Winglang – Elad Ben-Israel, Wing Cloud

This is techstrong tv. Hey everyone. Welcome back here to techstrong tv.

You know, I get excited when I, I find out about new companies doing exciting things, and I think you're gonna find this next story, uh, uh, very exciting one. I want to introduce you to iad Ben Israel. Iad is the CEO and co-founder for a company called Wing Cloud.

He's gonna tell us about that, but let's first welcome him. Iad, welcome to techstrong tv. Thank you, Alan.

Great to be here. It's great to have you here. Um, iad, let's start with your story.

Before we even jump into everything else, tell people a little bit about your journey and your, of course, the c e o Co-founder of Win Cloud. And, you know, let's hear what got you here today. Okay, I guess, I guess, uh, basically born and raised a programmer in the nineties, eighties and nineties.

Uh, so I'm, I'm pretty old. I've been, I've been around, uh, but, uh, really been like attracted to programming and software since, since when it was very unpopular at that time. Um, and, uh, you know, basically spent my childhood learning how to code and, uh, working a little bit and, and then doing the military service in Israel as a, you know, in the soft with software engineering.

And after I finished my Army service, I went to work for Microsoft and spent there mm-hmm. About five years, which was an amazing school for, you know, building software. And, um, uh, learned a lot there.

And after my time at Microsoft, I kind of went traveling for a couple years. And so I've been nomadic with my boyfriend at the time, and my husband now. Um, and we kind of traveled around the world and worked a little bit remotely.

And at some point, a friend of mine who moved to Amazon, uh, called me and said, uh, he's starting this new project at Amazon, and it's a, it's a secret project. And so he, you know, call we, we met and he said, uh, they're building a drone delivery system. And at that time, drones were der age.

They were kind of like this from, from from the future, right? Mm-hmm. And you know, this, this project really inspired me.

Obviously, you know, Amazon kind of delivering drones. It was a re at those years we were thinking, you know, in, in five years, we're gonna just see just drones in the sky, if you recall. Um, and so it pulled me back into corporate America, somehow big tech.

And it was actually really refreshing, to be honest. Like I felt Amazon had a very different d n a than from Microsoft. And so I was curious.

It was always, I'm, I'm always curious about how these, these amazing companies are built and how they're structured and, and how they operate. And, and the technology was really interesting as well, obviously. And so I stayed and spent their seven more years, uh, which was, uh, again, if you'd asked me before, I would never say that I'm like, uh, corporate dude, right?

Like, it definitely not my, but, uh, but Amazon was a really great place to innovate and build stuff. And, uh, I had an opportunity to, you know, the stars were aligning at some point, and I had the opportunity to start this really big open source project, uh, based on, on, uh, uh, something that I built internally. And that became, um, what, um, a project called the AWS c d k, which is, um, a, a framework for defining infrastructure using, uh, uh, programming languages.

And so that was, you know, my first step in kind like building developer tools for clouds engineers. And when I left Amazon, I really wanted to stay in this space because it's always been something that I was super passionate about, uh, because I'm a developer and, you know, I wanna build great stuff that I enjoy, right? Like, for me, it's like the easy, the easiest thing.

Um, and I, I still felt the cloud is broken in some way, right? Like, I felt like the, the promise of the cloud is amazing. You know, this ability to hand over responsibilities from my application to infrastructure services is economically makes a lot of sense, right?

Because now I have, I can focus on my business and I can focus on my users, and I don't have to focus on, you know, on, on something else. But the reality, and I think I've seen that at Amazon a little bit before, the industry is now kind of realizing it, the reality of developing those systems that leverage all these services is that you constantly have to stitch together those building blocks. And so you end up with an amazing promise, kind of like production economic promise, but horrible developer experience.

Like, horrible to the point where I'm yearning for my days with borland c plus plus in the nineties, you know, where I, where I'm able to actually like, write code and see, see how it behaves immediately on my machine, right? Like, and, and so the reality of the cloud is that every time I'm, I need to do an iteration that involves the cloud, I have to push my code into production or to staging and to wait for deployments to happen and to stitch together configuration files. And, and so you lose your flow as a developer.

And for me, developer, you know, software engineering is an art form. You know, like, you wanna be able to like focus and have a single threaded brain, right? When you're, when you're creating software and the glo and, and, and the reality of the cloud is that that's completely broken.

And we've seen in kind of recent months, maybe last past year, we've seen kind of company saying, we're leaving the cloud, the, you know, we're you, we're off serverless, we're not. There's this sentiment of like, the hype curve is over in a sense. Mm-hmm.

And, and I feel like I'm, I don't wanna give on up on the cloud, you know, I feel like it's, it's such an amazing innovation, multip, you know, like basically force multiplier for companies. And I've seen that happen at really large scale at Amazon that kind of like went in all in on the cloud and, you know, just didn't have a choice in a way. And I really wanna make sure that everybody has that opportunity and they're not giving up, right?

Like, you talk to customers and you constantly feel this like sense of, um, sense of kind of like compromise, you know, like, you know, we gotta get developers the ability to develop, so we're running around them and we're setting up those environments, and they don't touch these configuration files, or if they need to touch them, they, they just go and tweak a little bit, right? Like, you, you kind of, you constantly see the sense of like, we're, we really don't, we're, we're not happy with our environments. We're not happy with how we're developing across the industry, no matter what tool you're using.

And, and so that was, that was basically the mission, um, that I kind of, okay, let's try and fix this. You know, the c d K has been a first step in a way because the c d k, the tool, you know, the open source product that I built at Amazon basically brought in the infrastructure to the programming language world. And once it's in the programming language world, you get the ab the ability to model your infrastructure the same way you're modeling the rest of your application, but they're still very kind of divorced, right?

Like, there's still two different spaces. And, and so they don't know about each other. And a lot of the friction that we're seeing in the cloud is a result of the fact that the application leans a lot on infrastructure.

And so every time it needs to do something with infrastructure, it, it kind of leaves the comforts of the compiler, right? Because the compiler doesn't understand there's a cloud resource on the other side. And, and I, I kind of reconnected with, uh, an old friend Hai, uh, who I worked with at Microsoft.

I actually met him at Microsoft, spent some time together there, and I reconnected with him, and he's been, he founded a company and sold it at that, you know, during the time that I've been hacking on drones. And, and it felt, and he was like so excited about like this mission because he's been building stuff on the cloud. And so it was like his pain first, first class pain, you know?

Mm-hmm. Uh, so we joined forces and, and started this company. And, you know, the vision for, for, for Wing cloud is basically to create this new kind of cloud in a way, right?

Like, it's not a cloud in the traditional sense where there's like data centers and, and, and, and low level services. It's a cloud in the abstract sense. It's a, it's a layer that says, okay, this is what a cloud is, right?

Like, these are the ingredients of the cloud. It's the same way when as we're developing for traditional computers, we have this abstraction of the computer, right? It has RAM and C P U and memory and processes and file system.

Like it has a bunch of resources that we as developers know that we can use. And I can actually write code today that runs on all the computers, all the Mac, all the types of machines, all the types of operating systems, completely transparently, right? Because I, I'm working against this abstraction.

I'm working against this layer that, that, uh, uh, uh, cleanly defines the idea of a computer. And so we just wanna do the same thing for the cloud. We just wanna say, okay, let's think about a cloud as this big distributed computer and define the primitives, right?

Like, what are the primitives of a cloud? Um, and what are the language primitives that I need in order to express these systems, right? Like, because these systems no longer run on a single machine, right?

They run on multiple machines and they're interacting with each other, and traditional languages don't really have the ability to express those distributed computation models. Um, and so we are starting from the language, but the idea is to basically slowly build a layer, right? Like slowly build the pieces that allows developers and teams and DevOps teams to treat the cloud as a, as a, as a, as a unified, as a thing, right?

Like, as an actual thing that you can, that you can reason about, not just as a bunch of internet services. Got it. Yeah.

No, it all makes sense. You know, look, so one of the companies I helped put together was before there was Cloud, right? We were what they called an ASP application service provider.

And we were, we were offering hosted, managed things like Lotus Notes and Oracle, and you know, PeopleSoft, this kind of thing. And of course, we didn't have a cloud, we didn't have the virtual machines, we didn't have, right? And then you, you, you know, you saw, I remember when AWS came out, the, the original premise for AWS was, you know, you wanna run on Amazon, like infrastructure run on what Amazon runs on.

Same thing with Google, right? Run on what Google runs on. And, you know, then we saw the, the kind of rush, let's say 2006 to this, you know, commercial public cloud.

But like everything else in this world, a lot, when people have inflated expectations that are impossible to, the same thing's going on with AI right now, right? AI's going to do all of these things. It's gonna replace you, it's gonna replace me, it's gonna make, it's either gonna be the greatest thing for humanity, or it'll be the end of humanity, right?

No, it's, it's somewhere in between, right? And it's the same thing with cloud too. I think people have to have realistic expectations of what it's good and what it's not good at.

I, for one, never thought of the cloud as a money saver. I think it's easier to do, you know, faster certainly than spending up your own infrastructure. But at scale, it's not necessarily cheaper per se, you know, it, and anyone who thought they were doing it to save money, I think this is why we have this whole finops thing now.

But anyway, I think it's, it's a total cost of ownership. Yes. TCO not about particularly saving hardware costs, because obviously you pay a premium and someone else managing your hardware.

So it's sure, it's, I think like expecting that that's gonna be cheaper. I don't think that that's the right expectation, but again, it's about the total cost of building. And I think if you're able to actually leverage existing, you know, tools, uh, and, and when I'm saying ex uh, leveraging existing tools, I'm, I don't mean just the code, right?

Or just the, the, the software, but also the operational side. The, the fact that there's someone operating those, those tools and operating those services is hard and expensive. And so that's, I think the, that's where you're saving, right?

Like you're saving by, by not having to build those capabilities. Absolutely. So let me focus now a little bit on Wing Cloud, and I want to talk about Wing Lang, right?

So Wing Cloud is this company that, uh, you and your co-founder, we had met at Microsoft, founded it. And the idea it is to make it easier or better to use a cloud, and whether you're using AWS or Azure or G C P, whether you're on, you know, cloud, native Kubernetes, serverless, what have you, let, let's talk now. And so what role does Wing Lang play in that?

So if you think about the software development life cycle, uh, from left to right for, for all intensives purposes, uh, the left both side is the dev develop the software, the programming language, right? Uh, but then you go and, and there's stages throughout the development cycle, like, you know, building your software and testing your software and staging it and deploying it and operating it, right? Like, that's basically the, the cycle.

And so winging is at the left to most side of that, of that story, right? Like, is where, is is the, is the language, the programming language, whether, whether the winging or not is, is the way, you know, developers express what they want their system to do, right? That's, that's what programming languages are about, right?

Like, they're about expressing how, how what I, what I want my system to do, or in the case of the cloud, it's also about how I want my system to be architected. And I think this is, this is one of the key points about Wing Lang, and the analogy I sometimes give, and let me know if it makes sense to you, right? Uh, is that traditional programming languages, or the programming languages that we use today are technically designed around the idea that they create a sequence of operations inside a cpu, right?

Like, you write code and the compiler takes your code and produces a sequence of operations, and then they get, give it to a cpo, and it executes it over time. So in a sense, traditional languages are describing the time dimension, right? Like they describe what happens over time.

Now, if you think about cloud systems, cloud systems also have a space dimension, right? Like, there's also an architectural dimension. It's like, oh, there's this service that accepts the request here, and there's, uh, you know, it sends the request over here and there's a queue over here and there's a database over here.

And like, there's also this spatial architecture that is an intrinsic part of the application, right? Like, it's, it's, it's, you can't disconnect it from the fact that, right? Like, that's the application.

And, and within that space, there's a bunch of machines that also execute things over time, right? Like, there's also these containers or serverless or whatever. And so one of the big things about Wing is that it adds, this space dimension basically adds another compilation phase, uh, execution phase that describes the structure of your system.

And then there's also a way for you to describe the behavior within the cloud. And so it's, we are calling it pre-flight and inflight, right? Like basically pre-flight, obviously all the puns.

All the puns, mm-hmm. I get it. So pre-flight is kind of like the infrastructure step, right?

Like you described the infrastructure and inflight is the code that runs inside your infrastructure. And, and the beauty of that is that now the compiler has a full picture, right? It doesn't have only a picture of your infrastructure or only a picture of the application that runs on inside your cloud.

It has a full picture. So the easiest example is like, if my container or Lambda function puts a file in a bucket, then now my compiler sees that relationship between my Lambda function and the bucket. And so it could automatically creates I am policies, right?

Because it knows that there's a specific interaction between those two things. And that interaction was invisible before, right? Before because there were two different systems.

And so winging is basically the first step in that, in that story, because we have to give developers a way to express the system, otherwise we're blind, right? Um, but the, the, the interesting thing about the way we are building it, and I've always, I've always built systems like that. We've built the CDK in the same ways.

And, and I, I believe in that, I, I approach is, and that's actually something I learned at Microsoft, by the way, is that you build software in layers, right? Like you, you really wanna make sure that you have foundations before you create abstractions. And that's, that's a mistake.

I, I've seen a lot of people do, kind of, and that's a very popular thing to do, especially when you're a startup company and you're like trying to solve a very specific use case. And so you're starting from that use case, and you're starting by creating the exact thing that solves that one use case, that tunnel vision, right? And the problem with that approach is that it, it normally creates solutions that are very hard to expand from, right?

And so, you see, I sometimes call it the Heroku trap because Heroku had actually done something very, had some had that type of problem, right? Like, you kind of talk to people who've been to Heroku. Yeah, first year was amazing.

Second year we spent the entire year moving off of Heroku. Um, and so what I'd like to, I like to do is I like to build those things in layers. And so the, the, the idea is that the programming language is just one piece of the story, but there's another layer that's really, really important.

And that's the, the, the what we call the Wing cloud library. And that's basically the standard library for the language. But the library is, is a set of, of is this, that set of building blocks that are really abstracting the cloud.

Like those set of building blocks are the thing that creates the layer. And, and so if you're building on top of those building blocks, whether you're using Wing Link or you're gonna be able to use other languages with that cloud library, you're gonna get a lot of the benefits that, you know, that Wing Cloud offers without necessarily adopting the language, right? Like, we know a language is, uh, it's a big, you know, it takes time for a language to mature and to grow, and sure, you wanna give it the opportunity.

Um, it, it means that you won't be able to do this in-flight pre-flight thing, right? Because in-flight, pref flights are something that, I don't know, another language it has right? Built, uh, uh, support for, but you're still gonna be able to enjoy some of the abstractions.

And, and again, if you're building on top of those abstractions, then suddenly your application can be ported across cloud providers. Uh, it can run locally on a simulator, which is as a developer is probably the killer feature, right? For, for Wing, uh, being able to actually build on your local machine and get immediate feedback on the entire cloud application, not just on your one small piece.

Excellent. Alad, we're almost outta time for people who wanna get more information, where can they go? cloud is our company website.

Um, and we're on Slack and on GitHub. Uh, and so please feel free to reach out and ping us and yell at us and help us. I think, you know, we need, we need a, we need all the feedback we can get at this point.

And we're an open source company. We really, really believe in open source as, as a way of life. Um, and so we're trying to be very transparent with everything that we do.

Almost everything that we do is public and open and, um, inviting people to collaborate with. I love it. I love it.

So all things, wing, wing, dot, cloud, wing, Wing, Excellent. A lot. Thanks for being on Techstrong tv.

Best of luck with Wing Cloud. Keep us posted here. This is an exciting, you know, great story for our audience of developers who kind of love this stuff.

Enjoy. Thanks again. All right, we're gonna take a break here on Techstrong.

We'll be back in a moment with another guest.