Werner Vogels’ re:Invent Keynote Pulled us Into his Matrix Vision of The Future (Live from Las Vegas with StackHawk)

Transcript

Rahul:

Hilary, it’s a last day of Reinvent and the keynote that literally everyone’s been waiting for.

Hilary:

I know. I want to introduce him by dipping into his LinkedIn profile, which is one of the best in history. 22 core competencies he’s listed, ranging from engineer to architect and scientist to troublemaker, revisionist, mentor, musician, and biker.

Rahul:

Don’t forget the programmer, audit analyst, or most importantly, the CTO.

Hilary:

That’s what we’re talking about. It is the final sentence though of his LinkedIn about that really sums him up. It’s written as an addendum and I quote, “I don’t deal well with bullshit and incompetence.”

Rahul:

Classic Werner Vogels.

Hilary:

Right. The Dutch Master himself.

Speaker 3:

Please welcome the Vice President and Chief Technology Officer of Amazon.com, Dr. Werner Vogels.

Werner Vogels:

I often say the world is acing…

Hilary:

Live from Reinvent in Las Vegas, this is AWS Insiders, an original podcast by Cloud Fix about the services, patterns and future of cloud computing at AWS. Cloud Fix is a tool that finds and implements 100% safe, AWS recommended cost savings. That’s fixes not just analytics. I’m Hilary Doyle, joined as always by Raul Subramanium and coming up, our special guest, Dan Hopkins, VP of Engineering at Stack Hawk. It is a packed show.

Rahul, day four at Reinvent and we finally got what we’ve come for. Verner Vogels, Amazon’s VP and CTO, reminded people why they’re here. He opened with a matrix takeoff where the red pill transported him into a synchronous world. Spoiler alert, he does not dig synchronous. It’s worth the watch when it hits YouTube. I won’t ruin it, but I will say it was a smooth teaser for the VP of Unreal Engine who walked in later in the keynote to talk about the Matrix Awakens build, among other things. It was also a great teaser for Vogels’ tongue in cheek suggestion that we’re all living in a simulation.

Rahul:

Of course. And the music and the outfits, Hilary.

Hilary:

I know.

Rahul:

They’re always memorable. And he hit the stage in sunglasses, a Morpheus coat, a lambda t-shirt.

Hilary:

That one got me.

Rahul:

So apt. And of course, bespoke clogs.

Hilary:

We’ll discuss his feet later.

Rahul:

And I am so glad that at the end of this show, he just reinvigorated all the technologists out here with his message and his storytelling and he just captivated the room. I’m just so impressed.

Hilary:

I have to say, it was actually quite lovely to watch you watching him because you were alive through that entire two hours. The macro themes he was talking about, learn from the principles of nature. Nature is asynchronous. It’s built on patterns and it’s constantly evolving. So what does this mean for a dev team? Well, focus on event driven architecture and asynchrony so you can continue to evolve and expand multidimensional projects. Basically, respect John Gaul’s law that all complex systems evolve from a simpler system that actually works. So start there.

There were a lot of announcements. I have no idea if any of them were novel. I just know that I want to listen to everything that that man says. So Rahul, how should we parse the last two hours?

Rahul:

I think Werner did an amazing job of nailing the path that AWS has pioneered and making it very clear to developers that they have to think differently when it comes to the cloud. Just restating that message over and over again at every moment and this one, yet again. The cloud is all about event driven architecture and if you’re not leveraging it, you are basically not leveraging the cloud at all. But at the same time, he laid the path for the future. He’s already thinking about what you’re going to be building for the next 10 to 20 years and giving you a treasure map that says, “Here are all the things you want to go look at and go find your own path through it.”

Hilary:

Vogels’ takeaway was simulate everything. So he said simulation plays a critical role in innovation and then he walked through the how. Has he given you a taste of that Amazon opinion you’ve been waiting on?

Rahul:

Not really, but he’s given us the Amazon version of it.

Hilary:

Yeah, sure. Okay.

Rahul:

So what he basically said is, simulation is really key to be able to really understand complex scenarios. And today we live in an incredibly complex world with an enormous number of variables. And the only way around predicting the right outcomes or potential outcomes is to be able to simulate them. And it’s a very balanced tactic in looking forward, while at the same time looking backwards or at the present by saying, “To be able to do this thing in the future, you need to go back to an event driven architecture. You need to look at everything as events so that you can simulate and figure out what the outcomes are going to be without actually living through years of those executions to achieve those outcomes.”

Hilary:

So interesting that his storytelling made the difference today because in a week of mild announcements, he had the same heavy lift and yet you were captivated from start to finish. He just showed you how to put the pieces together.

Rahul:

Correct. And I think I’ll go back to a question that you asked earlier about being opinionated. I think if you heard the talk, I don’t think there’s anyone who would’ve walked away saying, “Event driven isn’t what Werner Vogels was talking about.”

Hilary:

Right.

Rahul:

I mean, he was so clear that no other architecture is going to simulate reality or be close to what you really want. Event driven is the only way to do things, which will scale, which will be reliable, and be real to solve real problems.

Hilary:

Absolutely. Nature is asynchronous. He talked about this break between the physical world and the digital world becoming less pervasive. There’s more of an integration between the two. Okay. Werner speaks to developers and engineers, so we have invited another engineer to join you up here, Vogels style. Let’s give the man his walkup music.

Formerly of Splunk, Dan Hopkins is currently VP of engineering at Stack Hawk. Stack Hawk is a DAST and API security testing tool. Dan, we are delighted to have you.

Dan:

I’m delighted to be here.

Hilary:

That was a long walk up from the couch. It’s literally three meters less, like a meter and a half away, but somehow it’s hard to maneuver around this [inaudible 00:06:31].

Dan:

I guess I’m the guy that stands outside the batter’s box and really gets ready to get in.

Hilary:

All right. Enough with the fun, we have business to discuss. Okay. I want to understand practically how any of the announcements from today are going to affect your day to day?

Dan:

I think, there’s definitely a theme here. He talked a lot about event driven, and you all were talking about that a few minutes ago, but I think more than that, what I took away from it was a real focus on engineers and a real focus on productivity. And I think that that’s so important with everything that’s happening with the economy, that we’re making sure that the engineers that you do have are able to work as high in the stack as possible.

Rahul:

It’s really interesting that you say that, Dan, especially coming from security company, and I know you haven’t been here for very long, but engineers usually tend to stay as low in the stack as possible because if it isn’t built here scenario, they feel like they haven’t contributed enough. And therefore, you find such a proliferation of software that is written and built from scratch across all layers of the stack. What message do you have for engineers and what have you seen, especially in the security space, where there’s such a focus to get as granular as possible into the stack and build as low as possible?

Dan:

Absolutely. I think engineers have to trust what they’re building on top of, and a lot of the times they feel like they’re building on top of sand, and so therefore, they think they have to go to a lower level. But certainly, unless you’re Apple, you’re not building your own chips, you’re not writing assembler code, you’re not writing a kernel. So there’s many parts of the stack that people trust, and I think it’s about trusting higher and higher and higher into the stack. And if you look at engineers at big tech companies where they’ve built a lot of the pieces that help their engineers be productive, you see that those engineers spend all of their time working at the business level, and they’re highly, highly productive for that reason.

Hilary:

We had a conversation earlier about timing and engineer flows. Would you tell our listeners a little bit about how you’ve been taking in this conference?

Dan:

I think talking about the pandemic and the impact that that’s had on innovation, I think in 2020 when the pandemic started and teams got broken up and we went to Zoom, a lot of the innovation slowed down when we were on just doing meetings online. Now, I think that there were a lot of projects that were in flight, and those projects had been planned, but they got fulfilled over the next couple of years. And so there was some innovation in 2021 that came out. But I think that what we’re seeing a lot in 2022 is that a lot of those projects that were planned on whiteboards as groups have now been exhausted. And so there’s a little bit more iteration happening now instead of step functions where things are actually increasing significantly.

I think you see this a lot in iOS or the iPhone 14, for example, being very similar to iPhone 13. And I think you see a lot of this even in enterprise software and developer tools and things like that. But I also think, with that said, a lot of the tools have pivoted more towards what engineers want. One of the things we always talk about is how you need to have really good tools. You need to spend time sharpening your saw before you go and cut down a tree. And I think that over the last year when people were a little bit more isolated, they were a little bit more self-directed, and therefore a lot of the things that have been created are more engineering focused.

Rahul:

If I were to ask you a question, how do you see developer workflows change over the next 10 years? If you could look into the future, how do you see that happening? Are they going to be writing a lot of code? Are they going to be stitching together a bunch of these services to achieve outcomes? Or are we just going to be talking to an Alexa device and telling it what to do?

Dan:

I am really, really bullish on the way that GPT-3, GPT-4 and co-pilot the way that AI is going to help engineers to develop code. I think that’s going to be huge. I would say that was my biggest surprise from watching this keynote is that with so much conversation about engineering and productivity, there’s very little conversation about artificial intelligence.

Hilary:

Why do you think that is? Because we’ve been waiting for that all week long. Is this still the 2020 problem?

Dan:

It could be. It could be that it’s really happening this year. GPT-4 that has just come out really seems like it’s starting to crack the code. I’ve been watching demos of open AIs chatbot and the way that we’re bringing that to design, bringing it to writing content the way that it’s now copilot is bringing it to code. It seems to be happening in 2022. I’d be very interested to see what’s happening in 2023 around AI in AWS.

Hilary:

You work in dynamic app security testing. That sounds incredibly difficult. You’re in real time just trying to break into people’s systems. Is there an application for 3D or AR in DAST?

Dan:

Probably not 3D or AR, but I think when it comes to securing your code, one of the biggest challenges, what every security customer comes and tells us is, “Hey, I really want to empower engineers.” There’s this concept of shifting left that’s happening right now. 10 years ago, we went through a transformation called DevOps. We removed operational responsibilities from a dedicated team into the development engineering org. The reason we did that is multiple reasons, but primarily it’s because there is one operations engineer for every 10 engineers. It didn’t make any sense to have operations teams doing that. We’re going through a similar transformation today with DevSecOps, where we’re moving responsibilities out of security and into the engineering team now because there’s only one security engineer for every hundred engineers. This is a very big problem. And I think in order to do that, you need to empower engineers to actually be able to secure their code.

Rahul:

But isn’t that the big challenge of organizational inertia where getting development teams to adopt those standard tools that’ll allow you to implement all of these in the developer workflow are really not adopted? So how do you see changing developers’ behavior, especially when it comes to something as important to security?

Dan:

It’s a really good point. There’s a whole debate about Emacs versus Vim that you can get into.

Rahul:

Still?

Dan:

Yes. Still. Of course.

Rahul:

That debate was raging 30 years ago and it’s still raging. It’s still a fight today.

Dan:

Silicon Valley popularized the tabs versus spaces.

Rahul:

Correct.

Dan:

Which was an amazing conversation with my wife explaining why that’s funny. But no, I completely agree. Every carpenter wants to have their own hammer. So I think it’s about figuring out the right layer in the stack to come in and say, “You might have opinions about the way that we do security, but if it’s done the same, do you really care?” And there’s organizational change that’s a part of that. But if you can get through the organizational change and everybody’s doing this, and if you do it on the paved road, then you don’t have to care about compliance, for example. You don’t have to care about doing security reviews. If you have 40 hours worth of work a week or 50 hours worth of work and you’re spending five hours a week doing XYZ task, we’re going to take that off your plate and you’re going to get back to doing what you really love, which is building software, shivving it to production, making customers happy. That’s what engineers want to be doing as much of their time as possible.

Hilary:

Dan, thanks for giving us some of your time. This has been a wonderful way to add layers to what we just heard. Real pleasure to have you.

Dan:

Thanks for having me.

Rahul:

It’s been a pleasure, Dan. Thank you so much for coming over.

Hilary:

Dan Hopkins, everybody. What a pleasure he was to speak with. Rahul, after yesterday’s keynote, one of our brilliant producers called this year’s Reinvent Reiterate, and he wondered if maybe in its current form, it was time for Reinvent to retire. This morning, has really helped me understand the importance of this kind of experience. Can you give us your top takeaways from the week?

Rahul:

Okay. I think saying reiterate is a little too harsh, but here are my takeaways.

Hilary:

He’s very crazy. He cuts deep.

Rahul:

So number one, I think it’s really important to understand that the only way to get through this uncertainty that’s on our heads right now is to innovate through it. I think if status quo is no longer an option, it is, I think there was one statement yesterday that said evolve or die.

Hilary:

That was today. That was Werner.

Rahul:

Actually that was Werner today. Man, I’m dilating.

Hilary:

Time has no meaning.

Rahul:

Dilating for me. But evolve or die is I think the perfect summarization of how we should be looking at the next year or so ahead. We have to evolve. We have to innovate, and that’s going to be key. The second takeaway from the last three or four days, I forget how many, but the second takeaway is that data that you have is going to be key to that innovation. So get whatever data you have, put it in the cloud, use all the tools over there to create new innovation because that’s going to help you survive this uncertainty that’s going forward. The third one is you need a lot of governance and structure to make sure that you’re being as efficient as possible. It is a big part of Swami’s keynote yesterday, a big highlight, and I think that governance and structure and standardized way of doing some things without constantly copying and pasting and data and creating shadow IT, I think that is really key to being as efficient as possible.

And lastly, I think the doubling down on the event driven architecture today from Werner’s keynote is critical because that blueprint really serves you well. The event driven architecture is key to making sure that you can evolve slowly and [inaudible 00:17:01]. Nobody knows what you need to build for a year down the line, or two years down the line. You want to build something small and keep evolving and growing it. And the best way to do that without completely reinventing yourself every two months is by using event driven architecture. So I think balanced at the end of four days of this, I walk away with this clear story and this clear path, and I think overall, I’d still give AWS nine out of 10 for an amazing event that they set up, and I think developers will have a lot to think about and go back as they redesign and rearchitect their solutions.

Hilary:

Talk to me a little bit about that personally. How are you going back home and approaching your work differently after these four days?

Rahul:

Oh, I don’t know if you noticed this, but we had an entire team of our product managers over here through every keynote, literally taking copious notes. But I think some of these very fundamental principles, one of the things that we were just going back to is, we design a lot of distributed systems and distribute computing, but every now and then, we will make trade-offs about certain principles of distributing computing where we trade off. We say, “Okay, this is not something that we are going to see right now. Maybe when we scale to a certain point that becomes relevant.” But we never go back and review a lot of our architecture against that lens. But the fact that Werner doubled down and said, “Here are the principles. This should be your framework that you always look at.” Makes me want to go back to a lot of the things that we built and say, “Against this lens, where did we succeed? Where did we fail, reevaluate, reiterate, and build something new?”

Hilary:

That’s a really inspiring note to go out on, so we’re going to leave it there. That is it for us here on the rooftop, Rahul. This is a wrap, four days, four episodes. Thank you so much for inviting me to join you here.

Rahul:

It has been an absolute pleasure, and thank you all for listening in. And Hilary?

Hilary:

Yes.

Rahul:

Are you going to be back here next year?

Hilary:

Back here? I’ll still be living here in the bathtub of this suite.

Rahul:

Okay. That sounds really exciting. In the meantime=

Hilary:

I know. For me too,

Rahul:

Please send us your thoughts, your queries, your reactions, of course, to podcast@cloudfix.com.

Hilary:

Rate and review, five stars, as always.

Rahul:

Always. Until then-

Hilary:

Goodbye.

Rahul:

From Las Vegas.

Hilary:

From Las Vegas.

Rahul:

Awesome.

Hilary:

Terrific week.