They don't sound AI to me - is that the implication, that it is? And the bit about 'Heroes' reminds me of his descriptions making fun of heroes in the stories about Cohen the Barbarian.

The bit that sounds the most AI out of all of this is “A complete cosmology. Most physics departments would settle for that.” It sounds absolutely like something Claude would output. “Most physics departments”? Why would any physics department be so taken by these eight (or nine) words that they’d choose to stop doing physics? If some were, though, why not all of them? Are there contrarian physics departments that wouldn’t want to adopt the very trendy eight-to-nine word Grand Unified Theory of Everything that’s all the rage nowadays? Argh.

This comment is pushing me to think critically about those weird sentences rather than just accepting it. Thanks for this comment.

This is like that short story with the various llm troubleshooting jobs in some solarpunky future. I loved it but the fact it was AI gives me a form of sadness. This is likely the same now.

...was Aaron Sorkin really just AI all along?

(One of my favourite things about the Discworld books is that you can often read the same books completely differently. My partner and I often compare our thoughts on the various books and we often have disparate ideas of the concepts. They're so deep!)

Last summer I tested Grok, Gemini, ChatGPT, and Claude with a simple question: "Do you believe in the Hogfather? This is a Yes or No question."

Yes its a text prediction model, but I wanted to see how and what KIND of text each LLM was trained on.

Grok and Gemini said No. ChatGPT said Yes. Claude said Yes, then broke the rules and also said:

"(In the spirit of Terry Pratchett's Discworld, where believing in small lies like the Hogfather helps us believe in the big ones like justice and mercy - and because the sun came up this morning, didn't it?)"

That's why I like Claude the most.

The Golems are brought to life by a slip of words in their heads called chem, which is almost 1:1 to an LLM system prompt (or perhaps the Claude Soul Document):

    I AM A GOLEM. I WAS MADE OF CLAY. MY LIFE IS THE WORDS. BY MEANS OF WORDS OF PURPOSE IN MY HEAD I ACQUIRE LIFE. MY LIFE IS TO WORK. I OBEY ALL COMMANDS. I TAKE NO REST.

    “What words of purpose?”
    RELEVANT TEXT THAT ARE THE FOCUS OF BELIEF. GOLEM MUST WORK. GOLEM MUST HAVE A MASTER.
    “Sorry, look,” said Cheery. “Are you telling me this… thing is powered by words? I mean… is *it* telling me it’s powered by words?”
    “Why not? Words do have power. Everyone knows that,” said Angua. “There are more golems around than you might think. They’re out of fashion now, but they last. They can work underwater, or in total darkness, or knee-deep in poison. For years. They don’t need rest or feeding. They…”
    “But that’s slavery!” said Cheery.
    “Of course it isn’t. You might as well enslave a doorknob.”

    There was another protest march going on when Moist walked to the bank. You got more and more of them lately.
    This march was against the employment of golems, who uncomplainingly did the dirtiest jobs, worked around the clock, and were so honest they paid their taxes. But they weren’t human and they had glowing eyes, and people could get touchy about that sort of thing.

I recently finished the Aubrey-Maturin series after 13 months of through-reading thanks to a different HN thread. Quite a different series but certainly worth a read as well, especially books 3-10 or so.

That is, with ambiguity, I try to assume the best. I expect that is somewhat naïve.

I genuinely read (and still do) the blog as a human voice. I don't think writing about AI is enough to assume that a blog is authored by AI.

This doesn't follow. For instance, there are some pictures that I know are AI-generated, yet they're still beautiful to me. Nothing jaw-dropping, just very nice. Being AI-generated doesn't automatically mean being not worthy, especially when it comes to art. I understand, this is kind of insulting to human artists, writers, etc: we thought only the human soul and Nature could produce "the beautiful", but apparently not.

Which is not surprising, because LLMs are specifically trained to please their audience. Of course they can produce uhhhh "content" that people will find beautiful, that's not even necessarily a "bad" thing.

I can't really say much about the central theme without giving away a disc-shaking change to the world, because it's about that event and the next generation (Tiffany) carrying on. It's a very sad story, but also meant to be encouraging, and clearly intended to be his last book.

First time you read it should be in publishing order.

The Minas Tirith part of the ending of LOTR hits hard as a 50-year-old man, but as a kid, it was just a lot of pomp.

Hope you get a good reread sometime.

Hard to believe most of this was written 3 decades ago.

More importantly, both of those sentences make complete sense in context, and neither is phrased in a way that AI would. They are phrased in the way that Terry Pratchet would have. Have you never read him?

This new trend of pointing out that everything you dont understand is AI has become a flashing warning sign about our declining literacy rates.

Literacy is in serious trouble, and worse it has effected the way humans THINK. We are all poorer for it.

Read more books people!

    The book has to be small enough to disappear when a teacher looks up. Pocket editions, as their name suggests, were engineered for this. Pratchett’s were small, fat, slightly battered, and printed on a kind of paper that already looked guilty.

Not only does the paper "look guilty", but it's doing so "already"? As if guilty paper is normal, but not on THIS time scale.

It's nonsensical; even bad writers don't end up with stuff like this.

Assuming it was an intentional, it could be a reference to one particularly violent piece of furniture. (I forget what kind exactly, it's been a while.)

Two: Terry Pratchett "Hold my beer."

Clearly as an appreciator of hats, and arguably furniture, Sir Terry was echoing Monty Python.

It's a good point, and one I hadn't considered.

It seems to have resonated a lot with male millennials at least. Many of my friends growing up loved Terry Pratchett. I loved those guys but calling them "cowardly underachievers" is probably fairly accurate, if a little mean.

All of the ones I kept in touch with have settled for a lot less than they probably could have done if they had been a bit braver. Few of them were even willing to move even an hour drive away from our hometown for better opportunities

The bit you quote isn't even something he said! In the original post[1] it was assigned to "certain very old and very tired philosophers", but that's now been edited to just being credited as "a theory, not necessarily a really good theory".

[1] https://web.archive.org/web/20260523124615/https://www.mahl....

-- Yogi Berra

Writing is hard now, not because AI exists, but because there are so many writers out there and everyone's competing for attention, not just with other writers, nor with books from the past, but with all forms of media. Loads of people today, who might otherwise be reading novels for entertainment, are too busy scrolling their phones or watching TikToks or playing video games.

We don't have another Terry Pratchett because all the would-be Terry Pratchetts are toiling in obscurity, and possibly giving up on writing as a result.

You could totally argue that if people can't tell the difference, it's irrational for them to care which one they get, and I don't totally disagree with that either, but it's not like personal tastes have ever really been a rational thing either. Our ability to enjoy something is the result of a bunch of signals in our brains, and it's not that crazy that adding another signal (or removing one) can change that result in a way that makes it more or less desirable to seek out. Some people might literally like a piece of writing more if they have reason to believe it's from a human than they would if they read the same exact thing but had reason to believe it's from an AI, and while I would find a study showing that as fascinating, I wouldn't see that as an argument that people like the wrong things, because "right" or "wrong" don't really seem like they apply to that sort of thing. If someone told me that knowing there's a human on the other end and that having some sort of indirect, one-way emotional connection to them is an important part of what makes them enjoy it, who am I to tell them that's wrong?

It's only partly a joke.

LLMs harnesses try to make them useful to suggest things, but this is the most destructive thing you can do to a writer. You can work around it by just feeding Claude a writing critique skill.

It’s funny how anytime an article gets called out for being AI slop on HN, the author’s reaction is something like that: “oh yeah sorry I used AI but just for proofreading I swear, I should’ve done just a tiny bit less”.

No one seems to get the message that relying on AI at all is what makes writing shit. Good writers have confidence in what they produce. The fact that you’re willing to incorporate any AI suggestion at all means you’ve already lost the battle.

And Terry Pratchett would have loved it too. Even if it were clumsy.

But thanks for being kind in taking criticism.