February 23, 2026, 11:07 AM
sigmonkeyquote:Originally posted by mark123:...not use Reddit because that’s kinda worse than GIGO. It’s more like Anything In, Garbage Out”.
Yes, that's (f)actually, correct.
Yes, i steered the conversation in that direction. That was just for fun. However, I often use ChatGPT to check myself on history. History is a weak point for me. I’ve found that, while I had the gist of certain historical events, I’ve misunderstood nuances rather extremely.quote:Originally posted by Pipe Smoker:
^^^^^
Re: “Let’s say God lives in 10 spatial dimensions. What would it look like if He pushed 10 dimensions matter into our 3 dimensional universe?”
According to prevalent superstring theory our universe has 11 dimensions. one is temporal and the other ten are spatial. Of the spatial dimensions three are the familiar L, W, and H. The remaining spatial dimensions exist, but are of negligible extent.
Yes, this is what I’ve found as well. I wonder if I can ask it to not use Reddit because that’s kinda worse than GIGO. It’s more like Anything In, Garbage Out”.quote:Originally posted by sigmonkey:
ChatGPT and the similar "AI Modeling" does the same thing.
(10 dimensional thinking)
It takes what topic you start, and with continued "conversation", (meaning, you keep feeding back into it), the "machine shapes" the conversation taking "cues" from you and provides that which will keep you engaged, as well as stroking your ego. The ultimate feedback poop of mental masturbation.
Garbage In Garbage Out, it's just a better version of processing and regurgitating.
Everything in the "AI 'verse" is put there by us, and instructions are ours, no matter how much it becomes refined at doing it from continual "correcting", and no matter how much energy and computing power we provide, it will just be a newer iteration of the same thing.
Everything the ChatGPT" postulated, has been discussed in conversations for many years.
"While it may be entertaining, it's not life, Jim."
-Mr. Spockmonkey
quote:Originally posted by mark123:...not use Reddit because that’s kinda worse than GIGO. It’s more like Anything In, Garbage Out”.
quote:Originally posted by mark123:
I often use ChatGPT to check myself on history. History is a weak point for me. I’ve found that, while I had the gist of certain historical events, I’ve misunderstood nuances rather extremely.
quote:Originally posted by RogueJSK:
Not surprising, when you consider how these Large Language Model so-called "AIs" work.
They do not think. Period. They don't reason, or analyze things. It is not artificial intelligence, despite the label.
Instead, they are programmed to merely predict the most likely series of words to print in successive order, based on probabilities from the large amounts of text input provided for them.
A LLM is fed large chunks of text scraped from the internet, including the largest and most popular publicly available websites like Wikipedia, Reddit, Twitter, and the Youtube comments section, and then using those blocks of trillions of lines of text it uses software to predict the most statistically likely words to print in a specific order when given a certain prompt.
At its base level, it's just text prediction. Think of it like a fancier version of Autocomplete, like when your word processing software or text messaging app tries to guess what word you want next while you're typing, based on statistics from all the other text messages you and others have sent previously.
Now, who are the majority of people going online to the most popular websites to loudly and repeatedly express their political opinions on places like Reddit and the Youtube comments section?
It's not the quiet conservatives, moderate liberals, and centrists, who are too busy going to work, loving their neighbors, and taking care of their families.
Instead, it's the chronically online minority of toxic leftists, spending the majority of their time online ranting against ICE, screaming about how Trump is literally Hitler, cheering Charlie Kirk's murder, etc. and injecting their political beliefs even into nonpolitical discussions.
So even if those types of people are in the minority in the country as a whole, they have an outsized influence on the online political-related text available to a LLM, purely because of the volume of the text that they are putting out online.
When that makes up even a small majority of the text that is fed into the LLM, then that is what the LLM software is going to deduce is the most statistically likely response to print out when provided with a politically-related prompt.
And that's even before you get into the possibility of any deliberate manipulation by certain LLM corporations in an attempt to sway political opinions.