not using the cloud.
16:52.46
Paul Duffy: It’s funny that you and I, obviously from this conversation, both actually have that experience. I remember while waiting in the parking lot of my startup’s office for the UPS driver to arrive with a surfer. But that was literally delivering now.
16:58.78
Matt Watson: Yeah.
17:10.25
Paul Duffy: Had a funny experience. This goes into the next part of serverless stuff. I want to start talking about where we had some interns in our team this last summer. We took them out to lunch, and one of the interns was chatting to me. He was saying, “What does a server look like? I’ve never seen a server. Is it rectangular or is it square?” And this gentleman had studied computer science. But for his generation, why would they ever think of using a server? It’s one of the things that’s an unusual thing. And it’s a fun thing about this job—goes back to something that Adam Slipsky—who’s our CEO now—said early on in my career where he . . .
17:48.54
Matt Watson: Yeah.
18:07.48
Paul Duffy: We want our customers to pay us the smallest amount of money possible for the services that they use. And I love the fact that me or my team can go to a customer [and say], if we can cut your bill in half, we’ll do it now. And if we can cut it even more than that, that’s what we’ll do.There are two gross ways of looking at this. Number one was the old world where our startups may have been in. When we had to buy and spend money on servers. Going to do that in the cloud was revolutionary because now, we’re paying for the hour. Now, you’re paying things for the minute or the second. But then, there’s also the next generation and the generation that will all come after that. When we launched AWS Lambda in 2014, that was kind of the next-level type of thing. Why do you even need a server? Just running this serverless architecture, you’re now paying by the milliseconds of execution time.
18:55.43
Matt Watson: Yeah.
19:03.48
Paul Duffy: One customer I worked with was called STEADI. It’s a really interesting startup that’s disrupting EDI, hence the name. The founder there was just quite a visionary in how he thought about doing serverless stuff. Because for him, if I ever have to deploy a server, forget for a second whether it’s in the cloud or not. Why would I do that? Why would I not rather go for this serverless stuff whenever I can? And I think, when you look at some of the largest startups today, sometimes we get younger startups.
19:30.64
Matt Watson: Yeah.
19:40.42
Paul Duffy: We want to talk to these people who did it three years ago, four years ago. We want to know how they did it. And sometimes, if you go and talk to those more mature startups, they’re like, “You shouldn’t listen to us because what we did three years ago, we would not do now. Because what’s available to us is very different.”
19:54.38
Matt Watson: Yeah, absolutely.
19:59.72
Paul Duffy: The thing that my team . . . part of their job is to [develop] state-of-the-art architecture. As we get the chance to start with a lot of these companies, who are literally at the earliest possible stages, it’s like we want you to do the minimum amount of work possible. So if we can get you using Lambda, get you using API gateway, get you using containerized stuff and as little infrastructure as possible and spending the smallest amount of money, that is so satisfying. The other thing is . . . if you’re a startup, you don’t have that legacy. You don’t have a whole team of people who are like, “Oh wow, we’re going to have to get out of the data centers due to the things I’m used to. What about my networking team who knows how to configure the CISCO routers? It’s just refreshing for those people. Like the guy I just talked about, he’s coming out at university. He’ll do a startup, maybe at some point . . .
20:53.39
Matt Watson: Yeah. So my last company, we spent over $1,000,000 a year on cloud hosting. We didn’t have a single server, admin system, and admin employee—not one. It was all in the cloud.
20:56.95
Paul Duffy: In his mind.
21:09.63
Matt Watson: To your point, our focus is always deploying the apps, right? We want to use services like AWS Elastic Beanstalk or Fargate or all these types of services where we’re just deploying our app and we don’t care about the servers at all. We just tell AWS we need seven instances of these things, and there are small servers or large servers [to] make it happen. However, it happens. That’s the beauty of it, and we don’t have to deal with the infrastructure—any of it. So it’s a very wonderful thing. I do want to take a minute to remind everybody that today’s episode of Startup Hustle is sponsored by FullScale.io. Helping you build a software team quickly and affordably. I also want to give a huge shout out to AWS because I really appreciate the programs you guys have for startups. And here, in Kansas City, I know you guys have been to several events that I’ve been to and given away huge amounts of credits to local startups here. Big thank you for that. I know, as somebody who spent like millions of dollars on cloud hosting before, and we were an anomaly, by the way. I mean, most people probably spend hundreds or thousands of dollars a month. We were a massive scale. We were an anomaly but the money that you guys provide people is huge. And for a lot of people, honestly, the credits could probably last them one, two, three, or four years. It’s amazing. How inexpensive using the cloud can be? And the credits that you guys provide can go a very long way.
22:40.41
Paul Duffy: Yeah. That’s one of the . . . it sounds unexpected to some people, but one of the most important things the team does is help people spend those credits as slowly as possible. If we can get it, we do that on your program thing.
22:40.70
Matt Watson: I really appreciate what you guys do for the start of community.
22:54.48
Matt Watson: Sure.
22:59.97
Paul Duffy: The focus on startups for us isn’t part-time. So everybody on my team, the broader team of account management people, [and] business development people are working in the startup community. It’s a full-time job. It’s not like in the few hours of the day that we have spare where we’re not working with enterprises. Everyone on this team kind of has a startup in their title because that’s what they do. Beyond the community thing, we have a bunch of programs. One thing that we just launched a couple of weeks ago in April was the IMPACT Accelerator. And that is a program that we’re investing $30,000,000 for startups led by under-represented founders. So Black, Latino, LGBTQI, and women founders. For those startups, not only they’re getting up to $100,000 in service credits for the platform. They’re also getting $125,000 in cash, and a whole kind of set of all the things that help them, [such as] business guidance [and] technical guidance.
23:51.72
Matt Watson: Wow.
23:57.70
Paul Duffy: Peer community. But we’re gonna take them to Seattle where they build a customized agenda for the stuff that they want to get out of the Accelerator. They’ll repeat that in New York or pay for the members who are going to do that kind of thing by 2006, [which is] totally changing the game. In the old days, you had to have money. If you didn’t have access to capital, the network, and that kind of stuff, you would protect . . .
24:21.97
Matt Watson: Massive amount of capital. Back in the day, it could be [that] you need tens of thousands of dollars just to buy the infrastructure and servers that you may need, right? Like, “Oh, I need SAND, and I need multiple servers because I need high availability. So I need more than one.” If you’re trying to do things the right way, it was very expensive buying the network equipment, racking it, stacking it, and configuring it all. The software licenses and everything else, it was crazy.
24:44.10
Paul Duffy: I still have the scars. And that’s the thing that in 2006, we felt the game was really changed for startups. Because now, literally, it’s like sign up for an account, get some credits, and you can get going. It’s like the little dog with a megaphone that makes it sound like a deep-voiced, bigger dog. The cloud enables those folks to do that because they’re getting the same new, high-quality infrastructure that anybody, like Amazon scale, could build and we’ve helped you.
25:20.14
Matt Watson: Yeah.
25:22.62
Paul Duffy: Talk about eBay, Stripe, Snowflake—all of these companies that have been built in that way. And this program is just part of continuing to level that playing field. So that founders who always have the most interesting ideas can grow those businesses regardless of gender, sexual orientation, ethnicity, or race. So we’ve had our landmark kind of Activate program that’s available to a very broad range of startups that gives up to $100,000 credits. Whether you’re a bootstrapped entrepreneur, or you’ve got investment through an accelerator or an incubator, or a VC firm.
25:58.17
Matt Watson: So can anybody apply for AWS Activate? That’s what it’s called, right?
25:59.73
Paul Duffy: With this but second.
26:03.49
Paul Duffy: Yeah, so Activate is the program that gives people access to different things. All those credits. It’s got a console experience that gives people technical templates to help them get started as quickly as they can and to point you. We’re talking about all the decisions they have to make [and] make it simple. It gives them access to third-party offers and there are a few different configurations. So that headline amount, it’s up to $100,000. Not everybody gets that but our idea is that if you and I were bootstrap founders, we can apply to this kind of the founders tier.
26:31.50
Matt Watson: Sure.
26:41.33
Paul Duffy: Without having to have external investment. And then, as our idea gets traction, we get investments from the kinds of community of investors that get access to more benefits with Activate. What we’re specifically doing with the IMPACT Accelerator, which is this thing, will be launched in April. Meaningful $30,000,000 commitment for startups by under-represented founders. Which not only does it get those people in Activate to try and level the playing field and give them the credits. But it gives them cash and it gives them a whole host of other benefits too. No doubt that we’ll continue to kind of build on the next up.
27:14.16
Matt Watson: So is that more of a competition? Or how does that work for these companies that want to apply for the impact part of this?
27:23.35
Paul Duffy: No, I wouldn’t describe it as a competition. We’re calling it the [AWS] IMPACT Accelerator. Yes, folks have to apply. So applications are open for the first round that will kick off in June, and I will have 25 US-based startups. Those folks submit their applications. There’s a pretty diverse and well-represented panel that will look into that. But after that point, we are there to basically try and accelerate those. So, for example, they’ll be in Seattle depending on what their startup is. It could be B2C. It could be B2B. It could be that they’ve got really hard technical problems. It could be that they’ve got distribution problems. You name it, and they’ll be able to kind of work out this customized channel curriculum. So we will be sitting there to help them along their path and to accelerate that sort . . . We’re investing deeply to level that playing field and to surround those folks with assistance. It means that all of the technical challenges [are] minimized as much. My team will be helping out office business development people. We’ve got a whole bunch of SS and mentors of every kind of description you could think of to help. And that’s just an additional film that we’re doing in the community because startups are so important.
28:55.76
Matt Watson: Well, it’s fantastic that you guys not only have the Activate program, but also have this IMPACT program. Like I said, somebody like myself, who spent a ton of money on cloud hosting . . . My company was able to take advantage of some of this type of incentive stuff before. [It] was a huge deal. When you’re first starting out, and you don’t have a lot of money, or maybe you’ve raised a small seed round from friends and family. At least knowing that you don’t have to spend maybe a few thousand dollars on hosting over the next few months as you’re trying to get this off the ground is a big deal. It’s a huge deal. So thank you guys so much for doing that. I know, once people start using your service, they [are likely to] never go away. So I know this is a great thing for you guys. You guys have been the leader in the cloud from the very beginning.
29:51.97
Paul Duffy: We’re glad it was helpful. We’ve got these leadership principles and customer—is that the kind of number one that comes to mind for folks, which are working backwards from our customers to do the right thing. Yeah, we want to build these long-term relationships and part of our . . .
30:08.98
Matt Watson: Yeah.
30:11.80
Paul Duffy: Our mode is like, if we have a choice between you spending $10 and $5, we’ll always make it the $5. If we can reduce it lower because we can improve your architecture. That is what we do. That’s kind of a fun thing to be able to go and do with customers. It’s like how can we cut your bill in half or a quarter?
30:28.98
Matt Watson: And I’ve been on the other side of that. Where you accidentally enable something and then you get a bill for like $10,000. I’ve been on that side of it, too. Whereas, as a developer, you have all these tools available to you. And you’re like . . .
30:30.99
Paul Duffy: Or stop for . . .
30:44.65
Matt Watson: I’m going to enable this extra logging or this extra database tier. Or scale up or whatever, and then you forget about it and you’re like oof.
30:48.75
Paul Duffy: Well, it’s kind of funny. That’s one of the things that we are focused on doing at the meeting. And working with these customers as early as we can, whether that’s at a scale thing later on. So, for example, when you start with Activate, a couple of things that we want to tell you to do. For example, one might be enabling multi-factor authentication on your account to make sure you’re starting off with the best security posture possible. Then, the second thing is setting up a budget, so that you’ll get an alarm.
31:12.54
Matt Watson: Yeah, yeah.
31:25.23
Paul Duffy: Those are always the conversations and someone’s like, “Oh, of course, I want to set this thing up. I’ll do that.” And we’re just trying to, almost like, help people eat healthily from the start. It’s like let’s do these two things first. Let’s teach you about these most important things, so we don’t have folks having surprises.
31:45.12
Matt Watson: What keeps customers happy? The last thing you want is for people to get a surprise or feel like, “Oh man, we’re paying so much money on this thing. Maybe I should look around.” You want people to be happy. And we use AWS at Full Scale, at our company. I’ve used the cost management tools, a bunch of times, and reservations to save us money and all that stuff. [I’m] glad you guys do all that . . . have great tools. So I really appreciate it. So how else can startups take advantage of AWS?
32:10.44
Paul Duffy: Good for you.
32:22.47
Matt Watson: You know, all the different things that you guys have. Or are there other side things besides Activate and the IMPACT program? Or are there other things that people should know about?
32:32.66
Paul Duffy: Yeah, I mean, as well as the IMPACT Accelerator and Activate. We also have certain specific accelerators. We have one for space. For example, we have one for Sustainable Cities. The best way of describing them is that they’re kind of trying to cluster customers together who have similar use cases, so that they can be put in a room of people who are experts. Those are a couple of other ways to connect with us. We’ve got a couple of physical AWS lofts that we call Startup Lofts, which are right now, we’ve got two in the U.S. One of them is in New York, but one of them is in San Francisco. They are kind of spaces for startup customers to come and do a whole variety of things. Come and get information. Meet people from my team at some coworking events. And I think, something that a lot of people have missed over the past few years, just a chance to network with folks and have representation. All of that kind of the AWS summits, so these are free two-day events that we do in a bunch of different places. We just did one in San Francisco a couple of weeks ago. We’ve got one in Atlanta coming up. We do them in New York, Toronto . . . all the kinds of cities where you have a population of startups. And they are free events where you can get everything. From networking to technical sessions in a 1-to-1 where you can ask experts. But Activate is the biggest single program. Anybody who is a startup who hasn’t joined Activate . . .
34:06.56
Paul Duffy: We say that you please join. That’s our way to be involved with you from a programmatic scale point of view. We have a virtual version of a loft that’s all about programming for startup. Specific stuff, like how you can reduce your bill and be as cost-optimized as possible. How [can you utilize] serverless architectures? What should you think about containers? All of that stuff. Some customers might have a chance to interact with the team directly on a kind of one-to-one, in-person basis. Some people might not have the time to do that. So we’re trying to deliver all of that stuff to customers at the place and time [that] makes sense for them. So yeah, Activate definitely being the kind of the lead program that . . .
34:53.81
Matt Watson: So how do you see the future of the cloud and AWS as services, and how people are using them today? Do you see serverless as being the future? Or was that a wave and now, maybe things are going a little bit back to containerization? What do you see in the future for developers and how they deploy their apps?
35:18.62
Paul Duffy: I don’t know. I don’t have a superb crystal ball that’s going to definitively answer that. I think that serverless is definitely not a wave. I was talking to one of the leaders in the serverless organization who came to talk to my team the other day. And we were reflecting that when we launched Lambda, the term serverless wasn’t really an existing thing at that point. And we launched a couple of different services at that event. We launched Aurora, which was MySQL compatible. Super high-performance, low-cost database.
35:53.19
Matt Watson: Ah.
35:56.24
Paul Duffy: We launched the ET2 container service for container orchestration. And we launched . . . now Aurora, [which was] kind of an obvious thing to people. It’s like, okay it’s a relational database thing. It’s got awesome storage and reliability performance and price characteristics. I get that EC2 container service made sense because containers were becoming and continued to become popular with that. But there’s ECS, whether it’s Kubernetes and powered by Fargate, underneath Lambda, was an interesting one because I think it wasn’t quite entering that sort of defined space.
36:29.95
Matt Watson: It was sort of an experiment at that point, right?
36:34.51
Paul Duffy: And now, we’d work back from customers because customers were like, “Okay, we don’t want to run servers in a physical data center.” Okay, cool. But we’ve given the ability to do it in the cloud. Yeah, we want persistent storage with different performance levels. Okay, we’ve improved DBs with . . . added all of this with Lambda.
36:47.41
Matt Watson: Is it Leather?
36:53.35
Paul Duffy: In some ways, it was responding to this thing. Customers were like, “No one ever grew up thinking [that] I want to run a server kind of thing.” That was not what they wanted to do. So it was very interesting that, after we launched Lambda, you’ve seen the reaction from customers . . . [it] was kind of awesome and these new serverless architectures . . .
37:10.67
Paul Duffy: Started to become more and more popular. So I definitely don’t think it’s a wave. I also don’t think that things switch overnight in a business as big as the cloud. Because we don’t just have stuff for customers running in the cloud. We’ve got enterprise customers. We’ve got people and migrating mainframes to the cloud. I think for startup customers, if you were starting on day one, you have this beautiful, clean sheet where you don’t have any legacy. So you could start with serverless, and I think my team, when they’re advising customers, it’s like if that’s the way that you can start and it will meet your needs . . . Probably just going to see more abstraction and less of the low-level heavy lifting. And you can even see that, like since 2006, since I started in 2011 to now. If I started my startup in 2011, it’s like wow, EC2s, 3RDs, all these kinds of things. If I was starting my startup, and if any of my team are listening to this, I’m not planning to quit again and do a startup. But part of me is just so excited by that. Because the candy store analogy that you give [has] all of this higher level abstraction.
38:20.97
Matt Watson: Yeah, right.
38:26.41
Paul Duffy: If you go to talk to us . . . if you guys talk to a startup that was founded in like 2011-2012, open the hatch, look at some of their tech stack, and you look at some of the stuff they’ve had to build, they would not build that now if they were starting today. Didn’t really have any option at the time. The same way that you went 10 years past, and you talked to startups who built data centers who didn’t really want to do it, they didn’t have a choice. So I don’t have a super fine, great answer for predicting exactly what the feature is. Or maybe, I’d be a startup investor, but I think we’re just going to this.
39:01.61
Matt Watson: So much about software development these days feels almost more like assembly. It’s like all these things exist out there, and I’m just putting them together to build a solution, right? Like the different building blocks.
39:02.39
Paul Duffy: Tide is going to continue to rise and make it easier for people to do lots of that low-level stuff.
39:21.16
Matt Watson: We’re using all these open-source packages, frameworks, and tools in the cloud to assemble our product the way we need it. And AWS is a big part of that. And the struggle, sometimes, is trying to figure out which pieces to use, right? Even from a database perspective, you have RDS, DynamoDB, Redshift, MemoryDB, DocumentDB. And, of course, PostgreSQL and Oracle and all these things. The only problem is the developer, sometimes, [has to figure] out how to navigate this. What product do you use? But the great thing is [that] you guys provide all of them.
39:50.48
Paul Duffy: Life.
39:58.44
Paul Duffy: Right? My team works very hard on that to be . . . I don’t want to use the word concierge because it sounds like it simplifies the job they do. But their job is, a lot of the time, like you look at [the tools] and we try to do some scale with it.
39:59.36
Matt Watson: Which one do I use?
40:16.39
Paul Duffy: Look at the interactions that work the best. Sometimes, when a solution architect is in a room with a startup customer, with a whiteboard, and it’s like, tell me about the things you try to achieve. And then, they’re like, this is the way that I think it’s pragmatically the thing to do for you now.
40:25.19
Matt Watson: Yeah.
40:35.39
Paul Duffy: [We] always like to make sure [that] we give customers a really broad choice. It’s almost like the compounding pharmacy where the pharmacist can walk around and kind of choose what they want to mix together into the treatments. We want to give those people who are [not] really experts in these different areas a lot of laterality to make the choice that they want to make. But then a lot of the work that we do with solution architecture stuff [is] that we publish repeatable architectures for customers to accelerate them. So we certainly don’t want customers kind of coming into the store and being overwhelmed and spilling the whales, especially for the startups. We don’t want those startups to go down this path and then six weeks later, they’re like, “Oh, wow. I was not planning to refactor, but I just hadn’t thought about this and I jumped.”
41:18.14
Matt Watson: Yeah.
41:29.67
Paul Duffy: My team’s job is to get in . . . [so] we can physically be in the room with them. Sometimes it’s on a video call. Sometimes it’s going to be some kind of scale thing through the Activate program . . .
41:33.82
Matt Watson: Yeah.
41:48.23
Paul Duffy: That says, “Hey, look. We’ve seen this pattern before. Let us help you. We want you to do this. You [can make it] better, faster, cheaper, [and] more secure.” That’s the kind of [how we] teach [them] to do.
41:55.64
Matt Watson: That’s one of the problems as developers that we have, right? We chase shiny objects and we’re like okay, we’re going to use AWS. I’m going to try all 27,000 different services they have and become an expert in all of them. And build this beautiful DevOps pipeline that does all this stuff and whatever.
42:13.66
Paul Duffy: One of my favorite stories about this is a devtool startup. We have a lot of container orchestration to do, and our product wasn’t meeting the needs of customers. And we were not getting closer to product market fit.
42:13.97
Matt Watson: It’s like, “Hey, dude. Maybe you should just deploy a drop.”
42:32.13
Paul Duffy: One of our lead engineers, who’s a great individual, wanted to build this kind of beautiful container orchestration system. We could have probably used ECS or something else instead. But the thing is that it kind of made us have to go and have the unpleasant conversations with customers who thought our product wasn’t ready.
43:04.56
Matt Watson: Yeah.
43:09.58
Paul Duffy: Nerding out with those . . .
43:11.93
Matt Watson: Well, once again, a big thank you to AWS and everything you guys do and the AWS Activate. Thank you to Full Sale for sponsoring today’s episode. Helping you build software teams quickly and affordably. Please join the Startup Hustle chat on Facebook and join us for some enlightening entrepreneurial conversations. Matt, of course, he’s always stirring the pot there and starting some fun conversations. So please join us. Paul, [I] really enjoy having you on the podcast today. Again, big thank you to what you guys do for startups. So we talked a lot about AWS Activate. What’s the best way to find that? Just search the internet for AWS Activate, I’m going to guess.
43:59.90
Paul Duffy: Search the internet. Head over to http://aws.amazon.com/activate. Sign up and find it all there.
44:06.79
Matt Watson: And then, you guys have this other special thing you’re doing right now. Remind us about that again as well.
44:13.55
Paul Duffy: That’s the IMPACT Accelerator. That’s the initiative for under-represented founders and you’ll see a link to that if you go to http://aws.amazon.com/startups. We’d encourage anybody—anyone who’s qualified [or] interested to apply.
44:23.83
Matt Watson: Okay, perfect.
44:33.00
Matt Watson: All right. Well, thank you guys so much for being on today. And thank you so much for what you guys do for the cloud and startups. Been a big benefit to me in my career, and like I said, we use AWS at Full Scale and in another company that I work at. We use AWS every day.
44:33.13
Paul Duffy: Like to get into that.
44:52.55
Matt Watson: Big thank you for what you guys do. All right. Take care.
44:52.63
Paul Duffy: Love it. We’re glad to help and yeah, really enjoy talking to you.