I really see this as trying to buy enough time for someone to rescue the investors, maybe the founders and definitely not the employees with an acquihire.
Docker is a good example of a company that (IMHO) should never have raised so much capital. It just doesn't have the moat to justify the valuation.
HN has posted several submissions (eg [1]). Containers aren't new. Anyone can do it. So where is the moat? Possibilities include orchestration (which they lost to Google's Kubernetes). There's no barrier to creating images or even having a public registry of images.
It always seemed tike containers were just going to be another feature on cloud platforms. Don't get me wrong: I think containers are a really good technology, for building, testing, deployment and so on.
Docker never had a clear value add and over the years has failed to develop one.
While I am normally critical of the unicorn pizza-kitchen-on-wheels type of excess, I think this is only true in hindsight and has a lot to do with Docker lacking a commercially minded founder. Open source isn’t a business model. There are very synergistic business models that go hand in hand with OSS but that distinction is important. Github does not really have a defensible moat. I think the network effects there are mostly trying to back-solve for its popularity. But it’s just a good product that became synonymous with modern version control. Docker had similar potential, but needed more commercial creativity.
Github is a good comparison here because obviously anyone can run a Git server but Github did create a lot of value adds and the UI helped build a network effect for the ease os use, cloning, PRs and so on.
What's more, Github became the engine for dependency management. Go springs to mind here. I actually thought this was a terrible system (eg putting repo owner names in import strings) but it speaks to ubiquity of Github.
But what are Docker images? Maybe a few hundred lines of Dockerscript at the end of the day.
Losing in orchestration I think was the obvious big fail. But they had an uphill battle here anyway because you really need to integrate such a thing with cloud platforms.
I'm really not sure what Docker could've done differently here.
> I'm really not sure what Docker could've done differently here.
Completely agree, it’s trickier than GitHub. But this is why founders of these companies can potentially make billions: if it were easy, everyone could do it.
I think they realized the CI/CD potential far too late. In another universe you push to GitHub, Docker builds and tests your images and deploys to your provider of choice. Their potential was probably not directly tied to containers but tied to their position in the engineering process between commit and before deploy.
> I'm really not sure what Docker could've done differently here.
I think they should have realized orchestration was “the” thing for production much sooner. It’s not like you can’t integrate with cloud vendors on your own; there are plenty of managed service providers where you can get hybrid cloud solutions, Docker could have bet big on this.
Instead they came with swarm, which was focused too much on self-managed “on-prem”, while people really wanted something more complex, managed and with a healthy ecosystem of service providers.
Docker got stuck with being a software vendor, but they should have pivoted to being a service provider much, much sooner.
Docker is a good example of a company that (IMHO) should never have raised so much capital. It just doesn't have the moat to justify the valuation.
HN has posted several submissions (eg [1]). Containers aren't new. Anyone can do it. So where is the moat? Possibilities include orchestration (which they lost to Google's Kubernetes). There's no barrier to creating images or even having a public registry of images.
It always seemed tike containers were just going to be another feature on cloud platforms. Don't get me wrong: I think containers are a really good technology, for building, testing, deployment and so on.
Docker never had a clear value add and over the years has failed to develop one.
[1]: https://news.ycombinator.com/item?id=22244706