I'll try this out.
Just find it amusing that a skill file has
* A repo with 61 files
* 80% python
I-have-ADHD: A skill to stop coding agents from burying the answer
331–340 of 390 posts
Re: I-have-ADHD: A skill to stop coding agents from burying the answer
#332[flagged]
Re: I-have-ADHD: A skill to stop coding agents from burying the answer
#333[flagged]
Re: I-have-ADHD: A skill to stop coding agents from burying the answer
#334Earlier quoted context omitted.
I've had better success with having it write and then rewrite rather than trying to prompt the first write into following my style rules.
also by abusing linters and other guardrails you can basically automate the rewrite
I mean that’s not abuse? That’s what mechanically checkable invariants are for! They constrain the space the model has to work in.
Re: I-have-ADHD: A skill to stop coding agents from burying the answer
#335The biggest "Claudism" that I have a hard time getting the LLM to stop doing is its insistence on talking about what it didn't do in addition to what it did. "I edited this.py and that.py but I did not edit README.md and I did not commit." or code comments like "This code invokes foo on bar and returns the results directly -- not through a callback." "This code returns true if the user clicked on a button -- not on t…
Grok does the same thing.
We'll discuss a feature implementation with various options for design, settle on one of them, and then it will write in the doc comment all the designs we considered but dropped.
Re: I-have-ADHD: A skill to stop coding agents from burying the answer
#336The biggest "Claudism" that I have a hard time getting the LLM to stop doing is its insistence on talking about what it didn't do in addition to what it did. "I edited this.py and that.py but I did not edit README.md and I did not commit." or code comments like "This code invokes foo on bar and returns the results directly -- not through a callback." "This code returns true if the user clicked on a button -- not on t…
It’s a model issue. I’m in the process of switching my company’s primary AI provider after several days of testing Astra.
Even Fable feels like an idiot now. It’s not the code quality, it’s the improvements in communication and judgement. It is an absolute breath of fresh air. I was spending a lot of tokens and building special workflows to reign in Claude’s horrendous prose.
Astra just communicates well out of the box!!!
Codex has worse UX, but Astra has fewer qualms about building you a custom harness overlay.
Re: I-have-ADHD: A skill to stop coding agents from burying the answer
#337Seems everyone on Hacker News has already forgotten about Anthropic’s amazing new watermarking algorithm. You don’t need to look any further to understand why it writes so awfully - it literally swaps words programmatically all over the text from the result of a hash function…
This was an issue before watermarking was publically announced. I'm also not an expert, but aren't bits of a hash indistinguishible from uniform randomness?
Depends on the particular hash. Some hashes are designed to produce similar outputs when the input data is similar (for example, SSDeep). I'd imagine hashes designed to survive light copywriting would also not be entirely indistinguishable from random.
Re: I-have-ADHD: A skill to stop coding agents from burying the answer
#338Earlier quoted context omitted.
Relevant anecdata because I've burned many a Claude sessions on this. If you're using Claude Code, then it's in the harness. At the close of many sessions, I would start a meta conversation over why the LLM would consistently break certain rules. What it found when debugging itself is that some of the "contradicting" rules that I had were in fact, not from my rules. Instead, the instructions from its own harness had…
You can't trust it's account on why it did something, it does not "remember". It will just make up something plausible sounding.
I know this is a popular position and it makes sense at face value when you think of LLMs as autocomplete systems. But it’s genuinely wrong.
Relevant reading is most notably anthropic’s research on the J-space. LLMs will plan ahead of time helped with CoT, get to a plan and “store” it in j-space, and execute on that plan which means they can in fact “backtrack” and give you reasoning on why they did something, because it IS part of their state.
Re: I-have-ADHD: A skill to stop coding agents from burying the answer
#339> 3. End with one concrete next step.
I mean maybe I have ADHD idk i'm not diagnosed, I don't generally have issues understanding Claude's output (even though I do think it's unnecessarily verbose), but I think this one is a general good advice which Claude ignores for some reason, and is especially annoying during debugging sessions, where it asks for 3 outputs, I paste one for it to process while getting the others, and then it immediately forgets it wanted the 2 others, and starts cooking up some wrong conclusions...
Re: I-have-ADHD: A skill to stop coding agents from burying the answer
#340In the example, I notice that only the original answer explain what is the issue. In this specific case it may be trivial, but I prefer to read something too verbose that explain things than just following instructions blindly (especially if it's for telling twice in a row to open the same file at the same line :D.