shykes

Docker worked well with lots of people.

Red Hat was frustrated with us because they were used to calling the shots. They couldn't fathom that we would refuse a pull request from them. Well guess what, every project is entitled to saying no. And Red Hat's contributions were just overall not good, or they didn't bother to consult with the maintainers on our design principles - like independence from the underlying OS distro.

You're reminding me why working with that company was such a pain. Always with the fucking superiority complex and passive aggressive drama.

As for a "security disaster waiting to happen". My view was, and remains, that the only isolation you can trust is the hypervisor. Kernel level isolation is not strong enough, and unprivileged user namespaces never actually mattered in cloud architectures. Your work on that feature was just security theater and ultimately useless, so spare me the lecture.

I do think Docker embraced user namespaces after my time but I couldn't care less.

Thanks for sharing! Docker has changed my life too :)

And yes the community has always been the best part of Docker...

Fun fact, we were almost acqui-hired by Engine Yard about a year before launching Docker.

I continued riding regularly for 5 years. First on a ninja 250, then a honda cbr 600rr. Stopped riding 8 years ago when I became a dad.

Oh that's so cool! And also making me feel very old :)

What are you up to these days?

At some point I emailed Linus to ask for his advice but he never replied. That was a personal initiative though, not on behalf of Docker.

I did speak to Greg several times.

I think over the years there was natural overlap, people who contributed to Docker would also be kernel contributors. Especially when we spun out runc, containerd, which created a more vendor-neutral space for maintainers to focus on the bottom of the userland stack with minimum drama.

To the extent that Red Hat employees overlap with kernel folks, that made it more difficult to collaborate since Red Hat was overtly aggressive towards Docker, and generally made our engineers feel unwelcome on "Red Hat turf". I discovered that they had devolved into a toxic company with a facade of benevolence. This made me sad because Red Hat was an inspiration to me. One of those "don't meet your heroes" situations.

Not sure what you meant by the "ignoring" part.

Thanks Pedro! Looks like we got buried on HN, oh well!

> One use case I've found for it besides CI was building an eval sandbox for testing different coding agents' performance using CLIs, APIs, and coding tasks.

Very nice! Would you be willing to do a video interview and show a demo for the community?

In my opinion evals and CI are converging into a common execution and data model for all pipelines involved in shipping your product. It's Continuous Engineering :)

> while the API and CLI are beautifully made, the documentation and the webapp scare the normies and could be improved.

Thanks for the feedback! Yes, we are working on the docs, it's not perfect yet but trending in the right direction I think. We are also betting that everyone will soon be asking their agent to setup Dagger, so we are investing heavily in making sure agents understand and appreciate Dagger, so they can set it up properly and explain it to their drivers. What a crazy world we live in!

None of those things were obstacles to commercialization. None of them were common requests from enterprise customers except private registries, which Docker did build.

Podman has never been commercialized.

Docker's difficulties in commercializing were self-inflicted. Enterprise and open source teams were culturally siloed and as a result the enterprise products were not "native". There was a lack of focus and a "checkbox" approach to building product. There was no singular "dictator" who could enforce a coherent product design and vision, and it showed. etc.

Hi Josh!

Hi Anil! Welcome to the 2026 Docker reunion thread :)

How did I not know about this video until today!

Hope you're doing well.

We are working on making Rust builds much faster, and easier to distribute on a cluster.

When self-hosting Dagger, you can run it on Kubernetes. Beyond that, we don't use Kubernetes at all ourselves.

We acquired a small team of Cambridge-based computer science & systems engineering wizards, called Unikernel. We then worked together to apply systems engineering to improving and expanding Docker in various ways.

The most successful project that we worked on together was "project Piñata" which later became Docker for Mac. It turns out bundling a Linux VM in a native Mac app, and integrating it into the host filesystem, networking stack etc requires lots of magic..

Now, most of the Unikernel team were very active OCaml developers, and created MirageOS among other ocaml projects. While working on Piñata, they simply reached for the tools they were most familiar with.