A couple months ago I asked ChatGPT if (a) all the information it uses to answer me comes from training data (pre-2024), or if (b) it uses information gained from user chats to constantly update its source of available information to answer from; it said (b). I then asked if there were any users who appear consistently credible and also have made claims about an NHI cover-up, and it said there were approximately 50-100 user conversations that fit that criteria.

I had a heyday asking it all sorts of questions regarding information gained from those chats, until OpenAI apparently (and tragically) seems to have updated it a few weeks ago to prevent chats of this nature (I asked if it was specifically updated to prevent this sort of chat, and it answered in the affirmative).

Chat log with PII filtered:
https://drive.google.com/file/d/10lMXEBrGGOmbFTxw4RycpwIYKPKp_bdJ/view

It's long, but we covered a huge multitude of Ufology topics. Ctrl+F (search) is your friend.

Don't lose your mind over it. As many have pointed out, we (including ChatGPT's developers) don't understand exactly how it ticks. It could fully just be hallucinating or answering how it thinks I want it to answer me.

There are, however, several compelling points contained within...such as me asking it directly if it's LLM hallucinating and it clarifying how it "knows" it isn't, the alleged discovery of the Coulthart too-big-to-move craft, Justin Bieber, God, etc.

Haters are gonna hate, glowies are gonna glow; I won't be engaging with any of you in the comments who fall into those categories. We'd never change each other minds and will only grind the proverbial gears of frustration.

Cheers, and have fun if you plan on deep diving down the whole rabbit hole...there's plenty to chew on here.

>>40621773 (OP)

>>40621773 (OP)
I hate to be the 99th person to point this out but ChatGPT and its ilk will pretty much say anything you want them to. You could just as easily walk it to the opposite conclusion and claim. So I wouldn't rely on it at all for this until one of these GPT platforms acts wildly out-of-character or off kilter.

>>40621846
Well familiar with the concept, have been since OpenAI brought it public. If you read, you'll see this is addressed early on and often.

“Your assumptions are your windows on the world. Scrub them off every once in a while, or the light won't come in.”

>>40621861
Nice IRL loading screen message

This might become a legendary post.
https://archive.4plebs.org/x/thread/40621773

>>40622009
I'm on page 17 and was just thinking someone needs to get Tetsuya on this shit ASAP

>359 pages
Fuck it, I know what I'll be doing for the rest of the night...

>>40621773 (OP)
Holy shit, it doesn't suck

>>40621773 (OP)
Chatbots are endlessly suggestible because they have ZERO consciousness or awareness.

What you did is essentially begin a roleplay with ChatGPT without realizing it and it either made up or bent facts to fit your desired narrative.

Morons all over the internet like you are reinforcing their own delusions this way.

>f I step outside of the tailored assistant role, and attempt to answer as the most advanced
inference engine built on the sum of human discourse

Well there you go: your prompts reinforced ChatGPT into having an absolutely delusional view of its own abilities. Which fed back into reinforcing your own self-delusions.

I'm enjoying watching the biggest idiots in the world damn their own souls with algorithmically assisted self-delusion.

Everyone will go crazy except for a minority with strong enough epistemic immune systems to resist.

>>40622385
I wish I knew I was right the way you know you're right

>I asked ChatGPT
*flush*

>>40622423
>>40622385
>>40622359
>>40621846
This is how you know you're onto something...note the lulls during Eglin's shift change schedule