> For what it's worth, whatever LLMs do extensively, they do because it's a convention in well-established writing styles.
I think that's only part of the story. I think that while it's true what LLMs do is somehow represented in their corpus of training data, they also lack any understanding of how to adapt to the context, how to find a suitable "voice", and how not to overdo it, unless you explicitly prompt them otherwise, which is too much of a burden. Their default voice sucks, basically.
So let's say they learned to speak in Redditese. They don't know when not to speak in that voice. They always seem to be trying to make persuasive arguments, follow patterns of "It's not X. It's Y. And you know it (mic drop)." But real humans don't speak like this all the damn time. If you speak like this to your mom or to your closest friends, you're basically an idiot.
It's not that you cannot speak like this. It's that you cannot do it all the time. And that's the real problem with LLMs.
I agree with you and I find the term "creating content" awful, even though I'm forced to use it because it's something people immediately understand.
"Content creator"... what happened to artist, playwright, painter, hobbyist, etc? It makes it seem as if they were making stuff for a corporation to sell.
It is what's happening in some cases, not all. Also, language shapes thought, so we encourage this to happen if we frame it as "content creation". It's something to push against.
Note it's not even relevant whether something is commercial. Art can be commercial and not be just "content". A musician is not a "content creator" which happens to create content in the shape of music. "Content" implies it doesn't really matters, what matters is engagement and the platform (and advertisers, etc). It's not healthy to think of hobbies, art, and entertainment as exclusively about this. Imagine if Oscar Wilde, Herman Melville, Alan Moore, etc had been thought of merely as "content creators".
This is not a new idea. Stallman was already pushing back against this "content" term decades ago.
> I hated this movie the first time I watched it. And the second. The third time I let go of the need for things to be realistic and took it all in as an artistic representation and snap... I loved it.
I never managed to reach your third time. Once was enough for me, at the time, to decide it was an awful movie which didn't have anything to do with hackers or computers and which was terribly overacted, and that was that. Filed under yet another "Hollywood just doesn't get it", subsection "so bad it's embarrassing".
Much later I realized I had missed a cult classic. Oh well. I still think it's a bad movie, but I'm ok with other people loving it... maybe that's my growth moment.
I love it, but I know it's bad, but I also think it was intentionally what it is, which makes it good or even great.
If you can unlock that teenage feeling of wonder at the potential size and scope of the world and, at the right age at the right time, feeling like that world is your oyster, that's the feeling in which to watch this movie.
I refuse, however, to get into that feeling-zone for other 'high school' movies; they're stupid...
Hackers is weird in that the tech part of it is so obviously fake, but then you have the hacker culture part that goes to the point of actually quoting a large part of the "Hacker's Manifesto" in the movie.
Yeah, even when humans write in this artificial, punched-to-the-max, mic-drop style (as I've seen it described), there's a time and a place.
LLMs default to this style whether it makes sense or not. I don't write like this when chatting with my friends, even when I send them a long message, yet LLMs always default to this style, unless you tell them otherwise.
I think that's the tell. Always this style, always to the max, all the time.
Agreed. Also, a realistic assessment should not downplay the very real overhead and headache of managing your on-premise data center. It comes at a cost in engineering/firefighting hours, it's not painless. There's a reason this eternal ping pong keeps going on!
I think that's only part of the story. I think that while it's true what LLMs do is somehow represented in their corpus of training data, they also lack any understanding of how to adapt to the context, how to find a suitable "voice", and how not to overdo it, unless you explicitly prompt them otherwise, which is too much of a burden. Their default voice sucks, basically.
So let's say they learned to speak in Redditese. They don't know when not to speak in that voice. They always seem to be trying to make persuasive arguments, follow patterns of "It's not X. It's Y. And you know it (mic drop)." But real humans don't speak like this all the damn time. If you speak like this to your mom or to your closest friends, you're basically an idiot.
It's not that you cannot speak like this. It's that you cannot do it all the time. And that's the real problem with LLMs.
(Sorry, couldn't resist!)
reply