"Stay away from my trash!" by Steve Ruiz https://tldraw.dev/blog/stay-away-from-my-trash

I've been thinking about this exact same problem of AI contributions in OSS, and I think he's onto something. "If writing the code is the easy part, why would I want someone else to write it?"

Stay away from my trash! - tldraw: Infinite Canvas SDK for React

The tldraw SDK provides tools, services, and APIs to build beautiful whiteboards and infinite canvas applications with real-time collaboration and a powerful React-based canvas.

"Once we had the context we needed and the alignment on what we would do, the final implementation would have been almost ceremonial. Who wants to push the button?"

This is what I've observed in my day-to-day work: the hard part is no longer typing the code but aligning on solutions, weighing tradeoffs, reaching a shared understanding, etc. For OSS I have no idea what this means except that the value of drive-by PRs has plunged to nearly zero.

I've long said that I saw little value in, say, PRs to my emoji picker to customize how the favorites are displayed, or whether the corners are rounded, or whatever – just use `patch-package` if you want to customize. Now this is even truer since developers are no longer limited by their familiarity with the codebase, coding ability, etc. If you want rounded corners just tell your LLM to do it – why should I accept your PR so that you can `npm install` it instead? It wouldn't even save you time.
This new era is so bizarre, and I don't even know what "open source" means anymore. I guess it means you've vetted some AI-generated code that passed your tests, which means you saved the next AI model some time in reinforcement learning when it slurps up your codebase? What a weird world.
@nolan sounds terrible to me.
@janl Not saying I like it! 😛
@nolan s/your tests/AI-generated tests sadly
@nolan I think writing code has gotten too much credit. The hard part has always been aligning on solutions, weighing tradeoffs, reaching a shared understanding, etc. Typing the code has always been like typing a book: if you already know the style, story and the language, typing it out is fairly easy. With AI, it's just easier to write the book in a language you are not familiar with. Or to let it write for you. But you still need all the other things, like we have always needed them.
@nolan What bugs me about this is that we reduce ourselves too much to just "production workers" in a factory line. We, developers, have always been so much more than that. The danger of this is that from a management perspective people can say: "well I have a machine for writing code now, guess I don't need you anymore." Or: "you must start producing more code because it is so easy now." It's like telling a chef he can cook better meals because there is a microwave now. It dismisses our value.