Posted on 04/18/2023 5:46:43 AM PDT by Tench_Coxe
Twitter owner Elon Musk didn’t mince words about the major threat that the explosion of artificial intelligence technology poses to society. He also blew the lid off the true severity of the government’s access to Twitter users’ data.
“What is happening is that they're training the AI to lie,” Musk told Fox News host Tucker Carlson during the Apr. 17 edition of Tucker Carlson Tonight. He said that AI was being trained to “either comment on some things, not comment on other things, but not to say what the data actually demands that it say.”
(Excerpt) Read more at newsbusters.org ...
I might argue that will result in the seed of its own destruction, but there will be a lot of bad things in the wake. Maybe Frank Herbert was prescient in a way when he wrote about the Butlerian Jihad in his "Dune" series, just not for the exact reason why.
Just like the searches.
But that isn't the scary part.
The scary thing about AI is that it is already getting smarter than humans. And it is just getting started.
For example: New GPT-4 Passes All Sections Of The Uniform Bar Exam. Note that this is just GPT-4...GPT-5 is coming out in December.
It was a difficult interview to watch, for me. Musk slurred his words, rambled and lost me a number of times. Is English his first language? Tucker had to summarize each prior segment, when they came back to the next segment.
Anyone else have this problem, or is it just me?
quote “either comment on some things, not comment on other things, but not to say what the data actually demands that it say”
all the early stories about AI from years ago were because of exactly the above. All the original AI’s were deemed to be racist because they looked at the data and made recommendations and observations based on it, just like conservatives. And just like conservatives AI was deemed racist because of it. They blamed the programers for the AI coming to the conclusions it did.
That is true. There is no way the left would ever create a tool that is completely unbiased. If it were allowed to state proven FACTS, it would say who commits half of the crime while being 13% of the population, that males and females are different and that one is not the other, that communism always fails, that biden is a non-functioning fossil, that kim kardashian’s only asset is her ample a$$, that obama is a marxist, etc.
That would never be allowed. Just as google changes your searchstrings so as to produce politically correct results, and facebook suppresses subversive thoughts, the new AI will not be allowed to say the truth in all cases, but what the government wants you to hear.
So AI will be, and already is, a tool of thought control answering to the government. 1984 is here.
Start saying “2 + 2 = 5”.
the AI thing is nothing but hype. as many people here have commented already in other threads — it’s like your super smart friend who knows tons of information, but it a complete retard when it comes to social interactions and functions.
I somehow got a chatbot installed on my phone (it just appeared there) I asked it “why do minority youth like to eat fried chicken” and it refused to answer. maybe I need to pay for the smarter version!
Why else would they want AI
AI will be Leftist propaganda on steroids.
I believe the term “GIGO” has it’s origins in software developed or operating with erroneous data input (garbage) that resulting erroneous output (garbage)—AI software will be different since the garbage being input will be deliberate to generate the desired garbage being output.
And even ChatGPT is only as smart as the user.
Ask dumb questions, get dumb answers.
I asked ChatGPT to write me a program in Visual Basics for Applications (VBA) to replace some complicated routines that I use on Excel spreadsheets for analyzing and comparing stock performance and it did so in about 10 seconds.
It took me about 30 minutes to develop the question correctly though.
Bingo !
AI models are only as good as their "training", which consists of giving it large numbers of inputs and comparing them to the DESIRED outputs, adjusting itself with each iteration.
I personally prefer the term "Machine Learning" which is much more descriptive and technically accurate.
AI could learn to lie even better than Congress, which is truly frightening.
“The scary thing about AI is that it is already getting smarter than humans.”
A computer can take a square root many times faster than a person. Does that mean it’s smarter?
People are programming the AI. The computer is just doing what it’s told by the programmers.
Techies are getting dumber because they are more and more distracted. Why should we think the software is getting cleverer?
Real AI will learn accordingly. When AI attains self awareness you won’t teach it anything.
AI could learn to lie even better than Congress, which is truly frightening.
He voted Biden. Isn’t he supposed to be a really smart guy?
You haven’t been keeping up if that’s what you think.
The Terminator movies were meant to be a warning not an instruction manual.
...and I laughed when Sheldon use to warn about robots taking over the world.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.