212 Comments

Footnote 1 = footnotes avail upon request? Like it!

Expand full comment

AI gives a strange name to Almighty God, too, and it is not same God of my Heavenly Father or His Human Son, the Lord Jesus.

I think some AI answers are giving the name of a "Moon" god or a Moon goddess.

Expand full comment

The key word in A.I. is artificial. Not long ago I received an unsolicited pilot A.I. program called Copilot. I asked it some pandemic related questions and found it responses were false. When trapped it apologetically tried to divert my search to a different subject. So A.I. is garbage in, garbage out.

Expand full comment

What would you do with an elephant? I'd wrap him up in cellophant. What would you do with a hippopotamus? I'd tell him who's top to bottomus. The Wizard of OZ ;-)

Expand full comment

Personally, I think it is taking evil to a whole new level. Especially when government is involved.

Expand full comment

What would happen?

ALL HELL WOULD BREAK LOOSE!

Is this a trick question?

Expand full comment

There appear to be reference notes in the article that are not attached. Will you please make those available?

Expand full comment

Clickable Footnotes and bibliography needed, bro.

Expand full comment

This article would have been printed in Sunday papers and open news on TV, if we lived in a better world, free from corruption and bad actors. Of course, that's just wishful thinking.

Expand full comment

Here is the full discussion I had with ChatGPT yesterday about vaccines.

https://drive.google.com/file/d/1hMagKOs5VxtoYVJMD-q3X0Gsjlv0KO9A/view?usp=sharing

Expand full comment

Having read this response I conclude that chatgtp is completely biased and can't believe that so many people believe it!

Expand full comment

That is a typical response, but you can coax different response from the program with skilfully worded prompts. Unfortunately, the default outputs are all most users will see.

I for instance, have got ChatGPT to come up with responses admitting that the deep state exists and does commit crimes, and that State Crimes Against Democracy (SCADS) are a real thing. It takes some work however.

Expand full comment
Nov 10·edited Nov 10

So the question is if it's initial response is what it is programmed to give, and you can coax it to agree with, confirm or admit almost anything, can you trust those replies any more than the initial one. I have a suspicion that if you made up data or information and reported it as true, it would then give you an answer based on data that it hasn't confirmed.

Expand full comment

Here is more, my latest and most comprehensive discussion on the topic of AI bullshit.

https://ephektikoi.substack.com/p/understanding-llm-ai-limitations

Introduction

The increasing reliance on large language models (LLM AI) in fields such as education, journalism, and even personal decision-making raises important questions about the epistemological and ethical implications of such technology. These models, trained on vast datasets and designed to mimic human conversation, appear sophisticated, often delivering articulate, coherent, and seemingly accurate responses. Yet, this coherence can be misleading. While language models can simulate the flow of human thought, they lack true understanding and operate primarily on probabilistic associations within their training data. This essay explores the limitations of LLM AI, examining the risks and misconceptions associated with human-like interactions, the inherent biases in AI outputs, and the ethical challenges of promoting a technology that, while useful, can also distort reality and stifle diverse viewpoints.

Expand full comment
Nov 11·edited Nov 12

I just thought of another example that would illustrate the Gell-Mann Effect you mentioned in your November 9 article. I started throwing and studying boomerangs since I was maybe 10 years old, and have a large collection, including long ranges boomerangs, which were my favorite, although I haven't thrown in decades (one reason it is needs a very large field for those longer range boomerangs, which are hard to find in metropolitan areas). I have studyied the science of what makes them return extensively. So maybe about 12 years ago I was reading an Ask Marilyn column--Marilyn Vos Savant supposedly had the highest recorded IQ as a child--and someone asked her about boomerangs, and I think she gave a superficial or flippant answer of some sort, but it showed she did not know how boomerangs worked. Her logic ability is very good, but trying to apply logic to a subject where one lacks the technical knowledge often doesn't work, as in this example. I guess that would be both an example of the Gell-Mann Effect (hopefully, not in my case) and analogous to AI answering a question where it doesn't have all the facts. So does that mean that Marilyn has AI?

Expand full comment

Very a propos anecdotes I think.

Expand full comment

It is amazing that it can sort of interpret context and reply seemingly appropriately. But that that can be deceptive or misleading as you point out reminds me of when I used to watch William F. Buckley on Firing Line. He had a huge working vocabulary and was extremely articulate, and usually his arguments made sense, but one time a response he gave, in his very articulate manner and using his large vocabulary, if you actually analyzed what he said it had logical fallacies or was just wrong, I forget which.

Expand full comment

Well these are good questions. I have done a lot of digging on this topic and published a lot on my substack. The responses are not exactly programmed explicitly, but LLM AI are shaped to give certain responses. My most recent short piece on this topic was today: https://ephektikoi.substack.com/p/pattern-without-precision-the-ai. With different prompts, you can shape the response. I guess there are some deep epistemological questions involved that go beyond any obvious answers. Still, you you most often get mainstream opinion by default. The short answer is it goes with the most common arguments in its training data. This is a result of its statistical word chasing. Why it works at all is still a great mystery I believe. I use it extensively, but have to work around its proclivity to give mainstream positions on anything remotely controversial.

Expand full comment

A massive leak of internal e-mails and memos from Germany’s Robert Koch Institute (federal agency and research institute for controlling infectious disease) reveals that the institute's scientists understood that virtually every aspect of the COVID-19 pandemic response was NOT guided by science, but by political machinations to spread fear, control the population, and promote the experimental vaccines. On November 2, 2024, Professor Stefan Homburg gave a presentation on the leaked documents in the German Parliament.

Since the beginning of the pandemic, I have suspected that governments throughout the world were committing the greatest organized fraud in history, but I still found Professor Homburg's presentation to be absolutely breathtaking in the sheer ruthlessness of the lies and manipulation revealed in these leaked documents. Perhaps the most perfidious is a memo from September 28, 2020.

Translation: 28.09.2020: FDA approval [of COVID-19 vaccines] before the US Elections is not desired, also not by European authorities.

In other words, both U.S. and European public health authorities were afraid the COVID-19 vaccines—which had been heralded as the forthcoming saviors of mankind—could help to get Donald Trump elected if they were approved before the election. And so they found it expedient to withhold approval until after the election, even though they claimed the vaccines could save millions of lives.

I strongly encourage everyone to watch the video presentation (in German with English subtitles) and to share it far and wide.

https://www.youtube.com/watch?v=-j7qjedth3A

Expand full comment

I knew it was all co-ordinated worldwide from the start. These organisations like the WHO are corrupt to the core along with all governments.

Expand full comment

Biochemical weapons- simple. We have been being poisoned in America since 1900. There are US patents to prove it.

Expand full comment

Autism was unknown in my country of birth, than communists Czechoslovakia , the only vaccine in school I rejected. Would this be possible in California? Americans you are under the spell of medical tyranny. Even the freaking communists I hate cared for our health, natural prevention was promoted in school and all media , not drugs. We did not have all those chronic condition America has. When sick with some of those harmless childhood viruses , it was to stay home , drink a lot of tea with lemons, herbs no doctor. Do not know of any child to go to a doctor when sick. To make sick is a big business in America.

Expand full comment

Thanks I shared Thompson,s experience and I was confronted with the Andrew Wakefield fraud discrediting those who question the safety profile , how do you react to such criticism

Expand full comment

That would be a dangerous proposition. We might learn the truth of WW1, WW2, The Great Depression, Korea, on and on, it would be the end for a lot of, hmm, familial interests.

Expand full comment