I agree with the author. Writing serves many purposes, and there is a continuum in how much humanity needs to be present in the text. A eulogy is not the same as a git commit.
I’m hoping he’s right that this is a passing moral panic, because the persecution of “AI smell” is exhausting. If it works, it works, right? Do we have to be so precious about how we got there?
> When the purpose is conveying personal thoughts, AI should be used lightly. When the purpose is to convey information, it can be used extensively.
I would argue that this is very wrong.
I think it's wrong because when we write, even for purely information dissemination, people still feel a lack of authenticity in obvious AI writing. It's also true that AI writing tends to be verbose, and over explain itself. When people talk about 'taste' in writing, sure some of it is the word choice, but a lot of it if taste in choosing what information is important to the audience, what to include or leave out, what can be assumed as shared 'tribal' knowledge that an AI doesn't factor for, and so ofter over-explains stuff that's a shared common knowledge.
This desire is what leads to verbosity.
Good writing is dense and concise, transmitting an idea or information efficiently. Virtually all AI writing fails that hurdle and does not try to be concise even when prompted to do so.
What does work is giving AI examples in context of past iterations of the same piece of work i.e. release announcements etc. that your LLM can copy. But it's still not perfect.
I agree with the author. Writing serves many purposes, and there is a continuum in how much humanity needs to be present in the text. A eulogy is not the same as a git commit.
I’m hoping he’s right that this is a passing moral panic, because the persecution of “AI smell” is exhausting. If it works, it works, right? Do we have to be so precious about how we got there?
> When the purpose is conveying personal thoughts, AI should be used lightly. When the purpose is to convey information, it can be used extensively.
I would argue that this is very wrong.
I think it's wrong because when we write, even for purely information dissemination, people still feel a lack of authenticity in obvious AI writing. It's also true that AI writing tends to be verbose, and over explain itself. When people talk about 'taste' in writing, sure some of it is the word choice, but a lot of it if taste in choosing what information is important to the audience, what to include or leave out, what can be assumed as shared 'tribal' knowledge that an AI doesn't factor for, and so ofter over-explains stuff that's a shared common knowledge.
This desire is what leads to verbosity.
Good writing is dense and concise, transmitting an idea or information efficiently. Virtually all AI writing fails that hurdle and does not try to be concise even when prompted to do so.
What does work is giving AI examples in context of past iterations of the same piece of work i.e. release announcements etc. that your LLM can copy. But it's still not perfect.
Says the author who didn’t learn to write.
Did you read the blog? I found it well written and easy to follow.