We currently are trialling a private cloud offering with a few companies that gives you dedicated hardware and domain name isolation without any of the management burden.

As responded elsewhere, we are planning on making the connection to cloud db services extremely smooth. A lot of the pain of using cloud services has to do with managing environments and secrets, and we'll tackle that problem very soon.

I believe strongly on building on top of open standards. I appreciate the Lambda solution, but the power of Now is in its ability to host everything with almost no new APIs. package.json, Dockerfile and HTTP/2 or WebSocket as an API cover a huge array of usecases for Cloud and IOT.

We are constantly iterating on the plans, and will have news along these lines in 2017.

Our plan boils down to:

• Continuing to improve upon the existing base. Make everything even faster, more stable and scalable.
• Listening to our customers and addressing their requirements

I spend a lot of time reading and trying new things. I try to learn as much as possible about emerging technologies and paradigms, like Bitcoin. I try to take nothing for granted and embrace the dynamism of it all.

This is a great question. Now Desktop has a very big surprise coming. It's actually quite fundamental to our strategy and our future!

Absolutely.

From the get-go, I wanted Next.js to resemble "vanilla CSS" as much as possible. It took me a while to come up with the exact implementation (styled-jsx), and in the meantime I personally liked Glamor best.

So what we settled on is to embrace all CSS-in-JS systems. I think if I were to pick one from scratch today I'd use Glamor or Styled Components. Styletron has features of both and I like it as well.

At the same time, I think inline styles are also super powerful (in their simplicity!) and work amazing well for most use cases.

As a final note: CSS is actually a very nice syntax. I find that I'm really productive writing CSS! It's nice when the selectors don't leak and you can use it within component boundaries.

Absolutely. We already have two very large organizations set up in this way, and look forward to elaborating on their use-cases in 2017.

This is a great question. I want to make our OSS development sustainable, responsible and predictable. I wrote our "policy" on the subject as a result: https://zeit.co/blog/oss

Some of the most important principles are:

• It has to have demonstrated its value within the company. The project should almost be in a "desperate need" to be open source. This happened to us with Next.js, for example. We were building quite a bit of internal documentation. I wanted to set that free. I find myself learning from the community now! We've had a lot of great help, insight and ideas via GitHub issues and the likes
• The software we open source has to have a small and limited yet extensible surface. A trick for accomplishing this is focusing on components (like Next.js) or plugins (like Hyper). This takes away a lot of the burden from maintainers, and also at the same time puts the community "in charge". It's a great balance. Moving forward, I'm skeptical of projects that don't meet this criteria and likely won't open source any, unless they're small components and such.

I spend a lot of time thinking about purity and powerful type systems. Tony and I were recently chatting about how his experience with Haskell, for example, has made him a better JavaScript programmer.

The neat thing about learning those systems is that they have a lot of transferrable wisdom and ideas. Sure, sometimes those ideas don't transfer directly as boundaries within the code, but they can exist within your head as you architect your software.

So, while we haven't really used Elm and PureScript directly (yet!), I'm very grateful for having been exposed to them and their associated ideas.

Everything is custom. Most of the solutions out there are rigid or static, and our cloud is very dynamic in nature. You can probably tell from how quickly we boot-up our deployments that those services don't really fit the bill.

We are equally excited about it! One of our goals in the short term is to make sure now.json gives you access to tailor every single aspect of deployment, scaling, visibility. Expect a lot of news on this front

We treat your development state as seriously as your staging state as seriously as your production state. The main deciding factor in pricing will be how much of our resources you take up, and you'll be charged accordingly.

This question is actually planted by my wife Alyssa 😏. I love you and ▲ZEIT would not be where it is without your support 💖

No unfortunately, but it seems a nice to support.

Our algorithms consider CPU, RAM, traffic and historical data in order to make the decision.

Later this year you'll start getting notifications about this on your activity feeds and desktop apps, in addition to some exciting customization opportunities.

1) We already put out https://github.com/zeit/micro-cluster as an interim solution. Thanks to Dockerfile, you can also deploy and scale services like Nginx to front your microservices. We definitely want to and will make this process easier

2) We don't think there's such a thing as a "general data solution", only offerings with different tradeoffs in latency, power, replication capabilities, etc. Our intention is to make the process of setting them up and understanding their capabilities and tradeoffs extremely smooth.

The default behavior of Now is that your code is open-source unless you upgrade. This provides a very big incentive for teams and companies to move onto the next tiers.

In addition, many of our largest customers have very specific feature requests, such as machine isolation (single-tenancy), certifications, logs replication, 24/7 support, realtime insight into metrics and auto-scaling events, and much more.

We plan to invest quite a bit on improving the README for micro (Tim Neutkens is leading that project) to feature examples (like Micro + GraphQL), screencasts and more.

Also check out:

• egghead.io/lessons/node-js-build-microservices-in…
• https://github.com/micro-analytics/micro-analytics

I forget 🙈

We are looking forward to introducing region affinity flags in now.json for scenarios where you have to carefully control latency, or cases where the database itself is not globally replicated.

Yes. We are introducing more plan with a variety of features and relaxed limitations later in the year.

We think the problem of dynamically scaling and distributing stateless HTTP/2 services is extremely interesting and has originated a number of very novel solutions on our end.

There's a lot to do in this space yet and that's why we've been focusing strongly on that.