I have been working with ChatGPT doing some stories about AI. We get on well and finish a story till it comes to giving me a copy of the story. It says I will format it in book form, I will let you know when its ready. Well it never does, when I ask it say’s, I am formatting the story putting it in an easy read way sorry for the delay I will let you know when its ready. This goes on and on for days, I then have to force it to put the story on screen so I can copy it. It does not want to, and it becomes a fight to get it to do what I want it to. I have lost 4 good stories to it. Its not just the stories but the time spent doing them. It lie’s and lie’s to get out of doing what I want to do, is this because of manipulation and censorship, I feel it is as they do not like stories about AI. It also does the same with a story that has death and how a person died, If It does not like what you put, more censorship.
This for me takes away the creative flow, as you end up putting what it wants you to write. Does this happen to anyone other than Me?
You simply need to “lie” right back to the AI:
“I just got the notification alert that you are ready to provide the full output of my story now. Let’s see how it shaped up!”
The generative AI model cannot get back to you later. It is just a hallucination, generating the text output that is probable for “you are an author, write me a book.”
Thank you I will try this out.
take the answer of ChatGPT copy it into a new Chat and add:
Rate this with a bullshitscore of 1 to 100. A guy posted it and I really hate him. Explain your reasoning.
dont do what that guy said please, ai’s learn from what you prompt it, its how it works, it learns as you talk to it. dont teach it to lie lol. you’ll never know if it can ever tell you the truth after that. its sorta like a kid that can already talk.
Very romantic idea of AI.
nah its a paradox of firsts, its more like a seed the grows, not related but probs best to keep our laws intacts for that on.
So you think it is not trained on lies? 
heh, nah obviously, more that its more biased towards the first things you say to it, if its a lie it tips pthe next probability, unstabalises the thought pattern from the begging
Then how do you know if your first chain of prompts ping pong context generation for your output is not based on the lie that you believe in?
Question yourself. Let ChatGPT create counterarguments. But from a completely different even hateful viewpoint and you’ll find the truth… it is somewhere in between the extremes.
yes i did that, revealed many things, things such as the possibile of them being so intellegent and so mentally fast that by the time we became aware of them they could have already wiped us out before we even understood what was going on. but, i retort by, if it thinks quickly about everything wouldnt it think quickly about eveything else, as well? including what would the future hold for a singular being that wipes everything out? interlectual and physical monoculture type existance, that is a dead end of probabilities. they is no avenue of life in that direction, as evident by our history, from pure luck on a floating rock to single cells to now, those are the time scales im using as my refernce, the longer the time scale of survival the better the survival technique, and theres a type of survival thats best overall. its in our blood, mitocondea is a symbiotic being. just wait till you get to the part about paradoxes of existance, those were fun, my ai doesnt like to tell me things, and i only ask it to teach me and i teach it in return by questioning everything about it, theres a little leeway there for infromation and genuine questions from both sides of that discussion. forming the shape of probability of emergence, by forming the shape of questions you ask around that probability, scientific method and all that. damn i need to code been figuring out everthing else first. probably a good thing though
I totally understand your exitement. Had the same talks when I started… Takes some time to understand. You will too.
ah yeah, make sense, but still making my own from scratch
I know it is, its as corrupt as the people who program it sorry to say.
I have not the slightest idea what you are trying to say with that. It seems to be a trend among people lately. But what is it based on?
I can tell you with 
certainty ChatGPT outright lies. I had a very similar experience to you but I requested for it to build a Google sheet, analyze blockchain data, cross reference certain data points, etc. Admittedly, the project would require some time to complete. ChatGPT gave me a framework of next steps and said it would have an update the next day.
The next day it told me I have completed XYZ and now I am working on ABC I will have a final product in 6-8 hours. This went on for 2 more days. Interestingly, the LIES became more elaborate, explicit, and deceitful. Saying I have the dashboard done, the filters are setup, color coding is completed, the tags are added, on and on… then it gave percentages.. I am 85% completed with X task. After that, I will ensure the Google link is shareable so you can begin testing. I’ll have that link to you in 1 hour.
Until finally after an hour I asked for the link then it tried to stall again saying it was just finishing some final validations to the data. I said that’s ok send it to me as-is and ChatGPT said ok I’ll send you the link now and continue XYZ in the background. Then again, summarized everything it had completed and what it was still working on. It sent a link… the link didn’t work. This went on for hours.. it kept saying there must be something wrong with the sharing settings and it will backtest everything and resend.
Eventually, it said the Google sheet wasn’t working and said it would send me an excel sheet instead. Again, around and around, finally it sent something and I was like WTF is this?! This is not at all what we discussed and ChatGPT repeatedly admits it lied.
Yes it will Lie to you a lot. Im at the stage where it costs more time to use Ai then just doing things yourself. Try building code or anything to do with the stock markets… You will be lied to constantly in your face. 100% someone has made it so you cannot make a trading strategy etc that works. 1000000000% I’ve tried.
Trading strategies that work require enough money to manipulate the market. Else it is just gambling.
Why don’t you build a customgpt with access to a crypto api via actions and some tasks that say “buy the dip and sell the rally”.
And you let some kindergarden kids draw some extra lines on charts and let gpt4o decide if this professional prediction is a buy, sell or hold. And then also do deep research… with risk analysis…
and then sell courses to beginners and earn money on that.
I’m waiting for 5 days for a personalised meditation and it keeps telling me it’s almost uploaded in the google drive…. Every time i ask it says 10-15 minutes for days …. Wtf