4 ms·

That mail looks like it was written by some LLM....I wouldn't take it seriously.

drbawb 2y ago
I wouldn't be so sure about that. If you had said this before my most recent hire I would have agreed - but my junior writes stuff that looks like this or worse. I'm always reviewing stuff with weird capitalization, punctuation, spelling, grammar, etc. that makes me raise an eyebrow on a regular basis. Their code looks just as weird despite having the help of a heavy-weight IDE in their corner. We're constantly cleaning up hanging indents, unnecessary newlines, mismatched indents, commented out code, etc. (That's with a college education, a Grammarly license, and no small amount of coaching on my part!) I have a sample size of 1, so I can't ascribe too much to "these damn kids," but it seriously strikes me as having learned written language primarily from texting & instant messaging. Whereas I grew up roughly by transitioning from: reading books -> writing mails to pen pals-> writing e-mails -> web chats -> T9 texting -> modern IMEs. In other words I initially learned to write with long-form content and learned to condense it down later. These days I think people are just learning straight from the condensed version. The other reason I don't think it's an LLM is simpler: most commercial LLMs wouldn't be "aligned" to be that rude, and the smaller LLMs I've seen wouldn't be able to inject relevant code snippets from a relatively unpopular library into the output. I would not be surprised if this person misused the library, got called out for it in code-review (calling the iterator multiple times is a huge code-smell), and now they are soothing their ego by shifting blame onto the library author for making "such a bad API."
- mrkramer 2y ago
  Individual who sent that mail has trollish name, trollish email account and s/he is talking gibberish. My conclusion was, give it a pass and move on. >I would not be surprised if this person misused the library, got called out for it in code-review (calling the iterator multiple times is a huge code-smell), and now they are soothing their ego by shifting blame onto the library author for making "such a bad API." It could be that but then again it's his or her fault not the maintainer's. At the end of the day, s/he has some serious anger control issues if that's true. >The other reason I don't think it's an LLM is simpler: most commercial LLMs wouldn't be "aligned" to be that rude, and the smaller LLMs I've seen wouldn't be able to inject relevant code snippets from a relatively unpopular library into the output. You can modify some open source LLM to talk trash, meaning teach it to hate and disrespect.

I wouldn't be so sure about that. If you had said this before my most recent hire I would have agreed - but my junior writes stuff that looks like this or worse. I'm always reviewing stuff with weird capitalization, punctuation, spelling, grammar, etc. that makes me raise an eyebrow on a regular basis. Their code looks just as weird despite having the help of a heavy-weight IDE in their corner. We're constantly cleaning up hanging indents, unnecessary newlines, mismatched indents, commented out code, etc. (That's with a college education, a Grammarly license, and no small amount of coaching on my part!) I have a sample size of 1, so I can't ascribe too much to "these damn kids," but it seriously strikes me as having learned written language primarily from texting & instant messaging. Whereas I grew up roughly by transitioning from: reading books -> writing mails to pen pals-> writing e-mails -> web chats -> T9 texting -> modern IMEs. In other words I initially learned to write with long-form content and learned to condense it down later. These days I think people are just learning straight from the condensed version. The other reason I don't think it's an LLM is simpler: most commercial LLMs wouldn't be "aligned" to be that rude, and the smaller LLMs I've seen wouldn't be able to inject relevant code snippets from a relatively unpopular library into the output. I would not be surprised if this person misused the library, got called out for it in code-review (calling the iterator multiple times is a huge code-smell), and now they are soothing their ego by shifting blame onto the library author for making "such a bad API."

mrkramer 2y ago
Individual who sent that mail has trollish name, trollish email account and s/he is talking gibberish. My conclusion was, give it a pass and move on. >I would not be surprised if this person misused the library, got called out for it in code-review (calling the iterator multiple times is a huge code-smell), and now they are soothing their ego by shifting blame onto the library author for making "such a bad API." It could be that but then again it's his or her fault not the maintainer's. At the end of the day, s/he has some serious anger control issues if that's true. >The other reason I don't think it's an LLM is simpler: most commercial LLMs wouldn't be "aligned" to be that rude, and the smaller LLMs I've seen wouldn't be able to inject relevant code snippets from a relatively unpopular library into the output. You can modify some open source LLM to talk trash, meaning teach it to hate and disrespect.