Why Git is no "good" for AI-generated code
github.comWhen AI agents are generating some, most, or all of your code, then occasional git commits of the resulting source code aren't sufficient. You also need a tool that ties the generated code back to the prompts and AI interactions that generated it.
Here’s a short technical explainer video of GOOD, a Git companion designed for this: https://github.com/specstoryai/getspecstory/blob/main/GOOD.m...
The core tool will be free (as in beer), but we may or may not be FOSS. We’ll figure that out soon’ish.
I would love some feedback on this! You need to have something people can try right now for a Show HN - take a look at https://news.ycombinator.com/showhn.html You are correct! I removed the Show HN. This is really interesting. Could you use it even on teams that aren’t using AI to generate code? It seems like it could help clarify intent for any kind of team. How different is it from well commented code? Or could it be used to add comments to old, poorly documented code? Could the AI ever infer intent so developers don’t have to document it - they just have to validate that Good correctly inferred their intent? > Could you use it even on teams that aren’t using AI to generate code? That's a great question that we haven't really talked about. The part of the AI codegen workflow that prompted this (no pun intended) is that it forces you to say/type aloud all your incremental intentions. It's the first time a tool has access to what was often previously inner monologue for most developers. But I know several developers who think aloud even without using AI (rubber ducking). But beyond just thinking aloud, using an AI code generator like Cursor Agent lets you express intent aloud and then easily change your mind (e.g. abandoning a code branch or using Cursor checkpoints)... the informal intent is right there next to the code for the first time. Yes, comments could be that in theory. But in 70 years, I still don't think we've got a single fully commented codebase that includes fully documented product and technical intent. Inferring intent is _really_ interesting. Though of course we have many examples where the code does not accurately express the intent (thank you, bugs). But could we bootstrap an intent store from existing code and then allow developers to validate and augment it with a more verbose process going forward? I think so. Love this -- as AI generated code inevitably leads to more decisions being made with less intentionality/understanding, having the context that AI was given is key to not creating a mess over the long run! You're spot on. Sometimes with ai codegen we end up offloading some of the intentionality to the AI. They work surprisingly well at filling in the details for under-specified prompts. I think it's important that this tool capture not only the prompts and context given, but also the full set of responses back from the model. There are implicit requirements buried in those responses as well that are made explicit at the moment a developer decides to accept those changes. THIS!!! The ability to retain intent is huge. Also creates an objective measure of process that can ultimately be improved. Essentially "intent + prompts = outcome" over time. Love it, this is a great step towards making AI generated code part of actual workflows in applications that go beyond initial POCs. Exactly. We've now got help in the inner loop, authoring code. But the bottleneck just shifts. If you're using ai codegen to write production code, but then throwing it all over the wall at another engineer to review you're not getting the full productivity benefit. It's time to envision an ai first coding workflow that covers the whole development lifecycle. Love this idea -- can't wait to try it out @ Quome Specifically, shadow git to manage multiple variants of a generated prototype for AI-generated code will be super useful to do fast iterations -- very interesting approach to solving the "we build fast, we need to collaborate" problem in this new era of vibe coding.