Open Source, AI, and Trust
retained text
AI usage is everywhere in dev, and it's not going away. But AI makes trusting open source... complicated... (arguably in a good way though). A big part of open source projects (particularly solo projects) was that they were typically someone's passion project. They cared about what they made, and that extended to the code too. with attention to detail to make it easily maintainable (i.e. legible to other humans), performant, and free of anything malicious. A big part of the trust associated with open source is trust in the maintainers. But when those maintainers aren't actually writing the code themselves, that trust is shaken. I have no specific objection to open source maintainers using AI though, as long as they know what they're doing. AI is a tool after all. But it's a tool that should be used responsibly, and with appropriate oversight. In the same way as I wouldn't trust a random 12-year-old to responsibly use a chainsaw, my trust in open source projects disappears when I see that it's been written by a non-dev relying entirely on AI. I've been a dev for over a decade (always in commercial settings, so my contribution to open source has always been financial, rather than through contributing code). I'm fortunate that my team's attitude is that even if AI wrote the code, it's still that person's code, so they better be damn sure that it's written well, and performs well. Code quality is extremely important to us, and we know that even with the most detailed prompts and surrounding context (instructions, skills, etc.), that AI sometimes spits out absolute crap. So if AI writes our code, then the time we would have spent writing it is expected to be spent on planning, reviewing, and confirming its output. The situation is muddied by more and more people using AI to churn out code without really understanding what that code is doing. Those sorts of projects can be genuinely dangerous too. Especially in a world of automatic updates, I'm not going to trust an app from a maintainer that won't understand when an LLM has just put ads and trackers (or more malicious things, like external executables) into their project. I'm still perfectly happy to use and support projects that were made with AI help, but I have less (or usually no) trust in them by default, and need to spend the time to review what I can of the code myself first. I guess what I'm trying to say is: for all the AI slop out there, there are also a lot of devs out there who know what they're doing. Devs who still care about what they work on, and who may be (and probably are) using AI tools to help them do it, whether that's personal choice, or their company expecting/requiring them to. AI has really highlighted the importance of human critical thinking, which is a good thing. Please don't immediately disregard a project because AI was used. Instead, approach it critically, and sceptically, as you always should have been doing. Continue to call out the slop, but also praise what's been well-reviewed and made with care and attention (regardless of the tooling that was used in its creation).
document fetched 2026-10-27 00:39 · canonical https://reddit.com/r/opensource/comments/1xfufbc/open_source_ai_and_trust