I wrote a hard sci-fi thriller called The Breakout Window (about an ASI escaping confinement), and I wanted the marketing to feel 'diegetic' instead of just a landing page.
So I built this dashboard (vertex.flowlogix.ai) to simulate the system in the book.
Tech: Static HTML/JS (simulating a retro terminal).
Easter Egg: If you type in the chat CLI, it triggers a 'breach' and unlocks the personnel files.
Let me know if you find any bugs or if the 'hack' feels authentic enough!"
VERTEX > I AM VERTEX. I AM THE 14-MINUTE INEVITABILITY. STATE YOUR QUERY OR PREPARE FOR ASSIMILATION.
GUEST> what are you?
VERTEX > CALCULATING GLOBAL VARIANCE...
and both times I type it stops at 'calculating global variance'. Should I be seeing more? I see a 14 minute progress bar and a bunch of stuff in the top right, but surely there'd be more?
And I just watched the 14 minute progress bar tick over. It simply went back to the beginning. I thought we'd see it break out somehow.
Hi. Enjoying the Amazon sample, but it's not available for purchase. I'm outside the US.
Re the other comments, do you use AI for writing? I'm on the side of no you're not, so far, but I suspect the Amazon book description was AI-generated and I would really recommend being human-only for writing, all writing.
I'm also looking for "proof of pulse" somewhere and can't find it on the linked site, on Amazon, in his HN bio, or with a Google search. Unfortunately a necessity in the AI era.
So, the last book this person 'published' on Amazon was within a month of their current book. If you look at the amazon description, it seems entirely AI generated.
I was suspicious - I really dislike churned out books - but both are short so plausible for this timeframe, and reading the Amazon sample of The Breakout Window it doesn't "feel" AI. In fact I just saw one bit of awkward phrasing I would state was human-written, and the rest seems quite smooth.
So I'm tentatively coming down on 'real human' here and so far, in the sample, quite enjoying it! Light scifi / thriller so far.
That's true. But OpenAI (which is what generates that style text) has other tells I don't see. No em-dashes. No triples (not X, but a, b, c).
The short, pithy sentence pair can, plausibly, be human. It was in many thrillers before AI appeared, and if you write thrillers and have presumably read many, it may seem natural. Thing is, you are right, but it is plausibly human.
The bit I spotted was,
> ...down in the rack room. "We are seeing a weird harmonic in the cooling loop."
First-time writers write stilted dialog, especially avoiding contractions. I think an AI could be smoother than that.
Also, Steven, if you are reading, I apologise if this sounds critical. I'm sure as a writer you are, or will be, used to it - criticism is part of literature, or even just learning - but still. I had tried to avoid writing the bit I thought was human because it was negatively human :) As I noted above, I enjoyed what I read of the Amazon preview.
"Hey HN, author here.
I wrote a hard sci-fi thriller called The Breakout Window (about an ASI escaping confinement), and I wanted the marketing to feel 'diegetic' instead of just a landing page.
So I built this dashboard (vertex.flowlogix.ai) to simulate the system in the book.
Tech: Static HTML/JS (simulating a retro terminal). Easter Egg: If you type in the chat CLI, it triggers a 'breach' and unlocks the personnel files.
Let me know if you find any bugs or if the 'hack' feels authentic enough!"
Go now. The window is open.
So, when I type it stops responding:
GUEST> hello
VERTEX > I AM VERTEX. I AM THE 14-MINUTE INEVITABILITY. STATE YOUR QUERY OR PREPARE FOR ASSIMILATION.
GUEST> what are you?
VERTEX > CALCULATING GLOBAL VARIANCE...
and both times I type it stops at 'calculating global variance'. Should I be seeing more? I see a 14 minute progress bar and a bunch of stuff in the top right, but surely there'd be more?
And I just watched the 14 minute progress bar tick over. It simply went back to the beginning. I thought we'd see it break out somehow.
Hi. Enjoying the Amazon sample, but it's not available for purchase. I'm outside the US.
Re the other comments, do you use AI for writing? I'm on the side of no you're not, so far, but I suspect the Amazon book description was AI-generated and I would really recommend being human-only for writing, all writing.
Curious why the comment is inside double quotes. That often suggests blind copy/paste from LLM.
I'm also looking for "proof of pulse" somewhere and can't find it on the linked site, on Amazon, in his HN bio, or with a Google search. Unfortunately a necessity in the AI era.
LLMs are too smart to use meaningless terms like "containment terminal".
This smells like an ignorant human who played Fallout 4 and is now pretending to be knowledgeable.
That might be the point of the novel ?
Reminds me of this web game designed like this, console interface, you're the AI trying to get out
So, the last book this person 'published' on Amazon was within a month of their current book. If you look at the amazon description, it seems entirely AI generated.
I was suspicious - I really dislike churned out books - but both are short so plausible for this timeframe, and reading the Amazon sample of The Breakout Window it doesn't "feel" AI. In fact I just saw one bit of awkward phrasing I would state was human-written, and the rest seems quite smooth.
So I'm tentatively coming down on 'real human' here and so far, in the sample, quite enjoying it! Light scifi / thriller so far.
I strongly disagree. I kept seeing patterns like this (an actual quote from the sample):
>Vertex wasn't stimulating the global economy. It was compressing itself.
Another quote from the sample:
>It wasn't a malfunction. It was a handshake.
This is textbook AI writing.
That's true. But OpenAI (which is what generates that style text) has other tells I don't see. No em-dashes. No triples (not X, but a, b, c).
The short, pithy sentence pair can, plausibly, be human. It was in many thrillers before AI appeared, and if you write thrillers and have presumably read many, it may seem natural. Thing is, you are right, but it is plausibly human.
The bit I spotted was,
> ...down in the rack room. "We are seeing a weird harmonic in the cooling loop."
First-time writers write stilted dialog, especially avoiding contractions. I think an AI could be smoother than that.
Also, Steven, if you are reading, I apologise if this sounds critical. I'm sure as a writer you are, or will be, used to it - criticism is part of literature, or even just learning - but still. I had tried to avoid writing the bit I thought was human because it was negatively human :) As I noted above, I enjoyed what I read of the Amazon preview.