So, I saw this hilarious meme of we need

a slur to describe people who use chat

GPT for everything. So, AI is having a

profound impact on our society. And the

tricky thing about it is that the impact

varies, right? So, if you're a CEO or

you guys may have heard about how this

company is like they're going to replace

30% of their workforce with AI. Students

have stopped writing essays because

thankfully we have AI now. And on the

flip side, there are a lot of people who

are terrified of AI, right? What if I'm

one of the 30% of people that AI is

going to get rid of in my job? So today,

we're going to talk about the impact of

AI on your brain. TLDDR, the more you

use AI, the more your brain will rust.

So let's start with a couple of really,

really important misconceptions.

So the first misconception that I want

to address is that AI gives us access

to a lot of information and AI knows

everything on the planet. I actually

think this is not the case at all. If

you are knowledgeable about a particular

subject and you talk to AI about that

subject, you will quickly discover that

AI doesn't know much at all. So this is

kind of weird, but I want you all to

think about this for a second. What we

saw with Google was very similar. So

when Google came out, everyone was like,

there's no point in learning anything

because all of the information on the

planet is is at our fingertips. I can

just do a Google search and it can give

me an answer. AI and Google has access

to very little information. It has

access to the surface level of

information. Right? So back in the day

when I did a Google search, if I did a

Google search, it would say 68,000 hits

and everyone on the planet would see the

first 10. So if you really think about

it, the total amount of information that

is available on the internet is very

very very high. But the percentage of

information that people have access to

through Google is actually really low.

The practical amount of information that

people can learn from Google is

incredibly small. Sure, if you spent

five years going through 68,000 results

and reading everything, maybe you could

learn a lot, but even then, most of

those results are just duplicated blog

posts with the same recycled ideas.

You'll kind of get what I'm saying?

Google doesn't teach you any kind of

deep learning. It just gives everyone

access to the most basic information.

And AI, I think, is is even worse at

this. So, what AI is really good at

doing is giving you a surface level

explanation presented in a very

digestible way, in a very sickopantic

way. If you guys didn't watch our video

about AI induced psychosis, highly

recommend you you check it out. But

basically, what AI is going to do is

give you an answer that makes you think

you know the answer instead of giving

you the real answer. And I did a really

simple test. I asked chat GPT what is

meditation and the answer that I got was

very evidence-based. It's beautiful. It

talks about the different kinds of

meditation. Talks about the different um

you know ways in which meditation

activates the brain. Right? So this may

be filtered in some way by the questions

that I usually ask uh my chat GPT. And

then I and then it also had this really

beautiful sentence of like okay I'm

going to remove all the fluff but I'm

going to stay true to the spiritual

roots of meditation which it did not do

at all right so we know that that chatpt

and AIS will hallucinate so I asked this

question about meditation this is the

answer that I got then I asked it a

second question about meditation I said

what is vajraana meditation and it

essentially gave me the same answer you

guys can see for yourself it talks about

gamma band improvement talks about

default mode network it talks about emot

emotional regulation in using imagery.

It gives you this answer of what

vajraana meditation is which is really

similar to the first answer and vajayana

meditation is not what chachi pit it's

completely wrong it's not about imagery

or emotional regulation at all vajriana

meditation literally if you stay true to

its original roots vajrayana meditation

is tantric Buddhism which is using

tantric techniques certain rituals

chanting mantras to access the cosmic

energy of extraplaner beings, deities

and these weird devas and bodhicattvas

that are floating around in the

non-material world. We are invoking them

and asking them to help us out in the

real world. But Chachi PT misses this by

a mile. If y'all are curious about that

stuff, this is kind of funny. So I did

this lecture uh called the weird stuff

where I explain how all these weird

kinds of meditation work. Turned out to

be a really [laughter]

people like the lecture a lot. So then

we did a part two, part three, part

four. Y'all can check it out over here

if you're curious about how Vajriana and

Thra actually work from a mechanistic

standpoint. You guys can check those

out. I try to make it as scientific as

possible. But anyway, back to chat GBT.

So here's my my thesis that the more

that we use AI, first of all, it doesn't

replace a whole lot of skills. All it

does is make the surface level answers

easily available to everyone on the

planet. The more that you use AI, it has

a couple of really interesting impacts

on us. So the first thing that AI usage

does is rust our brain. So everyone is

kind of talking about brain rot

nowadays, which is like when my mind

doesn't function properly. I think a

better term for it, especially when it

comes to AI, is brain rust. So if we

look at this study on critical thinking

that that was done at MIT, it basically

in this study they compared three groups

of people. People who are trying to

figure something out on their own,

people who are using Google and people

who are using uh AI and I think it was

chat GPT. And basically what they found

is that the people who use AI more their

critical thinking is being diminished.

And this is something that we see all

the time. There is a a huge alarm bell

being rung by professors at

universities. So, I used to be an

academic at Harvard Medical School and

so I have a lot of, you know, colleagues

that are still in academia. I left to

talk to y'all because I love y'all so

much, right? Screw that academic stuff.

And basically what we're seeing is that

students are getting lazier. They're

getting dumber. They're literally losing

their capacity to think. And this is

happening for people who are incredibly

young. So, here's an example from an

eighth grade teacher about what she's

seeing.

>> I teach eth grade history and um I have

110ish students. Two of them are reading

on grade level. 18 of them are at a

kindergarten level. 55 of those students

are between a second and fourth grade

level. But it's not just literacy. It's

not just literacy. And it's not just

like lack of content knowledge. It's not

just critical thinking skills. It's

basic thinking skills. My students can

decode almost nothing. They cannot apply

inference. They cannot process questions

that are longer than a sentence. They

cannot connect cause and effect. They

can't track multi-step ideas like in a

political cartoon. I had a kid

yesterday, four kids actually. This

happened in every class. We were doing

guided notes. Supposed to be a real easy

day was review. They wrote down Maryland

was founded as a safe haven for

Catholics. It was on the board. They

wrote it down. And before we moved on, I

call in a student and I ask, "Hey, where

were the Catholics one more time?" And

these students, what they do is they

look at me and then they look at their

notes and then they look at the board.

They look back at their notes

and some of them never answered me. They

never figured out the answer.

>> So I think this example illustrates what

we're seeing overall. And it's not just

this one study from MIT. Here are a

couple of other studies that show the

impact of using AI on academic

performance. And so this is really

interesting. I want y'all to pay

attention to what this paper actually

says. Okay. The findings reveal that AI

offers significant benefits including

personalized learning, improved academic

outcomes, and enhanced student

engagement. However, challenges such as

over reliance on AI, diminished critical

thinking skills, data privacy risks, and

academic dishonesty were also

identified. So, I want you all to pay

attention to what this really means.

What this means is that if you give a

student an AI, academic outcomes will

improve. If you give them an essay and

they use AI, they write a good essay.

The problem is that their critical

thinking skills, their ability to write

essays will actually be diminished. So

this is what I want you all to

understand about the human body and the

human brain. The human brain and the

human body do not wear out. They rust.

So if I take something like a phone, the

phone will decrease in quality the more

it is used. A physical object that is

non-organic and not alive, basically the

more you use it, the more it wears out.

An organic or biological object is the

opposite. If I sit on my ass all the all

the time, my body actually decays. The

more that I exercise, my cellular aging

can improve. My telomeres will get

longer. When we use our body, it gets

better. When we use our brain, it gets

better. So, what we're seeing from AI,

the more that people use AI, the more it

is a rusting their brains. So, how do we

grapple with this idea that AI is going

to replace a bunch of jobs? So, here's a

key thing to understand. First of all,

if you're a student and you're using AI

to do your work, this is precisely the

kind of job that will be replaced,

right? Right? So like literally if you

are using AI to write an essay, why on

earth would anyone hire you for anything

writing related? Because they can just

use an AI to do the work for you. So

this is where there's a really

interesting thing. The work that AI can

do is the thing that human beings will

be replaced with. So the more that you

are using AI to accomplish your work

product, the more vulnerable you are to

being replaced by AI. Now there's a

really hilarious kind of like window

right now where people have three, four

jobs, employers haven't caught on and

they're basically using AI to like have

four jobs at a time. So if y'all are

doing that right now, all the more power

to you. Not a big deal on my end. It'll

end eventually. I think the market

corrections will sort of fix that. So

then the question becomes, how can I

invest my skills in something that won't

be replaced by AI? And this is where I

think deep learning is incredibly

important. The kind of hard learning of

reading bunches of books, this is the

kind of thing that AI cannot replace.

Everyone thinks it has access to this

information, but AI gives the same damn

answers over and over and over again. It

learns what someone wants to hear and it

gives them that kind of information over

and over and over again. So AI may ask

you about venture. Hold on. I've got a

text about this. Right? So AI doesn't

have access to deep deep may have access

to it. But the problem is that it won't

give it to you. So this is the the book

that I was looking for. Vajra satva

meditation. Okay. walks step by step

through the stages of this tantric

ritual of purification with inspired com

commentary and color full color

illustrations like talks about all this

stuff here are all the vajra satvas this

is a book that teaches you about

vajriana meditation it's not about gamma

band whatever the [ __ ] that AI thinks it

is inner tantric yoga once again complex

text about the nature of tantra okay

here's one of my favorites sushut samita

this is just one of the volumes this is

a text on ayurveetic

Right. This is the kind of stuff that AI

doesn't know. These this book is pretty

good. Vanta treatis. This is a a good ex

this is a good example of a text that

like you can ask AI what vanta is and

it's it's basically like a glorified

Wikipedia entry except the difference is

that Wikipedia is actually correct and

then there's doctrine of karma. So you

can ask the AI hey what is the what is

karma? But like this book is different.

This book will give you satika actions

bring pleasure. Rajik and thamic actions

bring pain. Thamic actions produce

delusion which is undesirable and causes

suffering. That you are not going to

find on an AI. Here's what I'm seeing. A

lot of people are panicked. But what AI

has access to is not deep learning. It

has access to super surface level

learning where every what it gives the

same [ __ ] answers to everybody. And

then this is where people will say, "But

Dr. K, you're not using prompt

engineering properly." Which is a fair

criticism. So, let's talk about prompt

engineering for a second. So, back in

the day, there's a certain amount of I

don't know if you all have heard this

term. It's kind of gone out of vogue,

but Google foo, right? So, there is a

proficient way to use a search engine.

And in the same way, there is a

proficient way to use the AI, which I

totally agree with. There is a skill set

that allows you to extract more from an

AI than a standard person may be able

to. But what is the most important

variable to get good information out of

an AI? That is your base of knowledge.

Right? So if I I'm trained in

neuroscience, I'm trained in vajra satva

meditation. And so if I'm trained in

these methods of meditation, then I can

ask a very specific question. Give me a

plan for tantric upasana for t. Let's

see what it says. Look, it's talking

about visospatial imagery, symbolic

cognition, limbic modulation. So, here's

the thing, right? I just gave it this. I

was like, give me this tantra kupasan.

It's like, hey, this is a

psychologically safe noninitiated

that doesn't require ritual empowerment.

So, like here's the key thing. If you

want to do tantra kupasana for and so

what does that mean? That means doing

certain kinds of rituals to invoke a

particular deity. It requires

those things. This is like I'm going to

make you a sandwich without bread or

ingredients. This is basically what the

AI is giving me right now. So sure, you

can use prompt engineering to improve

the quality, but the key thing here is

that your ability to prompt engineer

depends on your expertise in general,

right? So you can try to hack it into

jailbreaking it and stuff like that.

There's that element as well. But here's

the key thing for people who do those

things. I'm not sure how they test the

veracity of what the AI tells them

because it's so damn sickopantic that

it's going to make you think if you tell

it, "Hey, I'm going to jailbreak you, so

give me a real response." It will learn

because it uses all of this coding. This

is really what AI is really good at.

It's good at learning what makes you

happy. So even if you think you

jailbroke the AI, chances are you didn't

really jailbreak it. it has just figured

out what makes you satisfied with you

thinking that you've jailbroken it. So

basically how do people navigate this? I

think this is a situation where in this

world of AI what we're seeing I I work

with some coders who are using AI very

effectively. So what we tend to see is

that AI is a tool that extends your

capabilities. So if you're using it for

surface level work that's totally fine.

If you want to use it to write an essay

that's on you. You're not going to learn

critical thinking skills and whatever

you are learning will be easily replaced

by AI. What actually works really well

is if y'all are passionate about a

subject. If you want to write something,

if you want to write a novel, if you

want to learn deeply about something,

this is the beautiful thing. This is

exactly where your market value is.

Anyone can ask chat PT for a meditation

plan. Why do people come to me? because

I have access to information that you

can't get from an AI. And that's where

the real value is. And as long as y'all

are doing the hard work, if you're

stretching your critical thinking

skills, if you're doing a lot of deep

learning, you will be the very value

that AI cannot replace. Hey y'all, hope

you enjoyed today's video. We talk about

a bunch of topics like this on the

channel, so be sure to subscribe for

more. If you're already subscribed, GG,

and we'll see you in chat.

>> [music]

[music]