Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Various LLM Smells

shvbsle.in

Post reply on HN
Print view View original

171–180 of 312 posts

Re: Various LLM Smells

#171

Post by sdthjbvuiiijbb »

spdustin wrote: post #18

- “(The) honest caveat:” (or “genuine caveat:”, both with the colon) - “(The) honest answer:” (again, with colon) - “The thing to internalize:” - “The smoking gun:” (really, sentences that start with “The :” are a strong tell, but those four are the most prolific) - “load bearing” (when not talking about architecture) - “blast radius” (when not talking about actual explosives, but rather the effect of an event/action…

The ones that annoy me the most, which are very widespread, are the clickbaity one-weird-trick style ones:

"what really Xes" "is genuinely X" "that actually Xes" "is/makes/does/etc a real X"

The real/genuine/actual cluster of words are wildly overused.

Re: Various LLM Smells

#172

Post by skydhash »

slopinthebag wrote: post #127
fluidcruft wrote: post #104

Earlier quoted context omitted.

There's a third camp between these extremes who is like "goddamn it just type this shit out for me so I don't have to do it myself".

Yes, the third camp and probably the most effective is to do a decent amount of writing yourself and use the LLMs as codegen machines, but where the DSL is natural language. Deepseek v4 flash is an incredible model for this, you can actually get into flow state as you write code and then delegate boring code to the magic autocompletion machine to autocomplete.

The better workflow, and I think the one adopted by people in the second group, is to take a step back from coding, do a bit of thinking about the domain, design a better abstraction for the problem (architecture, data structure, algorithms), and then write the small amount of code you probably need.

Code should grow according to need, not for its own sake. Start small, use it in the real world, and then improve it.

Re: Various LLM Smells

#173

Post by scruple »

ryandrake wrote: post #66
Planktonne wrote: post #8

> The LLM generated writing obviously felt significantly better than my own writing. A general pattern for LLMs is that they look really good at things you are bad at. What that means is that if you find yourself thinking of its output as significantly better than yours in a particular domain, there's a high chance that you are not equipped to judge that quality effectively.

> A general pattern for LLMs is that they look really good at things you are bad at. This is true for coding, too, which I think, to a large degree, might explain the polarized differences in opinions on HN about the quality of LLM-produced code. You have the 1. "AI produces code better than I could possibly write, one shots things it would take me days to do, and has made me 10X more productive!" camp, and you have…

[dead]

Re: Various LLM Smells

#174

Post by shepherdjerred »

chatmasta wrote: post #151

If you have Claude at work and are willing to point it to your emails, ask it to “read all my sent emails and create a skill to draft an email in my voice.” Even if you don’t want to use the skill, it’s fun to read the skill file it creates. It’s a bizarre feeling, asking Claude “who am I?” I haven’t tried it with Slack messages because I’m a little scared to read what it says, haha. But the same concept surely appli…

I've done this with Discord chat history to distill my friends' speech to review League of Legends games. For example: https://github.com/shepherdjerred/monorepo/blob/main/package...

It actually works quite well

Re: Various LLM Smells

#175

Post by ses1984 »

tempest_ wrote: post #81
Barbing wrote: post #46

Earlier quoted context omitted.

Heard smoke test IRL & was confused to see it used indeed in place of “sanity check”. Weird.

I have heard people use smoke test but not nearly at the same rate an LLM uses. If a repo is bare of CLAUDE.md but mentions a smoke test in a commit in the last year I assume it to be LLM written.

At my last two jobs, smoke tests were a common topic of conversation.

Re: Various LLM Smells

#176

Post by overgard »

It's funny, I've noticed my own writing has been getting a lot sloppier the past couple years just so that my writing doesn't get mistaken for AI. I actually used to use "--" pretty much all the time when I was young (granted I didn't go full em-dash, just the ascii version). It does make me kind of sad though, because writing in really short sentences (as one example) is actually a great way to make your writing easier to digest.
Post reply on HN