The best way to lose any and all respect I might have had for you as an engineer is to reply to my code review feedback with an AI generated wall of text.
-
@BigShellEvent @faithisleaping I yell "maybe training it all on reddit *wasn't* the best idea in the world" at whatever LLM driven piece of tech is driving me batty this week and then go cry quietly in the corner for a bit
@mxjaygrant @BigShellEvent Hell, even training it on “reliable” sites like StackOverflow wasn’t a great idea. There’s a lot of absolute bullshit on that site, even if you weight your training based on what the users said were good answers.
-
@mxjaygrant @BigShellEvent Hell, even training it on “reliable” sites like StackOverflow wasn’t a great idea. There’s a lot of absolute bullshit on that site, even if you weight your training based on what the users said were good answers.
@faithisleaping @BigShellEvent oh yeah, for every time it helped me, StackOverflow merrily sent me down the wrong path 75% of the time
... I kinda miss it tho? A billion primates banging on electric typewriters at any given time was really sort of brilliant there for a moment
-
But go ahead, dude. Keep vibing. I’m going to remember all these MRs and this is what I’m going to cite when I roll out stronger AI policies in the parts of the project I maintain. Someone’s got to be the reason the rules for added. So, really, you’re doing us all a service. 🫡
@faithisleaping Thank you for holding the line 🫡 (just when I thought I couldn’t love you more 🥰
) -
First off, there’s the fact that you’re using the damn thing in the first place. That’s really not the flex you think it is.
Second, it clearly communicates to me that you don’t respect me. I put my own human time into looking at the code and giving feedback. You need to match energy. If you just give me a pile of LLM spew, that means you didn’t value my time and energy enough to respond in kind. This is basic human respect.
Third, it’s not Artificial Intelligence, it’s Automated Mansplaining. I’m an expert in my field. Mansplaining at me about my own field AND being obviously wrong about it is REALLY not the look you’re going for. (Like, the LLM “explanation” I got thrown at me this morning wasn’t even plausible. It contradicted itself in the first two sentences. It was 8 paragraphs.)
@faithisleaping I love you so much for introducing me to the phrase “automated mansplaining”! You nailed it!
-
First off, there’s the fact that you’re using the damn thing in the first place. That’s really not the flex you think it is.
Second, it clearly communicates to me that you don’t respect me. I put my own human time into looking at the code and giving feedback. You need to match energy. If you just give me a pile of LLM spew, that means you didn’t value my time and energy enough to respond in kind. This is basic human respect.
Third, it’s not Artificial Intelligence, it’s Automated Mansplaining. I’m an expert in my field. Mansplaining at me about my own field AND being obviously wrong about it is REALLY not the look you’re going for. (Like, the LLM “explanation” I got thrown at me this morning wasn’t even plausible. It contradicted itself in the first two sentences. It was 8 paragraphs.)
@faithisleaping saw this elsewhere, and it's oversimplifying what you've so eloquently explained, but it's succinct enough that those who would hand you a wall of slop text might read it: "I can't be bothered to read what you couldn't be bothered to write."
-
@faithisleaping
"I asked Claude--"
"Well _I_ asked _you_. I could have asked Claude myself and skipped you altogether."@tankgrrl @faithisleaping
I go to human SMEs for specialized situations where I want to make sure there's not something I'm missing or misunderstanding. Or sometimes I just want a sanity check on something.And then I'll get something clearly slopified, often contradicting itself, getting lost in a subquestion, simply restating all the information in my question as an answer, and speaking in ways that only LLMs do. And when I say "sure,I already said most of that, but <repeat core question>" and get a "You're absolutely right..." I just mentally discard that conversation as lacking actual value and have a bigger sense of unease at the lack of quality and validation checks for big impactful processes.
It's pretty demotivating, tbh.
-
@tankgrrl @faithisleaping
I go to human SMEs for specialized situations where I want to make sure there's not something I'm missing or misunderstanding. Or sometimes I just want a sanity check on something.And then I'll get something clearly slopified, often contradicting itself, getting lost in a subquestion, simply restating all the information in my question as an answer, and speaking in ways that only LLMs do. And when I say "sure,I already said most of that, but <repeat core question>" and get a "You're absolutely right..." I just mentally discard that conversation as lacking actual value and have a bigger sense of unease at the lack of quality and validation checks for big impactful processes.
It's pretty demotivating, tbh.
@h3mmy @faithisleaping
My new go to is "I asked for YOUR expertise. Is your expertise asking Claude?" -
The best way to lose any and all respect I might have had for you as an engineer is to reply to my code review feedback with an AI generated wall of text.
@faithisleaping Yup. The moment I get an obviously-wrong LLM response, the xerox machine that shat it into my inbox finds out what the bottom of my priorities list looks like.
There are crickets there.
The crickets don't like them either.
-
First off, there’s the fact that you’re using the damn thing in the first place. That’s really not the flex you think it is.
Second, it clearly communicates to me that you don’t respect me. I put my own human time into looking at the code and giving feedback. You need to match energy. If you just give me a pile of LLM spew, that means you didn’t value my time and energy enough to respond in kind. This is basic human respect.
Third, it’s not Artificial Intelligence, it’s Automated Mansplaining. I’m an expert in my field. Mansplaining at me about my own field AND being obviously wrong about it is REALLY not the look you’re going for. (Like, the LLM “explanation” I got thrown at me this morning wasn’t even plausible. It contradicted itself in the first two sentences. It was 8 paragraphs.)
@faithisleaping "Automated Mansplaining" is genius right there!
-
@faithisleaping @BigShellEvent oh yeah, for every time it helped me, StackOverflow merrily sent me down the wrong path 75% of the time
... I kinda miss it tho? A billion primates banging on electric typewriters at any given time was really sort of brilliant there for a moment
I used SO recently to recover a partially transmitted zip file that contained only one text file in it.
My experience has been there are some really dedicated users, and the longer the time differential between the question and a detailed answer the better the quality.
Dude dropped a 5 line perl script that recovered 5000 json records from a file unzip would just cry about no matter what arcane set of flags I used.
-
@faithisleaping Yup. The moment I get an obviously-wrong LLM response, the xerox machine that shat it into my inbox finds out what the bottom of my priorities list looks like.
There are crickets there.
The crickets don't like them either.
@drsbaitso @faithisleaping It’s impossible to hear this and not imagine those crickets in a most wretched hive of scum and villainy…
-
@drsbaitso @faithisleaping It’s impossible to hear this and not imagine those crickets in a most wretched hive of scum and villainy…
@transpirateradio @faithisleaping I do have a disconnect sentence on twelve networks.
-
J jwcph@helvede.net shared this topic
