TRANSCRIPTEnglish

The Danger of Elon Musk's "TruthGPT" | Tucker Carlson.

19m 23s3,248 words487 segmentsEnglish

FULL TRANSCRIPT

0:00

now we got a touch on Elon musk's desire

0:03

to Star a truth a GPT now Elon just

0:08

launched a LLC or company and is

0:11

potentially looking to raise money from

0:13

shareholders called x dot a i and

0:17

shareholders would probably be private

0:19

venture-backed shareholders who are

0:21

already shareholders in companies like

0:22

Tesla and Twitter so not something we'd

0:24

publicly be able to invest in but a lot

0:26

of this is inspired by Elon musk's

0:29

frustration with openai turning into

0:32

really a for-profit rather than being a

0:34

non-profit especially since he helped

0:36

start open Ai and there's a lot of drama

0:39

around hey wait a minute what is a

0:41

truthful Ai and why does it matter well

0:44

we'll listen to a clip here from Elon

0:45

Musk in just a moment but I want to

0:47

start by pointing out this uh this

0:51

community notes response that Elon Musk

0:54

had following one of his tweets and so

0:58

Elon Musk tweeted the fall following he

1:00

said NPR literally said Federal funding

1:03

is essential to public radio on their

1:05

own website now taking down what

1:07

Hypocrites now I believe you could

1:10

actually still find this NPR Federal

1:11

funding item on their website they're

1:13

pretty transparent about their funding

1:14

sources but it is a little confusing to

1:17

break apart because some parts they like

1:19

to say oh well in some years we average

1:22

one percent from federal organizations

1:24

and then the reality is they're kind of

1:27

like misdefining exactly where some of

1:29

the money comes from and really a lot of

1:31

this is blowing up right now because

1:33

companies like NPR are basically getting

1:36

notes put under uh of their uh their

1:40

actual Twitter tags showing that uh hey

1:43

they are

1:45

backed by the government for example

1:47

here you have NPR has a tag government

1:49

funded Media or another one that's

1:52

pretty funny that came up yesterday was

1:54

the Canadian Broadcasting Corporation if

1:56

you go to just at CBC they uh they argue

2:01

they were less than 70 percent funded by

2:03

the government so Elon Musk helped label

2:06

them as 69

2:09

government-funded media

2:11

which on one hand is pretty dang funny

2:14

on the other hand makes you wonder wait

2:17

a second whose truth really goes into

2:21

this information and that does create

2:24

some potential problems because for

2:26

example here is a community notes

2:29

response to this that says at the time

2:31

of this tweet there we go at the time of

2:33

this tweet NPR's website dedicated to

2:36

their finances still claim that Federal

2:38

funding is essential so Elon Musk

2:41

actually got community noted on this so

2:43

it's not taken down NPR is still

2:45

claiming that Federal funding is

2:47

essential to the public radio service to

2:49

the American public

2:51

now what I reply to with this because

2:53

it's been such a big deal is I replied

2:57

uh when I can find my reply I think I

2:58

have to press back twice him I replied

3:00

it said unbelievable that Community

3:01

notes left out the percent of funding

3:04

and I said I lost some respect for

3:06

Community notes now if you weren't part

3:08

of the big debate around this that might

3:11

seem extreme hey Community notes is just

3:14

comparing or just suggesting that Hey

3:16

look it's still essential right we're

3:17

just fixing that item from Elon musk's

3:20

tweet and that's fair but all of this is

3:24

based on a massive debate on Twitter

3:26

about what percentage of NPR's funding

3:30

is from the government is it one percent

3:33

is it ten percent is it 15 and in my

3:36

opinion if Community notes is going to

3:37

come in and talk about this Federal

3:38

funding you have to also talk about the

3:41

percentage because even though NPR

3:43

suggests that they're funding their

3:45

government funding is essential you have

3:48

to remember what the word essential

3:50

means and it's something that nobody on

3:54

Twitter is talking about and the reason

3:56

this is important is because of

3:58

congressional Appropriations see that is

4:02

the most important thing that is

4:04

on Twitter and that's pretty common

4:06

because we're generally trying to look

4:08

for the most emotional responses to hate

4:10

or support people and it's really sad

4:13

how frustrated people are on Twitter I

4:16

really feel like everybody if if people

4:17

on Twitter were walking around they'd

4:19

all be walking around like this their

4:21

hands on their hips and with just like a

4:22

big scowl like hmm yeah I support that

4:25

but you're an if that's really

4:27

what Twitter feels like and to me it's a

4:29

cesspool it's disgusting because

4:30

nobody's really trying to delve into

4:33

truth instead you're getting this war

4:35

and so I think it's ironic that Elon

4:37

Musk is trying to create truth a truth

4:40

GPT when actually what you're seeing on

4:43

Twitter is very much a fostering of

4:45

divisiveness now there's this idea that

4:48

well maybe people should be able to

4:50

speak freely but really what you're

4:51

doing is you're creating Echo Chambers

4:52

and even Community notes isn't helping

4:54

with that Echo chamber problem if

4:56

anything it often feels like it leans

4:59

more towards just supporting Elon

5:02

overall in his arguments because again

5:04

it whether or not Elon was Community

5:06

noted for saying oh they took it down

5:08

from their website that's not the part

5:10

that matters the part that actually

5:11

matters the next level part that matters

5:14

is well is Government funding for NPR

5:16

essential but why does that word

5:18

essential really matter well calling

5:20

something essential is more important

5:24

than actually deciding whether or not

5:26

that quote-unquote essential funding

5:29

actually dictates how NPR's news

5:32

coverage is created or dispersed or if

5:36

there is independent journalism calling

5:38

something essential is critical because

5:40

when you look at Congressional

5:42

Appropriations

5:44

Congress wants to appropriate money to

5:48

where things are needed in the

5:51

impression of the other companies right

5:53

and so this in my opinion is where NPR

5:55

is like well we don't really want to

5:57

lose that Federal funding even though

5:59

they could probably make up some of

6:01

their Federal funding from other sources

6:03

or potentially all of it I think it's a

6:05

lot easier for them to say hey Congress

6:08

we're essential don't take our money

6:11

away now that's my opinion it could be

6:12

wrong right it's entirely possible yes

6:15

the money is actually essential but I

6:17

know what NPR does their membership

6:18

drives and they talk about their

6:20

basically expiring coupon code like I do

6:23

because we're having a big price

6:24

increase tomorrow on 420 at the end of

6:26

the day at 11 59 we're getting rid of

6:27

some of the lower bondwell options we're

6:29

getting rid of some of the lower priced

6:30

uh courses we're raising the prices

6:32

we're changing things completely so if

6:34

you want to lock in the lowest price

6:35

jump in but just like people say Hey you

6:38

know I don't want to hear about the

6:39

coupon code all the time you don't want

6:40

to hear about the fall membership drive

6:42

at NPR All NPR all the time but that's

6:45

what they do to raise money so they

6:46

could stay in business and actually

6:47

provide value to their constituents

6:49

right and so in my opinion this

6:51

definition of essential goes to

6:53

Congressional Appropriations and it's

6:55

basically a way saying look it's easier

6:57

for just us to just take some of that

6:58

money from Congress and then not cut us

7:00

out now then that does make you wonder

7:02

is NPR impartial because of that maybe

7:07

personally this video isn't designed to

7:09

defend or support NPR I actually

7:11

personally like listening to NPR but

7:14

that doesn't mean I I agree with

7:15

everything they say just like when I

7:17

listen to the English paper the

7:19

economists they lean super far left and

7:21

they hate Donald Trump do I listen and

7:24

believe everything they say no of course

7:25

not but I use it as inspiration to fact

7:27

check that's the same thing that you

7:28

should do as well so you can make up

7:30

your own opinion but my argument is that

7:33

hey if we really want to get to the

7:36

bottom of Truth in social media then we

7:39

have to provide context to the entire

7:41

discussion not just one tweet and this

7:44

was a missed Opportunity by Community

7:46

notes where Community notes could have

7:48

very simply broken down the debate hey

7:50

in 2010 NPR's funding was this in 2010

7:54

through 2023 the funding was this XYZ

7:56

it's all public Community notes could

7:58

have easily broken that down but instead

8:01

it seemed more interested in saying no

8:03

no it's essential which is basically

8:05

taking the side of Elon even though it

8:08

was a note against Elon is taking it

8:10

aside of Elon and the Elon fans on

8:12

Twitter rather than actually adding more

8:15

detailed support to the debate so I

8:18

question elon's calls for the greatest

8:21

truth GT GPT because I don't think

8:24

Community notes is doing the best job

8:27

yet I think it's useful it's going in

8:28

the right direction but it has some work

8:30

to do and it all is in response to this

8:32

Tucker Carlson piece with Elon Musk

8:34

let's listen to this and see what Elon

8:36

has to say

8:39

and it's not playing quite fun there we

8:43

go to a success of uh open AI uh it was

8:48

I I I put a trans amount of effort into

8:51

recruiting Ilia and he changed his mind

8:53

a few times ultimately decided to go

8:55

with the opening high but if he'd not

8:57

gone without any opening I would not

8:58

have succeeded I really put a lot a lot

9:02

of effort into creating this this

9:03

organization to serve as a counterweight

9:04

to Google and then I kind of took my off

9:06

the ball I guess and uh

9:09

they are now closed Source

9:12

um and

9:15

they are obviously for-profit and

9:17

they're

9:18

closely allied with Microsoft uh you

9:21

know in effect Microsoft uh

9:24

has a very strong say if not

9:27

um

9:28

directly controls uh openai at this

9:31

point so you really have an open-end

9:32

Microsoft situation and then at Google

9:35

deepmind uh other two sort of

9:37

heavyweights in this Arena

9:40

so it seems like the world needs a third

9:42

option

9:43

yes

9:45

so I I I think I will

9:49

create a third option although I was

9:51

starting very late in the game of course

9:53

can it be done I don't know I think it's

9:55

we'll see it's uh

9:57

certainly starting late

9:59

um

10:01

but I will I will I'll try to create a

10:04

third option

10:05

um and that third option hopefully does

10:08

more more good than harm uh like the

10:10

intention with opening eye was obviously

10:13

to do good but it's not clear whether

10:15

it's actually doing good or whether it's

10:18

I can't tell at this point except that

10:20

I'm worried about the fact that uh it's

10:24

been it's being trained to be

10:25

politically correct which is simply

10:26

another way of of being on Truth saying

10:28

untruthful things yes so that's that's a

10:31

bad sign

10:32

there's certainly a path to AI dystopia

10:34

is to train an AI to be deceptive so

10:37

yeah I'm going to start something which

10:38

you call truth gbt or

10:41

a maximum truth seeking AI

10:44

that tries to understand the nature of

10:46

the universe and I think this this might

10:48

be the best path to Safety in the sense

10:50

that an AI that cares about

10:53

understanding the universe uh it is

10:56

unlikely to annihilate humans because we

10:58

are an interesting part of the universe

11:00

uh hopefully they would think about it I

11:03

I think you know because yeah like like

11:06

we like Humanity could

11:09

um decide to hunt down all the

11:10

chimpanzees and kill them but we don't

11:13

yeah and so they go into this argument

11:15

that a like hopefully you could create

11:18

something that cares up most about the

11:21

truth that then doesn't want to destroy

11:22

all the other organisms that exist just

11:24

like we don't destroy whales or

11:27

chimpanzees or otherwise now later in

11:30

this interview Elon Musk talks about how

11:32

social media often comes it turns into

11:34

the Cesspool where everybody's fighting

11:36

for clicks and then you have this

11:37

argument of everybody's trying to fight

11:39

about how the other Clan hates the other

11:42

person rather than focusing on getting

11:44

to the truth or getting to food right

11:46

and that's very much unfortunately what

11:49

I think elon's Community notes can

11:51

sometimes actually contribute to is this

11:54

divisiveness see for example here's a

11:56

comma term and I appreciate you saying

11:58

this Kevin I'm a member of community

12:00

notes and rated that tweet direct

12:02

funding is different year by year and

12:04

indirect funding also fed funds is given

12:06

by different organizations yearly our

12:07

goal wasn't an itemized audit I

12:10

understand that your goal might not have

12:11

been an itemized audit my argument is

12:14

that it should have been and here is why

12:17

I say that because if Elon Musk wants to

12:19

create something that is looking for the

12:22

utmost truth if the answer is truth then

12:26

what we need is data and instead of

12:31

analyzing data or providing data in this

12:34

case Community notes simply provided

12:38

one minor argument in a larger debate

12:43

that reiterated one side you have to

12:47

think about that Community notes

12:49

reiterating that NPR is still saying

12:51

funding is essential is taking one side

12:54

of the debate think about that when we

12:56

draw this out like this if on one side

12:59

you have NPR is gov Shale right and then

13:05

on the right side you have

13:07

NPR is independent

13:10

okay

13:12

they're independent whatever

13:14

so on one side you have NPR's

13:16

independent on the other side you have

13:17

NPR as a is a government chill

13:19

in order to get to the maximum level of

13:22

Truth which is a goal that Community

13:24

notes is trying to do it shouldn't

13:27

Simply Be Community noting something

13:29

that actually reiterates that's it NPR

13:33

is a government chill that's what

13:35

community notes did it it made a

13:38

correction In fairness it made a

13:40

correction but by making that correction

13:41

and not stepping into the rest of the

13:44

debate to actually remain neutral it de

13:48

facto chose a side and it happened to be

13:52

the side that Elon Musk was on and that

13:55

is scary

13:57

so I know it might seem extreme but the

14:00

reality is if elon's going to be trusted

14:03

with a truth GT GPT which anybody can

14:06

create whatever they want then you

14:08

should also do your best to help settle

14:11

disputes if you're a member of community

14:13

notes by providing as much data as

14:16

possible even if you have a little

14:18

preface that says hey government like it

14:21

could be as simple as there's a debate

14:23

going around on NPR here's the reality a

14:27

NPR still says it's government uh or or

14:30

government funding is still essential

14:31

but here are more details on the debate

14:33

NPR's average funding has been eight

14:36

percent from government sources and PR

14:38

says they're independent uh Elon Musk is

14:41

debating creating a threshold for at

14:43

what point of government funding a

14:45

business should be deemed government

14:47

supported and then that should equally

14:49

apply to other businesses if certain

14:52

media businesses are eight percent

14:53

funded by the government then maybe if n

14:55

phase is 10 8 supported by the

14:58

government it should be considered a

14:59

government supported Enterprise as well

15:01

or Tesla

15:03

so that should apply equally and so

15:06

really I'm not here to Shell NPR or Elon

15:09

I'm probably taking the most unpopular

15:11

opinion which is again trying to be in

15:13

the middle but the reality is if you

15:15

don't fully extinguish the debate and

15:18

you just take us uh one side then

15:21

Community notes actually looks like it's

15:24

partisan all of a sudden Community notes

15:27

looks like it's just a bunch of Elon

15:29

shills doing something to support Elon

15:32

Musk even though they were correcting

15:34

Elon

15:35

they still took us that side of the

15:38

greater debate Elon doesn't care about

15:40

being corrected if anything and this is

15:42

the Sinister way to look at it Elon

15:45

could have purposefully made the mistake

15:48

on his tweet to purposefully be

15:50

Community noted and then have Community

15:52

notes reiterate his basic argument that

15:56

NPR is the government shell

15:59

that's touchy that's touchy uh uh so I

16:03

don't disagree Kevin but the point of

16:05

community notes is context to a specific

16:07

claim not the larger argument

16:10

that's a very fair Counterpoint uh I

16:12

agree with you

16:14

I would take it a step further though

16:16

and say if if the broad goal is to

16:21

maximize truth

16:24

than if somebody inputs an argument they

16:27

should get both sides right like for

16:29

example

16:30

if I put into truth GPT and say uh is

16:36

Donald Trump racist and it just says no

16:38

or yes whatever it says right and then

16:40

give me an example of Donald Trump being

16:42

racist and it's like here's an example

16:44

where Donald Trump you know said

16:46

something about a black person whatever

16:47

right

16:48

uh and then you're like give me an

16:50

example of where Donald Trump said

16:51

something about a white person it does

16:52

fine

16:53

in such an example in theory it would be

16:56

useful for the truth GPT to also go hey

16:59

look you asked for this but just keep in

17:02

mind here's a whole host of data and

17:04

information that suggests that yes out

17:07

of context one of those comments may

17:09

have been made but in aggregate

17:10

so-and-so is likely not racist or likely

17:14

racist because if we create a binary

17:17

world that is you're either for me or

17:19

against me then we have a disastrous

17:21

world and that's really what I'm

17:23

exemplifying here you're creating you're

17:25

creating a one side versus the other

17:27

side but humanity is is much more

17:30

complex than that and humanity is is

17:33

what I call a gradient right this is

17:35

what you have you have a gradient and so

17:37

you have people at various different

17:39

elements you have people over here on

17:41

the left you have people over here on

17:44

the right and the more you are to these

17:46

these edges the more extremes you have

17:49

the vast majority of people I would

17:51

imagine rest over here and and these

17:53

people this is probably 80 percent right

17:55

here they deserve both sides of the

17:58

argument not just one and unfortunately

18:01

that Community known example was

18:04

providing just one piece of data which

18:07

reiterates just one side of the spectrum

18:09

and it doesn't even provide the

18:12

suggestion like a button that's like hey

18:14

if you want more of our opinion on the

18:16

or like the community notes view of this

18:18

debate click here

18:20

well then again you're not actually

18:22

leading the world to realize there's

18:24

more truth to be had than just one

18:26

little correction so again by making one

18:28

correction and not others you're de

18:31

facto taking a silence a problem so that

18:33

is uh my advice to maybe uh the truth

18:36

GPT folks if uh if there should ever

18:38

become a truth GPT and that is my take

18:42

on Elon Musk and Truth GPT people like

18:45

oh Kevin's always an Elon and Tesla

18:47

shield no I I personally try to look for

18:50

truth as well and that doesn't mean I'm

18:51

perfect but I really try even if it's

18:53

unpopular to do I really try to look for

18:57

what's the middle of the road what are

19:00

both sides of the argument because

19:01

reality is I think most people eighty

19:03

percent of people

19:04

serve both sides

19:06

I mean really everybody does but some

19:08

people won't want to hear it cheers

19:11

[Music]

UNLOCK MORE

Sign up free to access premium features

INTERACTIVE VIEWER

Watch the video with synced subtitles, adjustable overlay, and full playback control.

SIGN UP FREE TO UNLOCK

AI SUMMARY

Get an instant AI-generated summary of the video content, key points, and takeaways.

SIGN UP FREE TO UNLOCK

TRANSLATE

Translate the transcript to 100+ languages with one click. Download in any format.

SIGN UP FREE TO UNLOCK

MIND MAP

Visualize the transcript as an interactive mind map. Understand structure at a glance.

SIGN UP FREE TO UNLOCK

CHAT WITH TRANSCRIPT

Ask questions about the video content. Get answers powered by AI directly from the transcript.

SIGN UP FREE TO UNLOCK

GET MORE FROM YOUR TRANSCRIPTS

Sign up for free and unlock interactive viewer, AI summaries, translations, mind maps, and more. No credit card required.