Top
Best
New

Posted by Labo333 23 hours ago

Show HN: The load-bearing vocabulary of Claude(louisabraham.github.io)
459 points | 218 commentspage 5
glimshe 16 hours ago|
I had an idea for an experiment. Take a decent text, any one, and ask AI to rate it. Then patch that text by replacing words for ones that AI likes (honestly, load bearing etc) whenever possible and ask AI to rate it again to see how the evaluation changes.
ricardobeat 16 hours ago|
Claude usually rates its own writing very highly, which is a problem when trying to make it write differently.
ziml77 14 hours ago|||
I've tried to push AI to get rid of the AI-isms. And, despite being able to produce a skill file which described all the strange ways that AI "talk", the AI failed to actually make the output sound less like AI. Tried this with ChatGPT and Claude with similar results for both.
smj-edison 15 hours ago|||
Yeah, I once asked it to review its own writing with a subagent with a pretty strict writing guide, and it didn't see anything wrong, lol.
United857 5 hours ago||
Smoking gun should also be there as well.
tengbretson 16 hours ago||
Why are people getting so hung up on the "load-bearing assumption" turn of phrase that Claude uses? I get that it becomes cliche, but it is also a rather semantically dense way to communicate an idea that a lot of people run into.
zahlman 26 minutes ago||
Is "load-bearing" better than e.g. "critical" or "crucial" here? I'm not convinced.
cillian64 13 hours ago|||
Generally the claude jargon is valid jargon I’ve seen real people use, it just uses jargon so much more often than a human would. And because it has particular jargon it likes, and claude is widely used, you see so much of that particular jargon that you get sick of it.

It’s like having one coworker with a very particular writing style which is mildly annoying, but then it suddenly feels like half the internet was written by that one person and it becomes a lot more annoying.

tesin 16 hours ago|||
It might be, when used sparingly, but when it's the first sentence in a wall of text as Claude goes way off the deep end on a two page description it becomes the easiest readily available tell that you're about to be frustrated.

Humans are very good at pattern recognition - Claude is _incredibly_ repetitive in the way it starts to struggle to communicate. I think there's also a ton of overlap in the Jargon instead of Usefulness that developers see in annoying middle management/salespeople. Circle back, synergy blah blah.

I don't think the individual turns of phrase are inherently problematic - but the process is triggering.

beering 16 hours ago|||
ItMs because Claude sprinkles these words as flavoring without aiding understanding. It feels like Claude thinks of metaphors that don’t actually mean anything (or maybe only makes sense to itself).
tengbretson 16 hours ago||
I don't understand. Does the metaphor "load-bearing assumption" not actually mean anything to you?
ruuda 11 hours ago|||
"Load-bearing" is maybe not the best example, but Claude uses these anthropomorphized decorative synonyms everywhere. Claude's structs don't have fields, they _carry_ them. Its programs don't log, they _narrate_. Its arguments are not provided on the command line, they _ride_ it. Its modules don't depend on other modules, they _sit_ on them, etc.
stonogo 15 hours ago||||
It doesn't mean anything at all, since any assumption that is not 'load-bearing' is irrelevant and unlikely to be discussed at all. It's just noise.
applfanboysbgon 13 hours ago|||
Not when "load-bearing assumption" is misused, no. "Honey, can you take the load-bearing assumption to the mechanic for an oil change?" is an example injecting the phrase into a random sentence, obfuscating the meaning of the sentence and making it more difficult to parse. LLMs do this constantly. There is no theory of mind behind how words are generated. Some phrases simply have a higher chance of being generated in various contexts, even if it doesn't make any fucking sense, and this is exacerbated by bad RLHF.
rubslopes 12 hours ago||
That's the problem for me: "semantically dense". Sometimes a single paragraph is filled with several of this expression, one after the other, and there's a real mental effort to get what Claude is saying.
jfultz 6 hours ago||
Contrarian take.

So, yes, it's amusing to see clear Claude-isms like "load-bearing", "outright", and "genuine" in a [very nice] bit of analysis like this. And there's a (maybe negative? or not?) argument to be made about the world being filled with more Claude-isms or LLM-isms in general.

But I think the data say a second thing which is just as interesting and an absolute positive for the typical source code base. Look at the clusters that shrank significantly. Most of what you'll see in there is just incomprehensible...not even English. Cluster 4 has, after "pullrequest", a bunch of seeming usernames in the top tier. Cluster 6 seems to have names of repositories or tags in the top tier. Cluster 9 has branch names in it.

Meanwhile, keep going through cluster 1 and you'll see words I don't consider Claude-isms that really, really grow in usage. Words like "died", "nothing", "worse", "ever" all have well over 10x growth. This tells me something else. That the average commit log was BARELY ENGLISH. And then the LLMs came along and made commit logs that were ACTUALLY ENGLISH.

I count this as a good thing. I don't know the cross-section of repos chosen for this analysis, and I get it...some repos are garbage/throwaway, some commits come from automated processes that generate uninteresting commit logs, etc. But I've been benefitting from my work team's actually explanatory commit logs when doing code/bug archeology for decades, when doing PR review for the last decade, and I've even seen LLMs benefit from it in the last year (granted, not as often). A large part of professional software development is communication, and while the most important communication is via the code/comments, the commit logs are not unimportant. So, if this is making the average GitHub PR better (arguably more professional) by including actual English descriptions of code changes in commit logs...well, that's a genuinely load-bearing concept for me. :)

timtam 5 hours ago||
Why "eleven", though?
zahlman 23 minutes ago|
Programmers would write it as 11 (or maybe 0xB depending on context), whereas English style guides mandate using words for small numbers, and LLMs have presumably absorbed that advice. I imagine that up to "ten" the effect is not as strong.
b0ringdeveloper 16 hours ago||
I'd be interested to know what distinguishes the 7 pre-AI types. I wonder, are there certain buckets of 'Pre-AI' comments that we'd recognize?
Labo333 16 hours ago|
Sadly pretty uninteresting, you can probably just launch Claude code on the repo to see by yourself. One cluster is french and spanish, another is about design, another about frontend, etc...
bogdanoff_2 15 hours ago||
To all the people in the HN comments saying negative things about Claude: why don't you just use other models instead?
georgel 15 hours ago||
Work only allows the tools they buy. So at least for me, I am stuck with Claude. At home for personal projects I am a fan of DeepSeek v4 Flash.
rubslopes 11 hours ago|||
For me, there are two reasons: it’s what my work provides, and for some tasks, I want the absolute best model for coding.

Having said that, I just subscribed to ChatGPT yesterday, as I've become impatient with Claude for a text-dense project I'm working on.

threecheese 15 hours ago|||
Claude is all we have at $work, if you don’t count MS Copilot (and you shouldn’t count MS Copilot).

Imagine being “incentivized” to aggressively use a tool for your job, and that tool produces thousands of lines of text in Olde English which you need. You’d be griping too, methinks.

voidfunc 13 hours ago||
Copilot is fine? Just use it through whatever harness you want like OpenCode.
slashdave 13 hours ago||
Don't much care for the misdeeds of the alternative
sailfast 13 hours ago||
You’ve found the smoking foot gun! And that’s exactly the right question.
slashdave 13 hours ago||
No "smoking gun"?
eightnoteight 11 hours ago|
learning some vocabulary had been eye opening for me

like the stories behind when those words first appeared in the software engineering

like quiescence the most recent one i learned

More comments...