There's a guy I work closely with that runs things he sends me through an LLM. When he messages me privately, it has very poor grammar and no punctuation. The LLM assisted text is always verbose...
There's a guy I work closely with that runs things he sends me through an LLM. When he messages me privately, it has very poor grammar and no punctuation. The LLM assisted text is always verbose and not easy to read. Very annoying!
I think too many people have forgotten why we use grammar and punctuation. It's so associated with formality and attempts to align with social pressures that many make terrible decisions about...
I think too many people have forgotten why we use grammar and punctuation. It's so associated with formality and attempts to align with social pressures that many make terrible decisions about what to include and when.
There's a reason this perception has taken hold: plenty of our writing "rules" are arbitrary and exist mostly as a way to signal your education/familiarity with writing rules. Capitalizing proper nouns or correctly punctuating before a quote doesn't really help me communicate when I'm just messaging a friend or colleague.
However, I think many have thrown the baby out with the bathwater here and decided that "casual" means "no punctuation". I very much want anyone reading something I've written (formal or otherwise) to interpret it the same way I imagined it when writing, and punctuation is a fantastic way to aid that attempt.
noise: as always writing about grammar on the internet, I am sufficiently paranoid that I made some silly errors here... but hey, if you understood, then it worked!
When I use a messaging app, I find myself sending only 1 or 2 sentences per message. Especially if I'm having a conversation with someone. I also tend to not bother capitalizing or using periods....
When I use a messaging app, I find myself sending only 1 or 2 sentences per message. Especially if I'm having a conversation with someone. I also tend to not bother capitalizing or using periods. But I'll still use punctuation like the apostrophe or commas.
I'm with you - I generally try to be as clear as possible when I'm writing something. Generally, I mean exactly what I say. It's often hard to press submit because I don't want someone to misinterpret me.
I've never really thought about it but I definitely use "new chat message" as a way to communicate a sense of pacing to conversation. It's like a line/stanza break in poetry.
I've never really thought about it but I definitely use "new chat message" as a way to communicate a sense of pacing to conversation. It's like a line/stanza break in poetry.
I also have a coworker like this. I recently got re-orged so I no longer have to deal with it but it was infuriating working with them. They're a brilliant engineer in their own right but...
I also have a coworker like this. I recently got re-orged so I no longer have to deal with it but it was infuriating working with them. They're a brilliant engineer in their own right but literally everything they'd do was always through an LLM.
I 100% agree. This has happened to me many, many times. I ask someone who works for me, or with me a question, and I receive obviously AI generated output back. Or I recieve a question very...
I 100% agree. This has happened to me many, many times. I ask someone who works for me, or with me a question, and I receive obviously AI generated output back. Or I recieve a question very obviously generated with an AI prompt.
A few thoughts almost always come to my head:
You spent no time creating this. Why do you now expect me to spend time reading this.
But also
Do you think I'm stupid, because you don't think I wouldnt be able to detect that this extremely obviously AI generated content is in fact, AI generated, and somehow will think this is something you wrote?
Do you think I'm ignorant? As in, I don't even know that AI exists? Because if I did know that AI exists, in the time it took me to formulate the request to you, I could have also typed chatgpt.com into my browser and pasted my question. Do they think I haven't thought of that? LLMs aren't some magical unknown power that only you have access to.
It sort of feels like back in the day when you would ask a question online, and someone would reply back to you with a let me Google that for you link. Except you'd never do that to someone you work with. Sending AI slop responses at work has become endemic though.
The policy where I work is that you do not label AI generated content for internal work, which I'm in agreement with. The premise is that as long as it's not going out to the public, work is work,...
The policy where I work is that you do not label AI generated content for internal work, which I'm in agreement with. The premise is that as long as it's not going out to the public, work is work, and you're responsible for whatever you send out. Then you don't get the situation where people are sending unreviewed documents and pawning the responsibility off on AI. Well, they still might be unreviewed, but they have your name on it. I know some people feel more strongly about disclosing AI use wherever it's used, but if the premise is that people use AI often in their work, the disclosure would just end up on everything anyway.
I'm not understanding where your comment is coming from. It seems like you're implying the opposite of what I've said. The accountability is the normal accountability for work submitted by people....
I'm not understanding where your comment is coming from. It seems like you're implying the opposite of what I've said.
The accountability is the normal accountability for work submitted by people. There's no "AI screwed up" excuse because AI is not allowed to be the author responsible for the work, just people.
I interpreted the comment to mean "AI is a default at the company. Treat mishaps with it as you would a fax machine breaking down". But now that "fax machine" can be leveraged an an issue for any...
AI is not allowed to be the author responsible for the work, just people.
I interpreted the comment to mean "AI is a default at the company. Treat mishaps with it as you would a fax machine breaking down". But now that "fax machine" can be leveraged an an issue for any and all communication (regardless of the truth).
I came to that conclusion based on
but if the premise is that people use AI often in their work, the disclosure would just end up on everything anyway.
part of your comment. I'd personally would still want disclosure so I can get context on the nature of the communication. But others may think differently.
An outage is a reasonable excuse for a delay but not for shoddy work. I think the more appropriate analogy would be if the fax machine garbled some text, you didn't bother checking and asking for...
An outage is a reasonable excuse for a delay but not for shoddy work. I think the more appropriate analogy would be if the fax machine garbled some text, you didn't bother checking and asking for a resend, then you submit the report with garbled text to your boss. Sure, it's the machine that gave bad output, but it's your fault for pushing it forward without fixing it. "It's the fax machine's fault" is a poor excuse in this situation.
Yeah, there's a big difference between a reactive "it's the tool's fault" defense after getting called out on garbage results or blown deadlines, and a proactive "the tool we use at the company...
Yeah, there's a big difference between a reactive "it's the tool's fault" defense after getting called out on garbage results or blown deadlines, and a proactive "the tool we use at the company for this process is [unusable/not producing good results] for [already investigated reason]. We can wait for it to be fixed, or we can instead..." or just figuring an alternative solution out yourself without the broken tool to meet the agreed upon deadlines if the tool usage is your personal choice.
Not directly related to the topic it's posted in, but I've felt the exact same way about AI written emails. I don't begrudge people saving time to write longer text - I initially even supported...
Not directly related to the topic it's posted in, but I've felt the exact same way about AI written emails. I don't begrudge people saving time to write longer text - I initially even supported folks doing it! - but the loss of personal voice and clear failure to proofread outputs its really starting to grate on me these days.
There's a guy I work closely with that runs things he sends me through an LLM. When he messages me privately, it has very poor grammar and no punctuation. The LLM assisted text is always verbose and not easy to read. Very annoying!
I think too many people have forgotten why we use grammar and punctuation. It's so associated with formality and attempts to align with social pressures that many make terrible decisions about what to include and when.
There's a reason this perception has taken hold: plenty of our writing "rules" are arbitrary and exist mostly as a way to signal your education/familiarity with writing rules. Capitalizing proper nouns or correctly punctuating before a quote doesn't really help me communicate when I'm just messaging a friend or colleague.
However, I think many have thrown the baby out with the bathwater here and decided that "casual" means "no punctuation". I very much want anyone reading something I've written (formal or otherwise) to interpret it the same way I imagined it when writing, and punctuation is a fantastic way to aid that attempt.
noise: as always writing about grammar on the internet, I am sufficiently paranoid that I made some silly errors here... but hey, if you understood, then it worked!
When I use a messaging app, I find myself sending only 1 or 2 sentences per message. Especially if I'm having a conversation with someone. I also tend to not bother capitalizing or using periods. But I'll still use punctuation like the apostrophe or commas.
I'm with you - I generally try to be as clear as possible when I'm writing something. Generally, I mean exactly what I say. It's often hard to press submit because I don't want someone to misinterpret me.
I've never really thought about it but I definitely use "new chat message" as a way to communicate a sense of pacing to conversation. It's like a line/stanza break in poetry.
I also have a coworker like this. I recently got re-orged so I no longer have to deal with it but it was infuriating working with them. They're a brilliant engineer in their own right but literally everything they'd do was always through an LLM.
I 100% agree. This has happened to me many, many times. I ask someone who works for me, or with me a question, and I receive obviously AI generated output back. Or I recieve a question very obviously generated with an AI prompt.
A few thoughts almost always come to my head:
But also
Do you think I'm stupid, because you don't think I wouldnt be able to detect that this extremely obviously AI generated content is in fact, AI generated, and somehow will think this is something you wrote?
Do you think I'm ignorant? As in, I don't even know that AI exists? Because if I did know that AI exists, in the time it took me to formulate the request to you, I could have also typed chatgpt.com into my browser and pasted my question. Do they think I haven't thought of that? LLMs aren't some magical unknown power that only you have access to.
It sort of feels like back in the day when you would ask a question online, and someone would reply back to you with a let me Google that for you link. Except you'd never do that to someone you work with. Sending AI slop responses at work has become endemic though.
The policy where I work is that you do not label AI generated content for internal work, which I'm in agreement with. The premise is that as long as it's not going out to the public, work is work, and you're responsible for whatever you send out. Then you don't get the situation where people are sending unreviewed documents and pawning the responsibility off on AI. Well, they still might be unreviewed, but they have your name on it. I know some people feel more strongly about disclosing AI use wherever it's used, but if the premise is that people use AI often in their work, the disclosure would just end up on everything anyway.
And is there any accountability for if/when documents go unreviewed? Or is "the AI screwed up" enough to dispel that?
I'm not understanding where your comment is coming from. It seems like you're implying the opposite of what I've said.
The accountability is the normal accountability for work submitted by people. There's no "AI screwed up" excuse because AI is not allowed to be the author responsible for the work, just people.
I interpreted the comment to mean "AI is a default at the company. Treat mishaps with it as you would a fax machine breaking down". But now that "fax machine" can be leveraged an an issue for any and all communication (regardless of the truth).
I came to that conclusion based on
part of your comment. I'd personally would still want disclosure so I can get context on the nature of the communication. But others may think differently.
An outage is a reasonable excuse for a delay but not for shoddy work. I think the more appropriate analogy would be if the fax machine garbled some text, you didn't bother checking and asking for a resend, then you submit the report with garbled text to your boss. Sure, it's the machine that gave bad output, but it's your fault for pushing it forward without fixing it. "It's the fax machine's fault" is a poor excuse in this situation.
Yeah, there's a big difference between a reactive "it's the tool's fault" defense after getting called out on garbage results or blown deadlines, and a proactive "the tool we use at the company for this process is [unusable/not producing good results] for [already investigated reason]. We can wait for it to be fixed, or we can instead..." or just figuring an alternative solution out yourself without the broken tool to meet the agreed upon deadlines if the tool usage is your personal choice.
Not directly related to the topic it's posted in, but I've felt the exact same way about AI written emails. I don't begrudge people saving time to write longer text - I initially even supported folks doing it! - but the loss of personal voice and clear failure to proofread outputs its really starting to grate on me these days.
So yeah, completely agree. Great maxim.
https://lobste.rs/s/1yjftk/if_you_are_asking_for_human_attention