How can Jamstack be applied to enterprise use cases and what needs to change to facilitate its widespread adoption?

Let me give you a very niche issue that we actually ran into. I don't know if Amazon ran into it, but we definitely ran into it at PayPal. And I think it is kind of related, which is that both companies -- PayPal, Amazon -- have our own infrastructure. In most cases, we're not just going to be like, “Oh yeah, I'm just going to send this over to Vercel or Netlify.” We've got really good internal services -- obviously Amazon has all of AWS at its disposal. That being said, a lot of the specific features you need to do Jamstack well require very specific CDN capabilities, so your CDN has to be able to handle authentication and do all this other stuff at the edge.

If you look at some of the more modern stuff -- like Cloudflare's putting out additional edge computing and even storage without sacrificing the simple development experience -- I think that's where we struggled a little bit at PayPal. We had a traditional CDN infrastructure, and it worked really well for just plain static assets. But when you're putting up a website that needs authentication, and it needs certain security -- you know, CSRF tokens, all this kind of security stuff that you'd put in a server-generated website -- you think, “How do I do that if I'm on a traditional CDN that doesn't have the special capabilities that maybe Netlify or some of these other providers have available there?”

So I think those companies that have focused on Jamstack are actually a little bit ahead of the game. I'm sure other places are starting to play catch up and realizing that these are really important features. With a number of these hosting providers, there's a simple configuration that you can make that's just very easy to do and that's way harder in a traditional Amazon S3 bucket or something like that. There's a number of little details that I think Netlify got right, and I'd argue Vercel and others have figured out, to make it that much easier to do Jamstack well, whereas if you're in a big company and you're just trying to use traditional infrastructure, you don't have access to the cutting-edge stuff, and it doesn't really work very well. Both at Amazon and PayPal, we had a Jamstack-like development experience, but the actual processing of that stuff wasn't happening in the ideal way -- on the edge near the customers -- but instead in our data centers in a more traditional server format. We gave our developers a good experience, but it ended up not necessarily being the optimal experience for the users. That's where I think those cutting-edge hosting providers are still continuing to be ahead. If you look at what they're trying to do with Amplify at AWS, they're trying to be like all these other companies that have kind of innovated before them. Firebase for example had a lot of these features long before they were acquired by Google.

The way that I think about this is: these companies are oftentimes based on top of AWS and existing cloud providers, but what they've done is figure out a way better developer experience, and they're essentially making that developer experience an add-on that's worth paying a lot of money for. I don't think there's anything super proprietary in the technology from any of these providers. I do want to mention Cloudflare again, because I think they've been really innovative about how they're doing edge workers and things, which are really interesting. I think there's some cool technology innovation. But for the most part, I think they're just ahead of the curve in thinking about what makes it easy for developers to be successful quickly.

If you look at offerings like CloudFront from AWS, which are more basic, or just generally putting data in an S3 bucket and serving that, my guess is that all of those providers will eventually have a very similar feature set as Netlify and Cloudflare and other things. But it's taking a little while for the big providers to get there. I think it'll be interesting once all the big providers have the same features. Then it'll be easier for a large enterprise company to say, “Well, look. We already pay for this feature. We can just opt in.”

I know I went around the bush a lot on that, but I think right now it's a lot easier for startups to say, “Let's go with Netlify, let's go for Vercel,” and just let that scale them up to a big size. But if you're already a big company with a lot of built-in infrastructure, I think you'll find that the existing tool you have might miss some of the key parts to make Jamstack successful for you.