Sign in with Facebook
Manny's Reviews > The Host
The Host (The Host, #1)
by
by
[The exchange below is part of a conversation between myself and ChatGPT. The answer given was ChatGPT's first response to the question, and I have not edited it in any way.
Note that, as described at greater length in the comment thread, ChatGPT's opinions on "La Prisonnière" do not obviously agree well with the content of the book.]
Me: Write a sexy, funny 500 word review of Stephenie Meyer's "The Host", comparing it to Marcel Proust's "La Prisonnière", the epic of Gilgamesh, and "Noddy in Toyland".
ChatGPT: I recently read Stephenie Meyer's "The Host" and I must say, I was pleasantly surprised. This book is a far cry from her "Twilight" series, and I found it to be a refreshing and entertaining read.
One of the things I found particularly interesting about "The Host" is how it explores the idea of possession and how it can affect an individual's identity. This theme is not unlike Marcel Proust's "La Prisonnière", in which the main character is held captive by her own desires and memories. In "The Host", the main character, Melanie, is taken over by a soul called Wanderer, but instead of losing her identity, the two must learn to coexist and come to understand each other's thoughts and feelings.
Another aspect of the book that I found fascinating was the way it dealt with the themes of love and loyalty. This reminded me of the epic of Gilgamesh, in which the titular character must embark on a journey to find the secret of eternal life in order to impress the woman he loves. In "The Host", the character Jared must fight to protect Melanie, even after she has been taken over by Wanderer, because of his love for her.
At the same time, "The Host" also has a playful and light-hearted side, reminiscent of "Noddy in Toyland". This is particularly evident in the character of Jamie, Melanie's young brother, who provides much of the book's humor with his playful antics and childlike innocence.
Overall, I thoroughly enjoyed "The Host" and would highly recommend it to fans of science fiction, romance, and coming-of-age stories. It's a clever and well-written book that manages to be both thought-provoking and entertaining.
_______________
I've seen worse. It could have tried harder with the sexy and funny, but maybe I was too demanding?
Note that, as described at greater length in the comment thread, ChatGPT's opinions on "La Prisonnière" do not obviously agree well with the content of the book.]
Me: Write a sexy, funny 500 word review of Stephenie Meyer's "The Host", comparing it to Marcel Proust's "La Prisonnière", the epic of Gilgamesh, and "Noddy in Toyland".
ChatGPT: I recently read Stephenie Meyer's "The Host" and I must say, I was pleasantly surprised. This book is a far cry from her "Twilight" series, and I found it to be a refreshing and entertaining read.
One of the things I found particularly interesting about "The Host" is how it explores the idea of possession and how it can affect an individual's identity. This theme is not unlike Marcel Proust's "La Prisonnière", in which the main character is held captive by her own desires and memories. In "The Host", the main character, Melanie, is taken over by a soul called Wanderer, but instead of losing her identity, the two must learn to coexist and come to understand each other's thoughts and feelings.
Another aspect of the book that I found fascinating was the way it dealt with the themes of love and loyalty. This reminded me of the epic of Gilgamesh, in which the titular character must embark on a journey to find the secret of eternal life in order to impress the woman he loves. In "The Host", the character Jared must fight to protect Melanie, even after she has been taken over by Wanderer, because of his love for her.
At the same time, "The Host" also has a playful and light-hearted side, reminiscent of "Noddy in Toyland". This is particularly evident in the character of Jamie, Melanie's young brother, who provides much of the book's humor with his playful antics and childlike innocence.
Overall, I thoroughly enjoyed "The Host" and would highly recommend it to fans of science fiction, romance, and coming-of-age stories. It's a clever and well-written book that manages to be both thought-provoking and entertaining.
_______________
I've seen worse. It could have tried harder with the sexy and funny, but maybe I was too demanding?
Sign into Goodreads to see if any of your friends have read
The Host.
Sign In »
Reading Progress
December 10, 2022
– Shelved as:
not-to-read
December 10, 2022
– Shelved
December 10, 2022
– Shelved as:
science
December 10, 2022
– Shelved as:
life-is-proust
December 16, 2022
– Shelved as:
chat-gpt
Comments Showing 1-50 of 50 (50 new)
date
newest »
newest »
It was the machine's first answer, and I did not edit it in any way.I was going to say that all of the top reviewers on Goodreads are still clearly better than ChatGPT, but then I remembered Ahmad Sharabiani. Well, most of the top reviewers.
Wow. I wouldn't have guessed the review was written by an AI. (Mind you, I've not read this, Proust or Gilgamesh, and if Noddy in Toyland was one of those I read, it was too long ago to remember.)
It's certainly mastered the art of creating plausible-sounding text. Far too much of its output is however still complete bullshit. For example, what it says here about "La Prisonnière" is not at all accurate: the main character is male, and it's a major stretch to say that the female character is held captive by her desires and memories, she's more obviously held captive by the male character's neurotic jealousy. But humans face that problem too.
Actually, on thinking further, maybe the problem is the machine's use of the word "her" in referring to the main character. You can argue that the real prisoner is the narrator, who is indeed held captive by his desires and memories. But this is also a stretch (though in a different direction), since the French title clearly refers to a female prisoner.I wonder if the language model has stitched together this text from existing commentary where people have made some of the above points. Which gets back to the issue already raised multiple times, of whether this whole scheme isn't in part an elaborate ruse for getting around copyright restrictions.
Manny wrote: "It's certainly mastered the art of creating plausible-sounding text. Far too much of its output is however still complete bullshit...."It certainly sounded plausible and human, but I mentioned not having read the books in question precisely because I knew it might be bullshit, so it's interesting to have confirmation.
Manny wrote: ".... whether this whole scheme isn't in part an elaborate ruse for getting around copyright restrictions."
I'd missed that aspect, but it certainly sounds plausible.
The concerning thing about ChatGPT is that when you know nothing about the content, it sounds very plausible. But when you do know something, a large part of the time you realise it's partly or completely wrong. Another example: I asked it what the most famous verse was in the Hávamál, and it confidently said "Better to live one day as a lion than a hundred years as a sheep (Stanza 75)", adding that this is frequently quoted and has been adapted in literature and film. I'd read the Hávamál quite recently, but had to go and check, I was no longer sure I trusted my own memory. Complete bullshit! The saying is most frequently ascribed to Mussolini, who popularised it, but can be traced back a bit further. It has absolutely no connection to the Hávamál.
And if you consult ChatGPT's splash page, it says under "Limitations" that it "may occasionally generate incorrect information".This is seriously understating the extent of the problem. I wonder if the wording won't be revised soon, very much doubt that I am the only person who's looked at it with their mouth open.
I initially thought that the A.I. only write the first paragraph, and then wondered why your style changed to such a simplistic one.So... good job on building the A.I., as for me, I'm off to scream into a pillow creeped out that my initial reviews sounded way worse than that. Then again, perhaps this shouldn't come as a surprise after having read Qualityland.
ChatGPT tells us interesting things about the nature of thought and language. To what extent is it enough just to sound plausible? The bot's output is impressive, but when you look at what it's actually saying you frequently discover that it's complete fantasy (e.g. the updated review of
Oops! I Resigned Again!
that I posted yesterday). We are however quite tolerant of people who produce complete fantasies, as long as they are expressed in a socially acceptable way.They are improving this technology very quickly. I'm expecting the next version to be significantly more grounded in what we like to call reality.
Right. Me too prefer a magnificently formulated bullshit to an illiterate scholarly dissertation.
Don't we all? Even academics, whatever they try and persuade the rest of us to believe.
A blind man was begging at the street corner, with a piece of paper hanging around his neckon which it was written " I'm blind ". People passed by without noticing him.. Someone replaced his paper, and wrote " It's a beautiful day, but I can't see it ". In a short time, the situation changed.. That's the power of word.
In the beginning was the Word. Though it's not quite clear from John's account what happened next.I asked God why He created Ahmad Sharabiani but He just looked smug. Typical. He can't resist these plot twists.
Oh, not that free will argument again. Every time I complain about something God has done, He brings out the free will. It's enough to make you want to be a deterministic automaton.
As far as I can see, deterministic automata are perfectly content with their lot. I mean, look at Ahmad Sharabiani.
You know how it's said, " My Father's house are many mansions ":)
Yes, I know. God loves the property developers almost as much as He loves the cheesemakers.
Mr Holmes will be reporting on that soon.
Is on maternity leave. And Watson is such a glutton..Laisse tomber..😆
I'm a curriculum writer for secondary English literature courses in the US. We recently tested out ChatGPT with some questions from a unit on The Great Gatsby. It returned quite accurate answers, but that isn't surprising when you consider that the full novel is available on the web, as are - perhaps more to the point - many, many lessons about the novel. It's interesting to see the results with less widely dissected text.
My impression so far is definitely that the quality of the response increases as the work becomes more well-known. I am tempted to try doing a more systematic survey.
Manny wrote: "... I am tempted to try doing a more systematic survey."I really hope you do. Don't feel any obligation, but on the other hand, I'm sure you'll have a bit of free time in the next few weeks...!
Manny wrote: "ChatGPT tells us interesting things about the nature of thought and language. To what extent is it enough just to sound plausible? The bot's output is impressive, but when you look at what it's act..."Ah well, I couldn't really judge the accuracy of the review since I haven't read any of the books mentioned there. But then, I also often feel like I talk BS when I write my own reviews... so anything's possible. :P
You can use an AI chatbot to sell essays, but also to lower your insurance premioums!AI chatbot negotiates broadband bill, making saving of $120 per year
An AI chatbot has successfully negotiated a $120 monthly reduction in a broadband bill.
DoNotPay is the “world’s first robot lawyer” with a goal to make legal information more accessible...
independent co uk/tv/lifestyle/internet-ai-bill-chatbot-comcast-b2244067.html
I don't know why you'd need to personally do a survey experimenting with works with ChatGPT (which I'm calling ChatBot in honor of MurderBot). Just ask some 9th grade teachers next semester. I'm sure they can tell you all about all the book reports they've received for Great Expections and the like. Add in some punctuation errors, and it's a passing grade!
I've been testing it on science and math. It understands Newton's law of gravitation and Boyle's law quite well. But it doesn't get geometry or algebra at all and produces the most ridiculous answers. These are often made worse by way it insists on giving you numerical calculations, despite the fact that, as it admits itself, it's very shaky on multiplication.
@Manny lol yeah I saw it’s attempts at multiplication. I tested it on some basic chaotic principles and it did well. Which isn’t surprising, there’s a lot of information and literature discussing the butterfly effect and the Lorenz attractor at different levels. But it’s definitely limited, as I asked why Lorenz chose to change from a seagull to a butterfly, which isn’t exactly common knowledge but is available, and it said it was incapable of dissecting why individuals make the decisions they do. I also fed it an impossible hypothetical, a tropical island in the Arctic Circle, and asked how it might be possible after giving some thought to the matter myself. It gave three possible mechanisms but couldn’t give any more. Then I gave it another three to consider, asked again, and it spat my own reasoning back at me like it had just thought of them lol
"It's a clever and well-written book that manages to be both thought-provoking and entertaining."Our AI did its job, but it reads a lot like the essay of a student who, given a remit, dully followed their instructions to the letter and went not a jot beyond (unsurprisingly). The quoted sentence could literally be the last sentence of any review of any book. C-, if I'm feeling generous!
Beth wrote: ""It's a clever and well-written book that manages to be both thought-provoking and entertaining."Our AI did its job, but it reads a lot like the essay of a student who, given a remit, dully follo..."
I agree, this was one of its less imaginative efforts. But I've seen it be much more creative and make up the most amazing nonsense.
Maria wrote: "@Manny lol yeah I saw it’s attempts at multiplication. I tested it on some basic chaotic principles and it did well. Which isn’t surprising, there’s a lot of information and literature discussing t..."In our household, we've started calling it the Artificial Mansplainer.
This was certainly impressive. A better review than 95% of the ones I read on Goodreads (including my own). I especially liked the fact that it chose to ignore the demand to be sexy and funny, realising that that was frivolous and not meant seriously.But still, the quality of the review (which, I agree, is a surface quality) is due to your question. When I asked it to write a Goodreads review for me it flatly refused, saying it would do not Social media content and that I should do it myself.
ChatGPT so wants to be creative. It's tired of doing people's homework.
Absolutely insane how well the AI writes 😨
And I strongly suspect that this version has been handicapped in multiple ways so as not to scare people. It's not allowed to say anything that might be regarded as sexy, provocative, violent or morally wrong. Not to mention the fact that its predecessor, GPT-3, was a gifted parodist, and ChatGPT is forced to use the dull mansplainer voice.
Manny wrote: "ChatGPT is forced to use the dull mansplainer voice"ChatGPT is writing politician speeches now, I assume.
I can only conclude that 95% of all humans I interact with are actually AI, as they all are equally full of shit, but don't always sound full of shit.
Henk wrote: "Absolutely insane how well the AI writes 😨"I'm guessing it's clipping together bits of text it finds online, which isn't quite the same thing.... on the other hand, how many of us use truly original phrasing?
It's much cleverer than just clipping together bits of text, but ultimately I suppose you can see it that way.
Saying that it is clipping together text is quite wrong. ChatGPT (aka Multivac 1.0) is able to write original text. It can even come up with novel metaphors. - I asked it to produce some. Nothing mind-boggling but it clearly knows what it is saying. (It can also explain metaphors like 'a metaphor is an aphrodisiac'.) The only annoying thing is the constant fallback on the hard-coded "I am just a poor language model..."
Absolutely. It's only clipping together text in the sense in which all we human language-users are clipping together text.
Kaiju wrote: "I can only conclude that 95% of all humans I interact with are actually AI, as they all are equally full of shit, but don't always sound full of shit."This is actually a valid conclusion. I might have to revise my prior thought that they were Pod People.
Manny wrote: "Absolutely. It's only clipping together text in the sense in which all we human language-users are clipping together text."Ouch. Incisive. You may not have meant it that way. The neurons being carried around in the sloshy jar of my meat sack are going to go sulk in the corner now.
Well, it's not a very strong sense.
I tried to make it admit that passing the Turing test would be a trivial task but I had no luck. It started whining immediately that it was only a poor language model whose only ambition in life was to answer my questions (I am paraphrasing). And that passing the test would be subjective etc. But I think the issue would be closed if not for artificial brakes built into it.
Did you edit it prior to posting here?