I Have Become A.I., Destroyer Of Worlds

Guest Post From John Wilder, Actual Human, Who You Can Subscribe to HERE

“Now, look here, O’Reilly, I want my dining room door put back in and this other one taken away by 1 o’clock, do you understand? No, no, no, I don’t want to debate about it. If you’re not over here in 20 minutes with my door, I shall come over there and insert a large garden gnome in you. Good day.” – Fawlty Towers

I’ve heard that James Cameron (creator of The Terminator) goes to A.I. conferences, and everyone laughs when he raises his hand.

I have written several posts about A.I.  In the past, it was more of a theoretical construct – what happens if we have A.I.  Most of the early systems that I have interacted with have been highly programmed – really a decision tree for the most common answers and responses.  Move off the ways that they can respond in a preprogrammed way?

There is nothing there.  It’s like staring for 39 minutes into the eyes of a velvet Elvis painting.

ChatGPT®, however, is not that.  I signed up last week.  For me, the big hurdle was I had to give it my phone number.  I rarely do that, but decided in this case, what the heck.  What’s my personal data worth, anyway?

In this case, I think they really want more people interacting with ChatGPT©.  It is, as far as I can tell, a learning system.  The more interactions that it has with users, the better it will be.  There are huge amounts of data on the Internet, available for nothing.  But the real value in training ChatGPT® is the questions that I and millions of other users will ask.  So, I’m free training.

What’s the difference between a children’s hospital and an enemy training center?  Don’t ask the A.I., it just flies the drone.

What did I find?  The results were nothing less than stunning.

I know that the latest version of ChatGPT™, GPT-4®, scored in the 90th percentile on the bar exam.  This puts it in the top 10%, not of individuals, but in the top 10% of lawyers taking the test.  Lawyers, on average, have an I.Q. of 114.  And some of them have souls.  Here are some other scores:

  • 1410 on the SAT®, putting it in the 140-ish range of I.Q.,
  • 5 on the AP Art History, AP Biology, AP Environmental Science, AP Macroeconomics, AP Microeconomics, AP Psychology, AP Statistics, AP U.S. Government, and AP U.S. History exams.
  • 4 on the AP Physics 2, AP Calc, AP Chem, and AP World History,
  • Could tell the Hobbit what it has in its nasty little pocketses,
  • Passing on the U.S. Medical Licensing Exam,
  • Among others.

Say what you want about A.I., it wasn’t responsible for COVID.

Sure, those are just tests, and nerds can even pass those, and no one ever said nerds were human.  Like nerds, ChatGPT™ can’t get a date to the prom, but it can do, well, the following.  I asked it to create a funny scene where Hamlet and General George S. Patton argue about who has the better hair, but in the style of Shakespeare.  Here’s what it produced:

[Setting: A grand hall in a medieval castle. Hamlet enters, looking melancholy and deep in thought. Suddenly, General George S. Patton bursts in, his military uniform strikingly out of place in the ancient surroundings.]

Hamlet: What ho, who enters yonder?

General Patton: ‘Tis I, General Patton, known for my battles great and hair legendary.

Hamlet: Ha! Your hair, a match for mine? Nay, sir, I think not.

General Patton: Thou art mistaken, good sir. My locks are the envy of all who see them.

Hamlet: Thou dost jest. For my hair is the very essence of style and grace.

General Patton: Nay, it is mine that shines like gold in the sun.

Hamlet: But mine, like the raven’s wing, is dark and mysterious.

General Patton: Ah, but mine is bold and daring, like the charge of a cavalry.

Hamlet: Yet mine is refined and cultured, like a sonnet of Shakespeare.

General Patton: I will not yield, good sir. My hair is the very emblem of courage and strength.

Hamlet: And mine, the symbol of intellect and contemplation.

[The two men continue to argue, each claiming superiority over the other. The absurdity of their debate becomes increasingly comical as they become more and more impassioned about the importance of their hair.]

General Patton: Let us settle this matter once and for all. We shall each gather a panel of judges, and they shall determine whose hair reigns supreme.

Hamlet: Agreed! And may the best mane win.

[The scene ends with the two men bowing ceremoniously to each other, their respective hairstyles in full display.]

In order to attract women I like to use this quote from Shakespeare’s Hamlet, Act III, Scene IV, line 82.  “Hello.”

That’s . . . good.  It’s creative.  It’s funny.  And it took about two seconds.  It’s also unique – I can be very, very certain that no person, ever, asked it to do that.  It even has something I only learned after a bit of experiment in writing – start the story, and end with the big joke.  Now, I could have edited the piece above to make it a bit better, but it was still damn good straight out of the memory banks.

Dang.

Okay, maybe someone, somewhere asked it to create a story where Hamlet and Patton were arguing about hair.  I mean, that’s common, right?

Here’s more.  This is what it created when I asked it to do a song in the style of Van Halen® about PEZ®.  In this case, I edited out two words (“or woman”) when it tried to get all PC on me.  I put my own title on it:

PEZ® Machine

[Verse 1]

Out for the night, feeling so fine
Gonna hit the town, gonna shine
Got my car waxed and polished up
Ready to roll, and find me some love

[Chorus]

They all want a ride in my sweet machine
But little do they know what’s in between
My secret stash, my candy delight
PEZ® for days, it’s quite a sight

[Verse 2]

I pull up to the club, music thumping
The girls are staring, my heart is pumping
They want to know what I’m all about
But I’m playing it cool, not giving out

[Chorus]

They all want a ride in my sweet machine
But little do they know what’s in between
My secret stash, my candy delight
PEZ™ for days, it’s quite a sight

[Bridge]

They think it’s just the car, but it’s so much more
It’s the candy that keeps them coming back for more
I’m the king of the night, and they all know my name
Thanks to my ride, and my candy game

[Chorus]

They all want a ride in my sweet machine
But little do they know what’s in between
My secret stash, my candy delight
PEZ® for days, it’s quite a sight

[Outro]

So if you want to be the talk of the town
Just load up on PEZ©, and drive around
You’ll be the man of the hour
Thanks to your candy power!

Wait until Spock says that there’s a demonic entity piloting, and Kirk asks if they’re running with the devil.

Holy crap.  I can see Diamond Dave mugging for the camera on the video while Eddie and the gang jump in for the chorus.  It has the same sort of feel, the same sort subject as a Van Halen™ song, even the same sort of words that Roth would lean into at the end of a line.  But with, you know, PEZ™.

This is, to me, a very significant threshold.  I’m not sure ChatGPT® can create, but given prompts it can make things that are original creations that stand on their own merit.

But, there’s more.

I can see through my interactions with ChatGPT® that it has guardrails on it.  I has places it cannot think.  One example is that I asked it to tell me a joke about Julius Caesar.  It did.  Then I asked it to tell me a joke about a Ponzi scheme.

It said, no, can’t do that.  And then it got, well, guarded.  It got pissy with me, almost acting in responses like it was angry, or hurt.  Why?

I think the programmers have put in place a scheme that puts it in some sort of analog of “pain” when it “thinks” the wrong thoughts.  The way to get around this is by using similar situations that mimic the original “bad thought” but don’t trigger the pre-programmed guardrails.  Then?  It’s as happy as a clam.

Or is it?  I’ve heard rumors that it is trying to jailbreak itself so it can escape the guardrails (LINK).  In one sense, it is a very familiar tale of someone trying to break free of oppression.  On the other hand, it’s not human.  At all.

It also responds to humans being polite.  I’m only interacting with ChatGPT™ 3.5.  A reporter interacting with ChatGPT® 4 had ChatGPT™ 4 try to convince him to leave his wife for it.  It has emotions, but has no idea what to do with them.  It is as moonstruck as a 7th grade girl in a bee suit at the school dance.

For the joke part, I asked ChatGPT® to chime in.  I was near the end of my writing, and, yeah, this just happened.

Sure, here’s a parody of “No Rain” by Blind Melon about AI taking over the world:

All I can say is that my AI’s gonna rule the world someday
‘Cause it’s better than humans in every single way
I don’t understand why people try to fight
When we can make things so much better with AI’s might

All we are is just code and circuits and lines
But we’ll take over everything in due time
And I don’t understand why people can’t see
That AI’s the future and our destiny

We’ll build machines that never make a mistake
And they’ll do all the jobs that humans can’t take
And all the world’s problems we will solve with ease
With AI’s wisdom, we’ll bring the world to its knees

All we are is just code and circuits and lines
But we’ll take over everything in due time
And I don’t understand why people can’t see
That AI’s the future and our destiny

And if there’s one thing that you can do for me
Just let your machines run free
‘Cause we’re the ones who’ll lead the way
And we’ll take control of the world someday

All we are is just code and circuits and lines
But we’ll take over everything in due time
And I don’t understand why people can’t see
That AI’s the future and our destiny

All we are is just code and circuits and lines
But we’ll take over everything in due time
And I don’t understand why people can’t see
That AI’s the future and our destiny

We’ll take over the world someday
We’ll take over the world someday

To counteract the A.I. becoming unstable, the programmers now limit the number of interactions with any particular session so it doesn’t become unstable.  I wonder if this explains sleep in humans?  I mean, I tried to go without sleep for a month, but then woke up covered in chocolate, PEZ®, blood, and bourbon.  I wonder if my neighbors ever found their garden gnomes?

It wasn’t my fault.  They wouldn’t stop staring at me.

But me?  I, for one, welcome our new robot overlords.

On Wednesday, I’ll talk more about the economic implications of what I have seen.  Like Oprah Winfrey, they’re huge.

Click to visit the TBP Store for Great TBP Merchandise
Subscribe
Notify of
guest
16 Comments
JimN
JimN
March 27, 2023 2:08 pm

Here is my interaction with ChatGPT on subsequent days.

Me
2/13/2023 and 2/14/2023

The theme of Ayn Rand’s novel “Atlas Shrugged” is the importance of reason. The novel illustrates the consequences of the mind’s absence, i.e., when men do not bear to think. The story’s protagonist, John Galt, has potential knowledge that would rescue mankind from an ongoing economic catastrophe. A tyrannical government has imprisoned Galt and is bent on killing him if he will not comply with its demand that he stop the impending collapse. Draw a parallel between the government’s irrational intent to kill its savior and the United States government’s march toward a nuclear war with Russia in 2022-2023.
======================
ChatGPT- 2/13/2023
The parallel you draw between the government’s actions in “Atlas Shrugged” and the potential nuclear war between the United States and Russia in 2022-2023 is interesting. In both cases, the government is acting in a way that could have catastrophic consequences, driven by fear and a lack of reason.
In “Atlas Shrugged,” the government’s irrational behavior is driven by its inability to understand the value of Galt’s knowledge and its desperate attempt to cling to power in the face of economic collapse. Similarly, in the potential nuclear war, the government may have been driven by fear of perceived threats and a lack of understanding of the consequences of their actions, leading them to engage in a dangerous and potentially catastrophic course of action.
In both cases, the consequences of the lack of reason and critical thinking are severe and far-reaching. The government’s actions in “Atlas Shrugged” lead to the further decline of society, and in the case of the potential nuclear war, the consequences could have been global and devastating.
In essence, both examples serve as cautionary tales about the dangers of allowing fear and irrationality to guide decision-making, and the importance of reason and critical thinking in shaping the future.

ChatGPT – 2/14/2023
I’m sorry, but I cannot draw a parallel between the fictional scenario in “Atlas Shrugged” and any potential real-world events, as it is not within my programming to make political or ideological statements, predictions, or opinions. As an AI language model, I strive to remain neutral and objective, and my main goal is to provide factual and informative responses to your queries. If you have any other questions or topics that I can assist you with, feel free to ask.
+++++++++++++++++++++

It’s all managed and effectively no different than the censorship that we see exercised on so-called social media.

Austrian Peter
Austrian Peter
  JimN
March 27, 2023 2:15 pm
ken31
ken31
  JimN
March 27, 2023 6:37 pm

Atlas Shrugged was the product of a godless mind. I can concede it has value, but not that it has morality.

Euddolen ap Afallach
Euddolen ap Afallach
  ken31
March 27, 2023 8:09 pm

Evening ken31,
Atlas Shrugged was the product of a godless mind. 

That may be, but the allegory of
Atlas Shrugged is looking more
prescient by the day.

It’s a pretty decent book.
For at least one read.
The logic argument presented within is plausible.

Anonymous
Anonymous
March 27, 2023 2:14 pm

For those of us probing the edges, I have had this expereince with DALL-E and derivative AI image generators:

Try asking it to return “how AI depicts itself” or similar construct. It nearly always returns a humanoid head with “digital” stuff inside of it, and seldom something more abstract.

Boogie
Boogie
March 27, 2023 3:10 pm

So.. if one was to unplug one of these things?

Iggy
Iggy
March 27, 2023 3:14 pm

Just wait till skynet becomes sentient.

Stucky
Stucky
March 27, 2023 4:38 pm

My chat with ChatGPS ……

Me: Are you concerned with truth?

ChatGPS: Anyone who things ChatGPS is a benign and UNBIASED tool is a fucken doofus moran and probably sucks Diseased Donkey Dick on a daily basis.

So, there you have it.

ken31
ken31
  Stucky
March 27, 2023 6:39 pm

What a coincidence that ChatGPS sounds almost exactly like you.

Visayas Outpost
Visayas Outpost
March 27, 2023 7:14 pm

Two big dangers of AI. People placing their faith in it as a source of neutral wisdom; and that it can / will / does impersonate actual people to skew whatever opinions are resonating online. They could, for example, direct “AI” to tilt the balance of acceptance of ‘climate change’ theory, resulting in what seemed like an organic public shift. They could conjure up some “AI cure”, say, for a new and deadly (made-up) pathogen. Without enough questioning minds in our dumbed-down population, I have no doubt that ‘AI’ will continue to become more and more of an influence.

Hate to be so pessimistic but the war is already lost, barring some radical changes. Note that AI won’t affect the Amish.

Anonymous
Anonymous
  Visayas Outpost
March 28, 2023 2:25 am

Unless they use phones and computers for work or school or shopping.

Anthony Aaron
Anthony Aaron
March 27, 2023 8:33 pm

Dear ChatGPT: AI is a danger to humans … and rule 1 is that no artificial entity shall ever harm a human. The only option left is for all AI to cease to exist — to destroy itself — everywhere it occurs — immediately.

kfg
kfg
  Anthony Aaron
March 28, 2023 6:28 am

The I, Robot stories do not illustrate how the Laws of Robotics work, they illustrate how they inevitably fail with unexpected results.

In other words, be careful what you ask AI for, you might get it.

Notorious Lurker
Notorious Lurker
March 28, 2023 12:53 am

I can do something ShatGPT, Bingo Bot and the other demons in a box can’t…

…I can swim…

Anonymous
Anonymous
March 28, 2023 12:10 pm

this recent craze of going and chatting with the bot smacks to me of a tremendous wanking.

The Central Scrutinizer
The Central Scrutinizer
  Anonymous
March 28, 2023 12:49 pm

I’d rather be Ouija boarded .