Former Chef Software CTO talks IT automation, open source

In a market that moves as fast as enterprise IT automation, Chef Software, founded in 2008, is now a venerable elder, and its former CTO is on to the next big thing.

What exactly that will be remains unknown. Chef founder and erstwhile CTO Adam Jacob has no immediate plans to start another company, but when he does, the business model will be open source.

Open source software is near and dear to Jacob's heart. One of his first activities as a member of Chef's board of directors was to oversee the company's transition in April 2019 to a fully open source licensing model, where previously products such as Chef Automate and Chef Habitat ran proprietary code. Jacob foreshadowed the move with a personal manifesto of sorts, dubbed Sustainable Free and Open Source Communities.

Open source software and IT automation are in flux and in the news as Jacob makes his personal career transition. We asked him for his outlook on the future of the IT industry and where he thinks he will make his next contribution.

What's next for you and for IT automation? Is serverless the next big thing? Anything else you're tracking?

Adam Jacob

Adam Jacob: I'll eventually start another company and do something in enterprise software, because that's where my expertise is, and that's what I like. But I don't know exactly what it'll be or when.

We have to build the system that makes people effective at adopting new technology -- whatever it is, wherever it may be in the stack -- that they need to run their business more effectively, instead of just the next platform.

Things like serverless are interesting, because they point the way to the user experience, and they're going to get adopted and have value. Are they the future of enterprise computing? Maybe for a minute. But then, there'll be something else. And until we get good at navigating those transitions, which we're completely bad at right now, I don't know that it matters.

At Cloud Foundry Summit, the focus was on business outcomes and incorporating things like Kubernetes into the Cloud Foundry platform. But the attendance at that show vs. KubeCon or even last year's Cloud Foundry Summit was night and day. That doesn't seem like the industry's focus right now.

Jacob: No. But even the Cloud Foundry folks came to that revelation through something that wasn't the purest of doors. They were the ones a generation earlier telling everyone, 'If you just adopt Cloud Foundry, all your problems go away.' Now, they're like, 'Well, not all the problems, but that's fine, because we also sucked up Kubernetes for you.' Maybe. I don't know.

Meanwhile, there's a clash in some instances between vendor sustainability and open source business models. How do you think that plays out?

Jacob: There's no evidence that when things are open source, you make less revenue, which I know sounds weird, because open source means you gave something away for free. Look at Elasticsearch -- AWS launched its Elasticsearch service in 2015. [But] Elastic Inc. went public [in October 2018] in the face of competition that supposedly is an existential threat to its economic sustainability. I just don't buy it.

What people relate to, and what people who buy software want, is a trusted brand that's putting together a product with security guarantees, product wrapping, marketing that you can use to convince your boss that you should use this tool, and sales reps that have attestations they can make as to its provenance. Nothing changes about the need to buy a product, whether the product is open source or not. My hope for the future is that we get back to that truth. It doesn't have to be that hard.

If that's the case, why are companies like Mongo changing the terms of their licensing?

Jacob: I think they believe that their value is actually in the proprietary parts of Mongo. It's easy for [a vendor] to say that the open source channel isn't the most relevant, but I think they're looking at it wrong. The truth is that they have a much bigger channel in open source, even if there are people who don't pay them. The odds are more in their favor, because there are more of those people in the world.

When I read your post about Chef's changes and looked at the Sustainable Free and Open Source Communities' website, it seemed the free software product model was one that you preferred. Is there a right approach?

Are [things like serverless] the future of enterprise computing? Maybe for a minute. But then, there'll be something else. And until we get good at navigating those transitions, which we're completely bad at right now, I don't know that it matters.

Jacob: It is situational. There are examples where I wouldn't use that. But in the situation that Chef is in, where we are the upstream and produce the vast majority of [the code] and own the brand, when people say what the software is, they'd say, 'There's Chef.' And even if there was another thing -- and there's a community effort right now to build a community distribution of Chef; they're going to call it Cinc -- when you ask them what it is, the answer is that it's a community distribution of Chef. That's an important distinction when you're trying to build a product.

HashiCorp, Puppet, all the infrastructure management vendors, [they] own the [intellectual property] and the upstream, and it's not clear the model we chose before was better. I'm not convinced that the reason customers bought Chef was because of what I held back. Now, I hold back all of it. [Laughs]

It's clear how that approach helps Chef, but how does it help customers?

Jacob: All of the reasons you ever buy software from a vendor are as true as they ever were. One of the most challenging things about buying open source software is justifying the transaction. In our traditional model, I give you this free software that you can do enough with to be successful, but hopefully not so successful that you're not willing to buy my commercial software. For those customers, you help them quite a bit. What you say to them is, 'If you want to do the work for it to be free, do the work for it to be free.'

Those people will come to Chef and put in effort that they don't have to put in if I deliver it to their door, which makes the whole ecosystem healthier. Or, they can pay me money, use my version, and I will maintain it for them. That is a much easier thing to discuss. And if you're going to take a piece of software and stick it on every computer in your enterprise, do you or don't you want to know where it comes from? I bet you do. It also helps them in that the community around the software is healthy, and that software has a long shelf life.