Using AI to write posts
- Thread starter Cheimoon
- Start date
-
In the latest episode of the United Hour Podcast, Nik and Imran are here to talk about a frustrating 2 points dropped at Crystal Palace. They then discuss the regeneration of Old Trafford. Finally, Nik is joined by special guest Michael Statham from the Football Oranje podcast - who gives us a more in depth look at our EL opposition FC Twente, and then runs through the (many) Dutch connections currently at United.
We know. Your threads are there for all to see
AI on a forum is an abomination. What have we come to? What's the point of having individual posters? We might just as well have a load of bots interacting.
So at work we have this genius autistic guy (like, unironically autistic, not in an insulting way) who was quantitively the smartest guy in almost every room, but couldn't communicate for shit.
I literally mean that, he used to wind up every single person on every work comms platform because the way he typed messages was very abrasive. When he saw someone put something dumb or make a stupid suggestion, he would literally say, "Guys, this is insane, terrible idea, couldn't possibly work" etc.
He got that feedback from management, and because he was autistic we all gave him a lot of leeway because it isn't quite his fault his emotional intelligence was a bit lacking, but even then it could be very jarring. We used to keep him a mile away from any external clients or external comms facing meetings.
Anyway, he tried his very best to rectify this and decided to use chatGPT to solve the problem
So what he could do is he would type the message on Teams as he nominally would, and then copy and paste it into GPT with the prompt of "make this sound nice and polite."
So messages like, "Hey, this is urgent, please get this to me by end of day today. It's already been a week and I'm blocked without this data."
would turn into:
"Good morning! Hope you're having a wonderful day! If it would be possible, could you take a small moment of your time today to...[rest of paragraph]"
Anyway, for about a week everyone thought he was passive aggressively taking the piss and mocking us until one of us saw him copy and paste his messages to GPT.
Rule of thumb: feck GPT for human to human interaction.
You tell 'em Penna!
That's Chat Bot Penna, it's like an answering machine. The real Penna is somewhere in the world drowning kittens.
They squirm a lot, too.
I've started using long and winding chat gtp messages for people that I am required to interact with but have absolutely no interest in.
Oates told me that's why you enjoy doing it.
There is an argument for forcing some posters to use AI.
AI certainly does not provide lower quality interactions than for example posters whose post history is 95% tweets without a single personal comment.
At least the tweets are interesting.
I have absolutely no reason to believe that everyone else on this forum is some form of AI bot. Some are just very, very early versions.
They built me with old VCR parts.
Can AI write out responses to posts that have really poor grammar? This may be a way to outsmart the bots.
I assume so. You'd just run it through a spell check, which AI is surely capable of doing?
No I mean the AI post itself. Every time you read something written on chatgpt it is formatted perfectly and never has any type of grammar error.
If you asked it to write a response to this post and use incorrect spelling and grammar, can it do it?
Yeah it can do that. Just prompt it to make the grammar worse after you've got your initial answer.
Edit: using those prompts I just got this:
One fun fact about Roman Empyre is they had many roads, like over 250,000 mils long at peak, with about 50,000 miles paved. This road thing helped trade and moving soldiers, also made it easy for cultures and info to spread. Saying "All roads lead to Rome" shows how important roads was.
I also prompted "add 3 spellings errors" to get that result.
We are all doomed. Can't be outsmarted.
Those are shit, unlikely spelling mistakes though. Still doesn't read like a real person wrote it. "Empyre"? Come on.
I dunno, y and i and two spaces away from each other on a keyboard. It's not an impossible mistake if, like me, you have fat thumbs and a lack of coordynation.
They haven't written the content of the post in that case, though, have they? Whatever about using it to format and correct something you've written, asking it to literally construct your posts for you is basically just posting AI content.
Straw man argument invented out of thin air.
Last edited:
The rest of the content is just grammar. The point of communication is to convey your ideas across, and the purpose is achieved. The main problem right now is that the technology is not yet advanced enough to make the post appear natural.
It really isn't. Yet again you seem to be missing the point of a discussion. Let it go.
I'm not sure how in this ridiculous scenario someone is able to exclusively say key words that 100% portray their thoughts on a topic for an AI model to fill in the blanks. It's not happening.
You just lack imagination then. The examples shown so far are just really low level AI free for public use. Future AI models will be able to learn from humans through hours of coaching, and mimic their posting style. It doesn't even need to be 100% to appear totally natural.
It definitely can be done. But what's the fecking point?
This website is an unbelievable waste of time for all of us but some of the only upsides are practicing the ability to formulate and communicate your thoughts in a way that might be interesting to other people. That and the vague satisfaction when you occasionally craft a particularly well written post.
Out-sourcing all of that for a bot would tip the balance between redcafe being an almost complete waste of time to a 100% complete and utter waste of time. At which point we should all log off and leave the LLM drones to chat to each other.
I agree, I don't use it and haven't actually given it much thought until this post. My first reply was out of concern because it was starting to feel like the early stages of a crackdown in DPRK, and I wasn't sure if people were completely serious. I am a modern day Voltaire, after all. This thread did end up with lots of examples showing why its not great to use, and I think it achieved its purpose.
You don't seem to understand the point people are making, if someone contributes a small amount of actual info from their brain to a post, then they aren't really posting. It will end up like those IG comment sections with bots talking to each other.
Has your man with the terrible username and suspect opinions been banned?