Writing is fundamentally the transfer of information from your brain to my brain. If you have 1000 bits of semantic information you want to transfer, you can't give 300 bits of semantic information to an LLM and have it fill in the remaining 700, because it doesn't know what those 700 bits are. If it's able to guess those 700 bits correctly, then they aren't true semantic information, and you really only have 300 bits you want to transfer. You might as well transfer those bits to me directly, rather than having the LLM add on an extra superfluous 700 bits that I then have to filter out.
It isn't the transfer of information at all. What's actually happening is you're prompting experiences in a human instead of an AI using text.
Communication only works if you have multiple levels of representation and abstraction, including but not limited to - letter shapes, grammatical structures, style and register, stylometry, and subtext.
All of that is learned, and writers usually assume they can rely on that learning as the context for the text.
So you don't write to 'transfer information' like a network cable, you write to trigger experiences in the human version of latent space.
Factual information is one kind of experience. But even when that's the goal, there are always layers of implied relationship, social register, role, status, and other implications in everything that's written.
In normal communications the context - business emails, personal messages, mainstream journalism, fiction, and the rest - defines what acceptable language looks like.
The content fits inside that. But it has to fit the context, otherwise it lands in a semantic and psychological uncanny valley - like sending LinkedIn speak to a spouse on a wedding anniversary.
The real problem with LLM writing is that it's good at the technical layer - the grammar and spelling - and has some insights into the rest.
But the default content style is marketing and ad speak. And recently it's developed a weird and unique hybrid style which applies marketing fluff and pretension to technical content like code comments.
So you get one register instead of all of them. It can attempt others, but it's still too limited to generate them fluently. Sometimes the results are outstanding, but often it defaults to mechanical clichés.
So that's why it sucks and sounds so hollow.
Can it be fixed? Yes, but it's very hard work, most people don't have the skills, and it takes time - often too much time to be worth the effort.
That "marketing style" isn't a style, it's the lack of content itself. You just restated their original point despite trying to object to it, because it was correct and there is no way around that.
Sometimes Claude's problem, such as when I ask it to summarize a long, complex session back to me, is it's too information dense. It uses weird invented terms to gloss over complex parts of the architecture instead of explaining them.
But no matter what - too dense or too sparse - it always sounds like Claude.
It misses the forest for the trees. It feels the need to highlight details not understanding what details are most relevant to a human reader and how to survey the larger problems in a cohesive way that emphasizes the right parts without cliche and undue emphasis.
It's never too dense. There are only more words but not more information.
If I describe all of the individual muscle contractions and joint motions required to walk across a room, it's hundreds of pages of data to say almost nothing. That is the exact opposite of dense.
By contrast a poet can deliver many concepts and many layers and even practically a fractal choose-your-own-adventure in only a dozen words. That is what dense is.
hatthew · · focus · HN ↗
Writing is fundamentally the transfer of information from your brain to my brain. If you have 1000 bits of semantic information you want to transfer, you can't give 300 bits of semantic information to an LLM and have it fill in the remaining 700, because it doesn't know what those 700 bits are. If it's able to guess those 700 bits correctly, then they aren't true semantic information, and you really only have 300 bits you want to transfer. You might as well transfer those bits to me directly, rather than having the LLM add on an extra superfluous 700 bits that I then have to filter out.
TheOtherHobbes · · focus · HN ↗
Communication only works if you have multiple levels of representation and abstraction, including but not limited to - letter shapes, grammatical structures, style and register, stylometry, and subtext.
All of that is learned, and writers usually assume they can rely on that learning as the context for the text.
So you don't write to 'transfer information' like a network cable, you write to trigger experiences in the human version of latent space.
Factual information is one kind of experience. But even when that's the goal, there are always layers of implied relationship, social register, role, status, and other implications in everything that's written.
In normal communications the context - business emails, personal messages, mainstream journalism, fiction, and the rest - defines what acceptable language looks like.
The content fits inside that. But it has to fit the context, otherwise it lands in a semantic and psychological uncanny valley - like sending LinkedIn speak to a spouse on a wedding anniversary.
The real problem with LLM writing is that it's good at the technical layer - the grammar and spelling - and has some insights into the rest.
But the default content style is marketing and ad speak. And recently it's developed a weird and unique hybrid style which applies marketing fluff and pretension to technical content like code comments.
So you get one register instead of all of them. It can attempt others, but it's still too limited to generate them fluently. Sometimes the results are outstanding, but often it defaults to mechanical clichés.
So that's why it sucks and sounds so hollow.
Can it be fixed? Yes, but it's very hard work, most people don't have the skills, and it takes time - often too much time to be worth the effort.
Brian_K_White · · focus · HN ↗
howunfortunate · · focus · HN ↗
Sometimes Claude's problem, such as when I ask it to summarize a long, complex session back to me, is it's too information dense. It uses weird invented terms to gloss over complex parts of the architecture instead of explaining them.
But no matter what - too dense or too sparse - it always sounds like Claude.
therealdrag0 · · focus · HN ↗
Brian_K_White · · focus · HN ↗
If I describe all of the individual muscle contractions and joint motions required to walk across a room, it's hundreds of pages of data to say almost nothing. That is the exact opposite of dense.
By contrast a poet can deliver many concepts and many layers and even practically a fractal choose-your-own-adventure in only a dozen words. That is what dense is.