Discussion
(1/2)
@n_dimension
> That last paragraph is actually the most potent argument I have seen against vibecode in FOSS
The reason I linked that particular article from Pluralistic is that he goes into some details about why none of the most important challenges in software development consist of 'how to write more low effort code, faster'. In fact, more code is antithetical to good software engineering, which like good writing, aims to do its job with fewest, simplest lines of code possible,
Projects like fediverse.party and delightful.social.coding are going to be more work to research for. Because there will be a cambrian explosion of vibe coded projects, most of which will be create-and-forget experiments like ClownFlare's Wildebeest. But will still clutter up the search space, as they are for the IRC researcher I quoted above.
(2/2)
Theseus boat.
@n_dimension
> Theseus boat
How does Murphy's Axe apply here?
If the software is functional and fit for purpose.
What does it matter how it was created?
So fully human input was changed to partial human input.
At the end of the day, if it performs advertised function. It's useful software.
If you don't know if the software is made by machine. Can you still tell?
(1/2)
@n_dimension
> If the software is functional and fit for purpose
That "if" is doing a *lot* of heavy lifting, I hope it did some stretching first ; )
> if it performs advertised function. It's useful software
There are many good reasons to believe vibe coded software doesn't fit this description;
https://pluralistic.net/2026/01/01/39c3/
But even putting that aside, low effort software is much less likely to be maintained over time, or responsive in any way to the needs of anyone trying to use it.
That last paragraph is actually the most potent argument I have seen against #vibecode in #FOSS
And I speak as a vibecode practitioner.
As to pluralistic.
I have huge respect for his past work, but I think he, like many who don't actually use AI is way off the mark and increasingly drifting further off it as the tech advances.
His "reverse centaur" is just laughable (to me) and "asbestos in the walls" assumes a world of Butlerian Jihad and that IMHo is higly improbable.
But as well as an activist he is also an author and no once can accuse him of not working to satisfy his audience.
Its a bit, like a vegetarian writing about steak cuisine.
But, thanks for curating FOSS, it's a rarely appreciated duty.
(1/2)
@n_dimension
> That last paragraph is actually the most potent argument I have seen against vibecode in FOSS
The reason I linked that particular article from Pluralistic is that he goes into some details about why none of the most important challenges in software development consist of 'how to write more low effort code, faster'. In fact, more code is antithetical to good software engineering, which like good writing, aims to do its job with fewest, simplest lines of code possible,
(2/2)
@n_dimension
> If you don't know if the software is made by machine. Can you still tell?
That's more like 'if a tree falls in the forest ..." than Ship of Theseus. Vibe coding is more like a ship vaguely like Theseus', built by a bunch of people who have seen lots of ships, but have no experience with boatbuilding ; )
If the software is functional and fit for purpose.
What does it matter how it was created?
So fully human input was changed to partial human input.
At the end of the day, if it performs advertised function. It's useful software.
