> This competition didn’t hurt Hashicorps chances of profitability, they were factored in from the beginning.
They weren't, actually.
From a Hashicorp FAQ article (which is a transcript of a video interview which has since been delisted from YouTube) titled 'Why is HashiCorp committed to open source?':
> Mitchell: [I]t's always sort of been a default for me. [...] When we were starting the first projects, we both didn't intend to ever start a company around them. There was no monetization goal at all, and so I think open source was an obvious default then.
That feels very icky to delete that at this very point in time. It’s PR, it’s spin, very corporate, very untrustworthy. They could have left it up and responded to it, but deleting is manipulation in this case.
Agreed, although there's no good way for them to leave that up.
No matter what addendum they add, or how they mark it as archival, it reveals that HashiCorp was not actually committed to open-source like they claimed to be. Their word was 'committed', but in actuality open-source was just a feature of the license they happened to be using at the time. So that page will always invite questions about what else they say they're 'committed to'.
Leaving it up but changing the title would seem underhanded in the same way. Leaving it up and changing nothing would be confusing and look like an obvious oversight.
To complement sibling post, some other stuff that I think makes it clear that open-source was not one of HashiCorp's values even before this change:
DM@07:30: So how we think about it is, we're kind of obsessive about journeys. [...] And we decompose the personas into two: there's the practitioner, and then there's the decision-maker. The practitioner, you're trying to progress through a journey of discover and learn, try and trial, use and advocate— and you do not care if they buy anything from you. The decision-maker you're trying to progress through the why/try/buy journey. [...] It is a pre-dredged river, it is not fair; they do not know what is happening to them. They are not going where they think they are going— they are going where I want them to go. [...] And it turns out it works.
GS: Is there anything specific, do you think, about open-source that you layer in?
DM: For us, open-source is really just a distribution channel. (Also it's a development channel, but predominantly it's a distribution channel.) [...] So I don't think it matters whether it's open-source or not, as much as just being really, really obsessive about the digital videogame that you play.
--
On the one hand, it's nice that Dave McJannet is so up-front about these strategies, how manipulation is a part of them, and how much market positioning is about pushing people around to retain control of an ecosystem.
On the other hand, he sounds like a pushy, arrogant asshole tbf.
Sure, some of it. In Uppercase the most important strategy messages.
DM: Dave McJannet, CEO at HashiCor
GS: Glenn Solomon, GGV Capital Managing Partner
15:22 GS: "...reminds me a little bit of you know when you when you listen to Mitchell and Armand talk about the early days with the open source and how they were kind of you know hand to hand go really going going for meetup to meet up and trying to get feedback from users..."
15:43 DM "yeah I think open source lends us up really well to the rapid open feedback loops which is one of the reasons why I think the best products are built in open source YOU HAVE TO BE REALLY CLEAR ON WHAT THE MONETIZATION PATH IS but but the feedback groups are hard to beat um you know I think the the neat thing about that we've done is we actually uh we did it on six different products and now eight so we've kind of run the same..."
16:18 DM "how do you get that minimum viable audience right that there are different ways you can do that in the early days we did that through social media um I have to talk about that one um and you know content and entertaining people but I think this notion of also building Network effects into the products is the other reason so yes we went knocking door to door but it was deeply considered in terms of how you integrate..."
16:43 DM "think about how terraform works as an example there's terraform core and then there's a plug-in for every piece of infrastructure in the world THAT'S ON PURPOSE RIGHT AND THE PROJECT'S ACTUALLY CONSTRUCTED THAT WAY SO THAT OTHER PEOPLE CONTRIBUTE AND COLLABORATE so there are lots of different ways to do that but ultimately it's about Network effects and how you drive it and then..."
17:11 GS "Dave start with you maybe uh talk a little bit more about like uh what what that TERRAFORM PROVIDER COMMUNITY is built um uh how you get third parties involved and like what what's been some of the secret sauce that's made it so Central"
18:08 DM "... so I went to Microsoft and I went to Oracle and I said hey I think we can help you I went to Google we can really help you because you're way behind and and and we actually did BD deals with them to to get into their communities to make it relevant to their communities and IT WAS DELIBERATE there are only four communities that mattered at that time and we kind of played them off against each other then we invested deeply in this digital distribution channel OF A MALICIOUS USER JOURNEY to success for people it was a combination of that kind of the anchor tenants plus the big investment in the digital experience plus you know we you know we to this day we probably have 60 people..."
18:54 DM "today yeah so today there are about two and a half thousand terraform providers out there in the world we develop about five of them um because what we were able to do is WE'RE ABLE TO FLIP THE MARKET AGAIN THIS IS A LOT IT'S A LONGER LONGER DESCRIPTION IT'S LIKE IT'S IT'S VERY MALICIOUS IT'S THE RIGHT WORD so what we did is we said hey we decomposed the project into core Plus providers because that's how you do it in the open source Community we control core outright and play every committer but anybody can contribute to a provider but number one we control the certification process right nobody else can certify it number two we then when we got to you know we knew if we got to 200 The Fortune 500 using terraform to interface TO AMAZON IT WAS OVER BECAUSE THEY'RE IN CHARGE not Amazon and so you were then able to force every isv in the world to say if you want to be part of you know JP Morgan's Cloud program you have to build a terraform provider so it was a combination of a few things like that but IT'S VERY VERY DELIBERATE SORT OF THE ARCHITECTURE of the projects owning the certification process and then then owning the key communities that you care about..."
They should have been. Them being irresponsible early on is negative for the community.
Cynically, I believe they intentionally were stupid knowing that their tool would only become widely used if open source (who wants infrastructure as code locked into some proprietary thing).
One of the things that comes through for me in that FAQ, Mitchell's story about learning to code by reading and playing with open-source projects, and Armon's assertion that this license change is continuous with the company's original 'open-source ethos' is that this has always really been more about 'source availability' for them— I think they're credible, in way.
I still think of open-source in a traditional, historically informed way: 'open-source' is about software freedom at bottom, even though its name and common arguments in its favor are more practical and self-interested.
I get the impression that Mitchell Hashimoto never saw it that way, and certainly didn't buy into the actual 'open-source ethos' in the sense of the values of the people who decided the term and founded OSI, etc.
A source-available license like BUSL probably would have been a better fit for HashiCorp from the start, or as near the start as possible. But I don't know that 'source-available' as a label could have driven adoption the same way in the early days, and it seems plausible to me that the founders might've realized that too.
> A source-available license like BUSL probably would have been a better fit for HashiCorp from the start
My problem with source-available licenses is that it makes the software miss out on the network effects of contributions.
I’m not going to contribute to source-available licensed software just like I won’t contribute to Windows (source code is available to see but they don’t even accept contribs).
I like contributing my time to communities and to building things together. I make pretty minimal contributions because I’m paid to write software for an organization and don’t have time to contribute meaningfully.
Even if hashi worked out some way to compensate me for my contrib, I don’t think I would bother because the amount would be negligible.
So if the idea is to have high quality software due to network effects, source available doesn’t seem like a good fit for this.
It also seems moot to me because decompilers have been able to get me the source of things that aren’t open source. Being able to view the source code isn’t as important to me as being able to work on things as a community.
Having a single corporation reap all the benefits of a community is not something I want to work on.
Totally agreed. But we've also gotten clear signals from HashiCorp that they aren't very interested in community contributions for a long time now, like removing commit access from community maintainers years ago and no longer allocating company time for reviewing community contributions.
They liked the idea that customers could try before they buy, and that customers could inspect the source code for debugging or other purposes, but they clearly didn't buy into the whole catb thing.
From where you and I sit, it's a missed opportunity. Perhaps Hashi had good reasons to believe that they had to be the ones to drive Terraform forward themselves, and that community contributions would at best play a negligible role, whether that's because core Terraform code is hard to work on, because the contributions they saw tended to be minor, or because they wanted all the core contributions to be their own for the sake of retaining control. But whatever their reasons, they weren't on the same page as you and me about this.
They weren't, actually.
From a Hashicorp FAQ article (which is a transcript of a video interview which has since been delisted from YouTube) titled 'Why is HashiCorp committed to open source?':
> Mitchell: [I]t's always sort of been a default for me. [...] When we were starting the first projects, we both didn't intend to ever start a company around them. There was no monetization goal at all, and so I think open source was an obvious default then.
https://www.hashicorp.com/resources/why-is-hashicorp-committ...