3D39739091

Nobody can.

"Claude Code uses Bun" is the reason that's given. But even though Anthropic tells us that "coding is solved", instead of rewriting Claude Code in something other than JS, they bought a company that made a JS runtime and then did a mass rewrite of that JS runtime.

Brain dumping to a blank page and organizing your own thoughts is still an option.

The issue is that the people evaluating you don't know the difference between legit domain expertise and pure bullshit.

Real talk, though, you should already have talked to enough people about solving this problem for them that you have a list of contacts who are eager to sign up.

Sounds like you don't have that. I'm kinda going to guess that you fell into the trap of building something without validating the need for it first.

Because to be completely honest, no business would ever sign up for this, and no reasonable individuals will sign up to carry anonymous packages through customs. The people that would deal with this for whatever $25 you can pay them are exactly the people you should never trust to carry your customers' packages in the first place.

Drug mule as a service

Seems like the author's only talking about the extreme end of the spectrum. Not every fork is attempting to replace the forked project. There are so many other lesser degrees.

You can fork something just for a fun or experimentation. You could have a use case that the project doesn't handle and you aren't ready or interested in contributing that solution (especially if you only need this for a one-off scenario or a short-lived project).

This could also apply to needs that your client or company has, but it's out of scope for the original project, so you make a private fork that you and your team maintain internally. It could be that you DO actually make this public (either initially or eventually) so other people who have the same need can benefit and possibly contribute.

It could even be that, over time, the amount of users of your fork convince the upstream project that there is a need for this use case. Maybe they decide to handle it themselves or maybe your fork merges back in with the upstream. Sometimes projects just can't say yes to certain things because they wouldn't want to/be able to implement, maintain and support it. Seeing that you and others maintain a fork for a non-trivial amount of time can establish the credibility that there is indeed someone who will maintain this.

Sounds like maybe the next project should be a better bug report/issue tracker

> LLMs likely have a much better understanding of what our users need and want.

They don't.

Basically this sounds like Agentic Fuzz Testing. Could it be useful? Sure. Does it have anything to do with what real users need or want? Nope.

Why not just do your work in an organized way in the first place?

Yes.

This is exactly the best part about the Linux experience right now. There is nothing that's there because a PM is trying to get a promotion.

> Learn your tools, people. This is the equivalent of a contractor taping rulers together because they didn’t realize their bag comes with a tape measure.

I feel like this is one of the big things we lose when working completely remote and disconnected: missing out on all these chances to help each other. We don't know the things we don't know.

Imagine OP was going to do a rebase -i, sitting next to a teammate and said, "man, I would do anything to be able to use interactive rebase without having to learn vim", and their teammate turned back and said, "Oh, you can use any editor you want".

If anything, AI tools are doing the opposite. You don't acquire those skills by outsourcing your thinking to an LLM.