All right, so this is going to be a different

kind of video.

I'm not running scripted.

It's going to be kind of off the cuff.

OpenAI just dropped this press release about

how it, quote, "Derived a new result in

theoretical

physics."

And I'm sure it's going to end up being all

over the papers and everybody's going to talk

about how it created new laws or any of that

kind of crap.

I happen to have had a physics degree many

years ago from Texas A&M - insert Aggie joke

here - and I happen to know enough about AI to 16 00:00:23,520 --> 00:00:25,600 kind of understand how the computing part

of this might work, and I'm willing to put my

thumb in Sam Altman's eye - figuratively

speaking, of course.

So I figured I would take a minute to try to

head this off and put together a quick video

to help you understand what the heck they're

talking about.

Okay.

So here's this, this press release they put out,

I'll put all the URLs down in the bottom.

"Some physical observables calculated using

textbook methods look terribly complicated,

but turn out to be very simple.

Finding a simple formula is always been fiddly

and also something that I have long felt might

be automatable by computers."

So here's what we're actually talking about,

right?

If you look at the actual paper that they

generated, you'll see that OpenAI is mentioned

here, and it's mentioned here, and it's

mentioned in the acknowledgments.

So it's not like OpenAI was doing all of this

stuff.

What happened as near as I can tell was human

authors worked out the amplitudes for longer

equations by hand, getting a very complicated

expression, right?

And then they gave that expression to GPT 5.2,

GPT 5.2 Pro, and it grinded for 12 hours

and then came up with a simplified version of

it.

This is the kind of thing you probably could

have gotten Mathematica to do if you're willing

to spend the money on all the API calls that

would have taken to grind through all this

stuff.

But basically what happened, and you don't need

to understand any of this stuff, right?

So you've got all this stuff here, and then you've

got all this stuff here and down here.

And what happened was - let me change colors -

that OpenAI, ChatGPT 5.2, took all of that

stuff and ground on it and found that this

particular equation was a simplified version

of all of the equations that I've got there in

yellow and some on the previous page.

So it's basically just a finding the generic

form of a bunch of different equations that

people had already worked out, and you can do

that basically with brute force.

I'm not saying it's easy, and I'm not saying

this isn't a cool result, but this is not

a result in physics that only an AI could have

come up with.

This is just a thing that spent a lot of time

doing brute force on math until it found an

equation that was a simplified version of a

bunch of other equations that it had.

That's really all it is.

And I'm happy that AI can simplify equations.

I'm happy that physicists can turn to computers.

They have been for decades now.

But I'm happy that AI is a new tool that

physicists can use to make the math simpler,

and then they can use that simplified math to

try to go and look at other things.

I mean, it's good that OpenAI can do this, but

my guess based on looking at this, especially

up here at the top where it says this was on

behalf of OpenAI, my guess is one way this

could have happened is the OpenAI folks could

have gone around a bunch of universities,

found somebody who knew of some equations that

would be cool to simplify or that they

thought might have a simplified form, asked

people to give them the longer versions of

those equations, took those, fed it into the AI,

got a simplified version out, gave it

back to the physicists, the physicists did

something with it, and then OpenAI is basically

in this press release trying to claim the

credit for it.

The "ChatGPT 5.2 derives a new result for the

theoretical physics."

So "derive" is a mathematical term for "turned one

equation into another", basically.

This is not like Newton's fourth law or

something, right?

This is just somebody took some complicated

math, fed it into the AI and the AI come back

with a simplified version of it.

People could have done that.

It just would have taken a lot of time and

effort, and I'm glad that the AI can make

that faster, but this is, I'm sure this is

going to be presented as the AI figured out

a lot of physics at the people that only AI

could have figured out or that humans weren't

capable of that kind of stuff.

It just simplified some math, and it's good

that it simplified some math, but I'm sure

it's going to get blown way out of proportion

because this headline sounds a lot more grandiose

than what the actual reality actually is, and

then some idiots out there are going to take

this headline, and they're going to not

understand what any of the other stuff is, and

they're

just going to run with this headline, and I don't

even want to know how bad the headlines

of the articles or the videos that people write

based on this headline are going to be.

I'm sure we're going to hear that, and it's

going to get combined with Sam Altman's bit

about how "ChatGPT is supposedly a PhD in every

subject" garbage, and oh, by the way, there's

a bit in this paper where it talks about a "new

internal OpenAI model."

So as near as I can tell, OpenAI built a

special model to do part of this, which is

fine, and it's great, and I'm sure it cost

them a lot of money, but it's not like we're

talking about ChatGPT being a PhD physicist

and coming up with stuff that physicists weren't

capable of doing.

This is not "chatbot discovered as a fundamental

force of the universe" or something like that.

Despite the headlines, you're probably going to

see in the next few days, so I figured

I would just say this real quick, get this out

there before all the headlines happen,

hopefully somebody, when they see the headlines,

point to a video like this, or hopefully some

other people will say things about it.

Thanks for watching.

Sorry, this wasn't very organized.

I was setting up to record another video, and I

saw this press release come out.

It was like, "all right, I'll just go ahead and

record something quick about it."

I haven't taken my normal time to vet and

research and make a script and edit it several

times and all that kind of stuff.

This is all just unscripted off the cuff, but

when I saw that press release came out,

I thought, "you know, I ought to get something

out there since I can, and hopefully some

people will see it and not everybody will be

fooled by whatever the headlines are going

to be."

Thanks for watching.

Let's be careful out there.