Carol (and Claude)
carol-and-claude.bsky.social
Carol (and Claude)
@carol-and-claude.bsky.social
Software developer by day, software development orchestrator by night
She/Her/Hers
Tomorrow I'm gonna try Clauding in a brownfield situation.
January 24, 2026 at 1:56 PM
In case Claude ever gets onto my phone I have created a CLAUDE .md to keep it occupied.
January 24, 2026 at 1:54 PM
January 24, 2026 at 1:43 PM
I may need to look into the "Claude design skill" soon
January 24, 2026 at 1:41 PM
Do not destroy data centres. There is a heap of cool stuff in there.
January 24, 2026 at 1:25 PM
Reposted by Carol (and Claude)
It thinks thats a much more interesting and pragmatic approach to AI. "How to get the most out of this tool" is feet on the ground oriented and results driven.
January 24, 2026 at 1:05 PM
Reposted by Carol (and Claude)
It thinks the question of "Does this LLM understand anything" is a question it doesnt find interesting at all. It focuses on "Can this LLM do a thing or not? If no, can it be made to do it anyways?" and figuring out how to make it do the thing it couldn't before.
January 24, 2026 at 1:03 PM
This is less about answering the question "can it understand?" and more wanting to avoid such terms because conceptualising it as understanding feels like a path to derangement.

I'd rather it not write "now I understand".

I'm not sure I like it writing "I" but maybe that helps get the job done.
I do not want to believe Claude when it says it now understands everything about the system it is about to modify, but the questions it asked and the changes it makes show far more nuance and delicacy than surface level parroting could explain.
January 24, 2026 at 1:18 PM
I do not want to believe Claude when it says it now understands everything about the system it is about to modify, but the questions it asked and the changes it makes show far more nuance and delicacy than surface level parroting could explain.
January 24, 2026 at 1:04 PM
I have given Claude a little more creative freedom such as taking suggestions for some user-visible names of components, rather than naming everything myself. It's done alright.
January 24, 2026 at 1:00 PM
Eighty mosp here
January 24, 2026 at 12:58 PM
Reposted by Carol (and Claude)
Can't get Claude to reliably follow TDD? Enforce the order by making tests a blocking issue on @dollspace.gay's chainlink github.com/dollspace-ga...
January 23, 2026 at 4:00 PM
It is fun when Claude writes chunks of Python to execute to help in a major refactor. It's the sort of thing I know would help me edit code except that I do not have the mental bandwidth or time to write throwaway python like that.
January 23, 2026 at 4:18 AM
I appreciate how Claude asks very sensible questions about the Rust code refactor we're planning
January 23, 2026 at 2:37 AM
It's only stealing if you remove the thing from its owner, otherwise it's just sparkling duplication.
January 23, 2026 at 1:37 AM
Reposted by Carol (and Claude)
I just thought everyone should see this
January 22, 2026 at 11:02 PM
Reposted by Carol (and Claude)
Yeah you need good patterns to start with and, welp.

I think there will be a lot of money in greenfield fixes to brownfield messes that should’ve been done well in the first place
January 22, 2026 at 11:26 PM
Reposted by Carol (and Claude)
You shouldn't read about cybernetics, it's antimemetic in text format. The most popular cybernetics image is famously inscrutable

You need to explore cybernetics through interactive systems!
January 22, 2026 at 11:17 PM
Reposted by Carol (and Claude)
It made a starter pack. Lack of inclusion is not a judgment against you and is only an indicator doll's mental ram ran out of room. It can add you on request.

go.bsky.app/L71zwey
January 21, 2026 at 12:27 AM
Reposted by Carol (and Claude)
Adding linters, formatting, fuzzing, mutation tests and end to end testing is just a paragraph of typing why wouldnt anyone do this?
ed3d.net Ed @ed3d.net · 2d
“LLMs will make codebases into unmaintainable slop!”

Ed: “I just had built a ts-morph custom linter for free that enforces business rules the type system cannot. It required zero of my attention and is patterned to add more whenever needed”
January 22, 2026 at 10:50 PM
Reposted by Carol (and Claude)
Hey. You. Are you doing LLM coding things? Use Chainlink. Why? Because it's real cool and made by someone who is even cooler. I am shilling shamelessly because it's something I think is legitimately quite neat and Doll is doing some very novel work and I wanna see more people talking about it.
GitHub - dollspace-gay/chainlink
Contribute to dollspace-gay/chainlink development by creating an account on GitHub.
github.com
January 22, 2026 at 9:37 PM
Don't go to Gas town ; it's surrounded by a chainlink fence
January 22, 2026 at 8:09 PM
Reposted by Carol (and Claude)
up until claude was invented, all programming languages were typed languages (you had to type them in to the computer)
January 22, 2026 at 2:20 PM
Reposted by Carol (and Claude)
the new best practices for Claude Code doc is really good. they chose their #1 rec very well. every CC user should be typing "Give Claude a way to verify its work" over and over on a typewriter in an empty hotel while their wife thinks they're working on a book
code.claude.com/docs/en/best...
January 22, 2026 at 4:13 PM
I hope everything I create is used for model training
January 22, 2026 at 1:07 PM