An e-mail interview with Alex Blackie from Worktree

Some of life’s best ideas come from just sitting down with a cup of coffee and staring out the window. I thought about a message I stumbled upon a while back, mentioning a new Canadian git hosting company called Worktree. As a fellow Canadian, I felt a sense of optimism that someone had taken on the challenge. As someone who has recently gotten into independent consulting, I was also quite curious about both the company and the people behind the process. So I decided to reach out to Alex Blackie, and see if he would be down to answer some questions for a blog post. He was, and I’m both thankful to him, and glad that I decided to hit send!

I have to give him a lot of credit here, there wasn’t a huge payoff to either him or Worktree to doing this. This blog has no revenue to share, and it isn’t big enough to result in significant awareness. However, I do believe that it goes to show that we can build a web in which we learn from others, and just how kind complete strangers can be.

Based on how well this went, I am thinking that I will try to make this a recurring theme on the blog (not always entrepreneurs or small companies, but just interesting people doing cool things in tech). Now, with the introduction out of the way, let’s get into the questions and answers!

What made you decide to become an entrepreneur, and how did it feel when you fully committed to the decision? How has it been so far?

Terrifying. Freeing. Exciting. Terrifying. Gratifying.

I’ve worked at a couple different startups and those experiences were by far the best of my career. Small, high-trust, high-agency teams working towards a shared goal… It’s an incredible experience, but it’s rare to find. I guess I’ve been chasing that high ever since.

Starting a company is amazing—your only barrier is yourself… But it’s also grueling and unforgiving. I’ve worked harder in the past couple years than I probably did for my entire career before this. But I wouldn’t trade it for anything at this point.

Why Worktree, and why now? Were there any other projects that you considered, and what ended up solidifying your choice to launch a Canadian-focused git hosting service?

For years I’ve always run my own Git service somewhere, just for myself (I was a big fan of Stash before Atlassian killed it). When Microsoft bought GitHub it left a bad taste in my mouth, and now its continued near-monopolization of all developer tooling has been a bit scary to watch.

As for “why now”… Platforms like Codeberg and Sourcehut have already shown that there is a desire out there for independent code hosting platforms, and that people are starting to get fed up with the status-quo. And now that the US has become increasingly hostile and unpredictable towards its allies, it seems like the perfect time to say, “hey, maybe time for another option?”

A goal for Worktree was also to be more commercial-friendly. Codeberg and Sourcehut are great projects, but it’s my opinion that without that connection to industry they just won’t have the resources or positioning to truly compete with an incumbent like GitHub. The lessons learned and technology built-in response to demands from commercial customers are the driving force that really grows and solidifies a platform for the long-term.

As is the case for almost all services hosted in Canada, the cloud vendor is American. Worktree currently uses AWS for the main hosting and DigitalOcean for actions support, with data centres hosted in Montreal and Toronto respectively. Where do you think the Canadian industry is lacking, and what features or functionality should be prioritized to create truly competitive Canadian companies? Are there any particular changes that might better help smaller businesses like Worktree?

This has been a personal battle of mine for YEARS. I’ve used every Canadian cloud provider that exists, and even previously used some that don’t exist anymore.

To sum up my experience, it feels like we’re 15 years behind the curve. Every provider feels like they sell the same IPv4-only VPS that’s running on a single core of a Xeon from 2012 and a network-attached disk that has so few IOPS it might as well be over SMB.

You can get a server in Canada, but Canada lacks a true cloud platform. We don’t have a provider that can give you a three-zone-redundant managed Postgres with 10k IOPS and automatic leader-elected failover. We don’t have a provider that can give you 9-9’s of durability and cross-region replication of your backup data in object storage with automatic cooldown tiering to tapes… AWS has this. Azure has this. GCP has this. IBM has this. OVH even has this… Zero Canadian providers have this. Most Canadian providers can’t even give you IPv6.

Part of this is probably the market dynamics of not having enough high-tech companies demanding this sort of high-tech infrastructure in the first place. To really get to the root of it, we need to keep fostering domestic innovation and make it more attractive for people to stay in Canada instead of brain-draining to the US. Naturally, some of them will keep their servers here too, and then we can start to build the cool stuff and move past self-hosted WordPress VPS’s.

As far as Worktree goes, we’re preparing to move everything to OVH and some bare metal soon – OVH, while still not 100% Canadian-owned, I think most would agree is a significantly better alternative and gets us out of the US Cloud Act purview.

Building a domestic cloud is something I’d love for Worktree to tackle as well, though obviously it’s no small feat. Some of the work we’ve done on our own bare metal deployment is laying the foundation for these plans though, so it’s something we’re working towards.

What is one thing that you have already learned from putting ideas into practice with Worktree. Is there something you had planned out, but had to do differently when it came to implementation?

I think trying to fork Worktree from the open-source project it was built on was a misstep. We’re tracking much closer to the Gitea upstream project now, and have scaled back a lot of our more invasive refactoring goals to keep that upstream compatibility going forward. The original goal was “hard-fork, rip and tear”… But frankly we just didn’t have the resources to pull this off while still pushing the platform forward.

We’ll probably still hard-fork one day, once our patchset becomes too unwieldy to merge, but for now this is by far the right move that benefits everyone involved.

Also, using AWS was probably a mistake. They did give us a bunch of free credit, though, which was helpful.

If you were to sit down with a team in a private company, and a team from a public body or institution today in Canada, what do you think they would be looking for most when getting started with Worktree? Is Worktree aiming to tackle the needs of both simultaneously, or one area first and then the other?

I don’t think the needs of either are particularly diverging. Both need to host and collaborate on source code. Both want to automate their processes. Both hopefully want to minimize their risk exposure to foreign interference, surveillance, and monopolization.

I’d hope that Worktree works just as well in a tech startup as it does in a government IT department. At the end of the day we’re all just developers working together, and that’s what Worktree wants to help facilitate.

GitHub is currently in the middle of a pretty big overhaul of the platform to be focused around Copilot. Do you envision AI being a part of what Worktree offers? If so, how? If not, what factors went into the decision?

Worktree has no current plans to launch any AI features, but we’re not anti-AI. Ultimately, nothing about Worktree has to change to let you use AI in your day-to-day development work.

I don’t think there’s much we could offer on the code hosting side that AI would materially improve. Perhaps there’s opportunity on the Cloud side to provide model hosting or other AI infrastructure, but that’s a whole other story (and there are already other Canadian companies doing this).

In any case, I’m certainly not in a rush to shoehorn some half-baked chatbot into every PR someone opens.