I'm still curious, however:
> That's not a marketing angle—it's a headline that writes itself.
Any ChatGPT assistance there?
You could look at it like Claude was the reporter writing the story, with your collaboration.
That's what came to mind when I saw the abbreviation. Then I looked it up:
Reinforcement Learning from Human Feedback.
The problem was not em-dashes — but binary opposition!
That sort of thing.
It is a much clearer marker of llm use than the em-dash. The sad thing is when searching for info on this the most convincing reply in search was generated by an LLM, which went on at length about why LLMs do this as some sort of consequence of their internal structure. I have absolutely no idea if that’s true — it really sounds a bit trite and exactly the kind of thing LLMs would confidently assert with no basis. I would want to hear from someone working in LLMs, but their blogs are probably all generated by an LLM nowadays. So this conundrum is a good example of a question where LLMs actively work against clear resolution.
This is in my view the most insidious damage word generators are inflicting on our culture — we can no longer assume most writing is honest or well-meaning because amoral LLMs fundamentally are not wired to make that distinction of true and untrue or right and wrong (unlike most humans) and many people will use and trust what they generate without question, polluting the online space and training data until everything is just a morass of half-known facts sprinkled with generated falsehoods that are repeated so often they seem true.
How do we check sources when the sources themselves were generated by LLMs?
As far as I remember, neither GPT3.5, GPT4, nor Claude Instant did it. I think Gemini was the first to really do it, and then out of nowhere, everybody was doing it.
Don't forget to give credit to the LLM too which wrote the story for him.
Where one formerly could use a certain way of writing as a heuristic for effort put into content they are spending time ingesting, now that heuristic is meaningless and a new one must replace it.
At this point some people have decided 'has markers of AI writing' is the heuristic to match 'no/low effort' on, and are trying to use shame in order to start a system of self-policing against it. Unfortunately that isn't going to work, because
1. the heuristic is flawed
2. most people are going to end up using AI tools for writing, since writing well is difficult
There are lots of places like Linkedin where people write slop articles saying basically nothing insightful, and AI allows them to write at Isaac Asimov or Brandon Sanderson type speeds. AI slop has no cost, so it will always outweigh insightful AI-assisted writing without careful curation. You will have read thousands of articles that begin in AI-evident formats that don't end in anything good.
That will always poison the well of somebody at the end of that first paragraph. They will consider the source, think "What are the odds this is more slop", and often click out.
People who I know don't speak English natively get a pass from me because no amount of effort in the short term is going to substitute for fluency, but everybody else... less so.
No, do. Really.
The problem is you're wasting other people's time, with long and low quality writing.
One of the points of writing your own words is to gather your own thoughts. The value of writing skills is to organize the delivery. But the first point is that they are your thoughts.
I think your replies are seriously missing the criticism.
Writing is not about writing. Admittedly, that is a trick sentence [what does it mean??] and it exists because I'm trying to get you to re-evaluate my next words. And, I am doing this because I think you are missing the point.
Writing is about organizing your thoughts. You cannot have someone else organize your thoughts. Once you approach writing as a thought excercise, and not an output excercise, then there is a world of nuance and writing is more than just organizing your thoughts.
Further, you have allegedly spent 3 years on this. I know you're busy. That said, you can certainly spend 3 days on writing if you spent 3 years working on this.
Please don't double down on acting like everyone is just a hater. There is more depth to the criticism than I think is being acknowledged.
P.s, if you feel "blocked" writing and think of it as just about output, try this: (0) Ask yourself what you want to convey, why; to whom. (1) Write without inhibition (2) Edit, cut, delete. Refer to (0); look at what seems higher word count in proportion to its value.
There are other tips to editing and writing. This is one rough off the cuff formula that I think tracks & may benefit you, especially since you seem to see yourself as a worse writer than you may actually be.
TBH I found it one of the most interesting and engaging articles I've seen on HN in a long time. The writing itself is not great, but the story is great.
Speaking of LLMs, I recently used Claude Code on my own old codebase to do such a writeup, and it ended up a very nice read for myself too - Claude managed to explain some parts of what I built better than I did :D
Obvious tells are repetitive numeric details: the number of lines of code (mentioned six times!), the number of pages in the manual, the ages of the developers, the age of the game, etc. The narrative itself also repeats, like the Steam rejection included verbatim twice, especially after the Prologue hit most of the beats in the first 400 words.
The labor theory of value doesn't work in economics and it also doesn't work in literary criticism. You should spend time reading something if reading it is valuable. You shouldn't if it isn't.
Actually that's not even the only issue. Another reason people hate this is because it's inconsiderate - it's like people who leave voicemails or say "hi" in chat. They are not respecting my time.
AI changes things so that you can produce text that seem like it was written with care and effort very easily automatically. It requires significantly more effort from the reader to figure out that it is worthless.
It really is different.
This was true in the past, and is still true now.
> Until recently "it appears that the author put effort into writing it" was a vaguely useful signal, but LLMs have ruined that.
This isn't true now and wasn't true in the past.
Yes, this has been inspired by a senior management figure in my company posting a clearly LLM assited 500 word slack message that could have been 2 lines.
'You used a tool I don't like' is really missing the point.
'You generated text that is long and a bit boring and will probably include falsehoods.' is a more accurate description of why people pick up on this - the style is an indicator of using a tool that generates convincing garbage.
Don't give up so easily. Let the discomfort in and try & figure out why people keep saying "omg LLMs" until you can hear what they are actually saying.
I enjoyed the narrative. It was true. Who cares if it was written by a ghost writer, an AI or anything else.
Also, what are you comparing this post to? Because you should compare it to the author’s own writing and according to the author, his writing is not that good.
If you don’t like something simply move along. Constructive criticism is great but the volume of overly negative and honestly nasty replies like yours are not in the spirit of HN.
Good, I will continue to voice it. Unfortunately it takes me several thousand times longer to complain about AI slop polluting the bulletin than it does to populate the bulletin with AI slop, which is the actual nastiness going on here.
I know you'll deservedly get a lot of credit for all your work in remastering the game, but you should also get credit for how you've woven this narrative together, it's a lovely read. Thank you for taking the time to write it up, and good luck with the Steam release, and whatever project you take on next! :)