Hello! I have not been blogging for a while, but I have been watching this issue blow up for a few days, and I wanted to put down some thoughts.
Beyond the obvious surface question of whether the userbase wants or doesn't want AI in their terminal, I feel like this thread points out a really important meta-issue that's worth keeping in mind if you are part of the leadership team in an open source community.
They've been working on this feature for at least a year and clearly their entire userbase (including me!) had zero idea it was coming. They can say "we had it in a beta, and it was in the roadmap" all they want, but given the immediate emotional backlash, obviously most of their users just don't engage with the project on that level.
For me, iTerm2 has been a tool that Just Works for so long I really don't even think about it. As a user, that builds a huge amount of trust, and launching this feature in an update without the wider community being aware of it ahead of time is a great way to break that trust with a lot of folks, whether you think it's justified or not. I can sympathize that for the developers, it's surely frustrating to have so many people relying on your work that never even peek in at the project other than to download it, but the reality is that's the large majority of the userbase of any popular open source project.
Up until recently, I was at OpenNMS, an open source network monitoring project that was Doing It Right™ for a very long time, even with a commercial component. However, since being acquired, it has kind of lost its direction as it fumbled around trying to work on a new product, and I've watched it slowly sink into obscurity as it has failed to engage meaningfully with its community. It's still a great tool with a lot of features that are unmatched elsewhere! But without the open source community being an active part, it's just one more in a giant pile of enterprise network monitoring tools.
I'm now at NetBox Labs, working on putting together an on-prem enterprise offering. It's been a whirlwind of learning new technologies and getting comfortable at a new company after being in the same place for 17 years. I've enjoyed the heck out of it so far, and there are a lot of smart people doing good work here. Most importantly, though, the thing that attracted me most to NetBox Labs (besides having a friend who already works here 😀) is that we do really seem to be walking the walk on leaving the OSS version of NetBox to be its own vibrant community and tool. We employ the lead maintainer, Jeremy Stretch, and obviously he has a strong role in guiding the project, but we are regularly and actively reaching out to the other maintainers for things that might affect them and being super open with the community early and often. The community is active, engaged, and positive about NetBox Labs helping push the platform forward while building a company around it.
NetBox is doing a pretty fantastic job of being out in the open about its development (even the commercial stuff), so the idea that the iTerm2 developers could be so far from their users that they couldn't imagine this backlash is terrifying to me, and it's important to me that we make sure we don't ever end up going down the same path.
This should be a warning sign to F/LOSS projects that there is more to your user community than just the people that engage with the project, and it's as important (or more important) to work to understand them and figure out how to engage with them if you care about being a truly open project.
Top comments (0)