Real Coffee with Scott Adams - Episode 2304 Scott Adams： CWSA 11⧸26⧸23 Intelligence Is An Illusion, AI Proves It. So Does The News

00:00:00.000 Do, do, do, do.

00:00:02.460 Do, do, do, do, do, do, do, do.

00:00:06.080 Good morning, everybody, and welcome to the highlight of human civilization,

00:00:11.280 and possibly even the civilizations that built the pyramids, whoever they were.

00:00:17.340 If you'd like to take this experience up to levels that even Elon Musk can't reach with his best rocket,

00:00:24.220 all you need is a cup or a mug or a glass, a tank or a chalice or a stein,

00:00:27.380 and a canteen jug or a flask, a vessel of any kind.

00:00:30.920 Fill it with your favorite liquid.

00:00:32.340 I like coffee.

00:00:33.840 And join me now for the unparalleled pleasure of the dopamine,

00:00:37.720 the day, the thing that makes everything better, a little bit of serotonin today.

00:00:41.960 It's called the Simultaneous Sip, and it happens now.

00:00:45.360 Go.

00:00:49.480 Oh, that's good.

00:00:51.240 Savor it.

00:00:52.220 Savor it.

00:00:53.800 All right, good.

00:00:57.380 We're going to just gulp and not savor, and that would just, what a way to start the day, huh?

00:01:03.820 All right, let's talk about all the news.

00:01:07.040 I do have a theme which will emerge pretty soon, and the theme is,

00:01:13.560 Intelligence is an Illusion.

00:01:17.580 That's right.

00:01:19.120 Intelligence is an Illusion.

00:01:21.040 We'll see if we can find that theme as I go through the stories today.

00:01:26.500 I think you'll find it.

00:01:28.280 My favorite story of the day goes like this.

00:01:32.880 I'm no political consultant, but if I were, I would say to you,

00:01:39.140 if you're running a campaign, and the biggest topic about your campaign

00:01:43.780 is the question of whether Roger Stone called your spouse a four-letter word that starts

00:01:51.360 with C and ends with T, and that's the only news you're making.

00:01:58.820 That's Ron DeSantis' situation.

00:02:01.800 Ron DeSantis, the only story out of his campaign this week,

00:02:06.120 is that Roger Stone may have called his wife a C word.

00:02:10.400 He denies it.

00:02:13.740 He says he only said, see you next Tuesday.

00:02:18.620 So I guess we can believe him, and maybe he just plans to see you next Tuesday,

00:02:24.080 if you know what I mean.

00:02:26.460 All right.

00:02:27.920 Well, that's pretty funny.

00:02:30.140 So if I can teach you one thing, it would be...

00:02:35.080 That's the time to quit your campaign.

00:02:38.640 If he's looking for a sign, can you imagine DeSantis praying for guidance?

00:02:46.880 God, I'm trying to decide whether I should stay in the president race,

00:02:51.760 or should I get out?

00:02:53.620 Can you send me a sign?

00:02:56.640 Well, let's see what's happening next today.

00:02:59.420 Roger Stone saying some things about...

00:03:02.720 Okay, it's time to quit.

00:03:04.020 It's time to quit.

00:03:04.960 I'm out.

00:03:05.920 I'm out.

00:03:06.440 So, that would be the religious interpretation.

00:03:13.300 All right.

00:03:15.840 In the category of President Trump is right again,

00:03:20.320 I was watching the Amuse account.

00:03:23.460 That's the name of an account on X, Amuse.

00:03:25.840 And Trump has told Democrats that if you keep taking down statues,

00:03:32.160 pretty soon they're going to come for Thomas Jefferson.

00:03:35.240 Do you remember when you thought,

00:03:36.580 they're not going to come for Thomas Jefferson?

00:03:39.480 Well, the Thomas Jefferson statue has been removed from City Hall in New York City.

00:03:46.560 So, it's gone.

00:03:50.760 Now, I'm perfectly in favor of removing it,

00:03:55.200 but I think there's going to be a big sort of gap where there should be some artwork.

00:04:00.720 Is anybody with me?

00:04:01.920 Should it be George Floyd that they put back?

00:04:05.200 I mean, that makes sense, right?

00:04:06.220 Because, as we've learned, Thomas Jefferson was a horrible, racist piece of shit,

00:04:12.200 but not George Floyd.

00:04:15.180 George Floyd would be more of a hero situation.

00:04:19.540 So, everybody, George Floyd statue?

00:04:24.380 It's unanimous.

00:04:25.940 That's what we'd like to see.

00:04:29.900 Meanwhile, Business Insider is writing an article,

00:04:34.780 the title of which is,

00:04:36.240 Here's what happens if Trump dies while running for office.

00:04:40.780 And then they repeat,

00:04:41.880 dies while running for office over and over again in the article.

00:04:46.580 Do you know what that looks like?

00:04:49.000 Well, I don't know,

00:04:50.480 but it looks like a murder attempt to me.

00:04:53.660 Because if this reporting,

00:04:59.040 and this is just speculative,

00:05:01.380 but if this were influenced by any intelligence entities within the United States,

00:05:07.300 it would be for the obvious purpose of triggering the actual assassination.

00:05:13.880 So, this is the copycat problem,

00:05:17.480 except without the copying.

00:05:19.420 So, the copycat problem with serial killers

00:05:21.920 is that I don't believe anybody would even have the idea

00:05:25.160 of shooting up a school,

00:05:27.680 except that it's in the news.

00:05:30.000 Like, who would even have that idea?

00:05:33.060 Right?

00:05:33.520 So, these are clearly things that the news prompts.

00:05:36.760 So, knowing that the news can prime you

00:05:39.480 and prompt you into doing something you wouldn't have even thought of doing,

00:05:42.400 what happens if you start running a whole bunch of stories

00:05:46.360 about what would happen if somebody took a shot at Trump?

00:05:51.620 It's guaranteed.

00:05:53.680 If you tell that often enough to a large enough population,

00:05:58.120 it's not a maybe.

00:06:00.320 Does everybody get that?

00:06:01.300 If the population is large enough,

00:06:04.540 and you repeat that message enough,

00:06:06.200 even though you're not encouraging somebody to do it directly,

00:06:09.380 just putting the idea into that many people's heads

00:06:11.860 largely guarantees it happens,

00:06:14.860 or at least somebody tries or thinks about it really hard.

00:06:18.160 So, that's pretty close to a murder attempt.

00:06:20.680 Now, if this was just an idea that the reporter or the journalist had,

00:06:26.260 well, then it's not,

00:06:28.040 unless the journalist was intending to make it a murder attempt.

00:06:32.220 But there's no way to prove that.

00:06:36.020 To me, it looks like not an organic story.

00:06:39.880 No way to know.

00:06:40.960 But to me, it looks like a story that somebody in the blob,

00:06:46.220 you know, we call the blob whoever's running the country

00:06:49.660 from the intelligence and media and Democrat side.

00:06:53.120 It looks like they, to me, it doesn't look organic.

00:06:56.940 But it's hard to know.

00:06:58.420 Just a guess.

00:06:59.960 Well, there's this hostage release situation happening

00:07:02.800 over in Israel and Gaza.

00:07:06.280 And there's a reason to believe,

00:07:08.240 we're told by Jake Sullivan, National Security Advisor,

00:07:10.900 that at least one of the three Americans

00:07:14.440 might be part of the ones released.

00:07:16.180 I kind of hate that Hamas has us playing this game 0.77

00:07:21.940 where, you know, you have to think about and worry

00:07:25.840 and speculate about which hostages get released.

00:07:29.340 I feel to some extent that's sort of playing

00:07:31.700 into their terrorism model,

00:07:35.120 that it makes you think about it and get terrorized again.

00:07:38.020 Because if you're hoping your loved ones get released,

00:07:41.180 but then they're not released,

00:07:42.420 that would be like terrorism all over again.

00:07:46.160 It's like they're milking this frickin' thing

00:07:47.760 as hard as they can.

00:07:49.540 Now, I don't know,

00:07:50.720 I'm not saying there's an alternative.

00:07:54.880 Honestly, if I were in charge of everything

00:07:57.640 that Israel is doing so far,

00:07:59.780 I would do it exactly the same.

00:08:03.120 I have not personally,

00:08:05.480 you know, I'm no military expert,

00:08:07.580 no expert in the area,

00:08:08.600 but when I watched from the beginning to the end,

00:08:11.280 from October 7th on,

00:08:13.200 I didn't see anything I would disagree with

00:08:15.500 in the way that they would handle it.

00:08:17.880 And what I mean by that

00:08:18.820 is it's the way any country would handle it

00:08:20.720 if they had the resources.

00:08:22.880 So the reason that I don't criticize them

00:08:25.400 is that literally anybody

00:08:27.040 would do the same thing they're doing.

00:08:29.000 How do you criticize that?

00:08:31.760 You know, you can make some kind of, like,

00:08:34.220 hypothetical, philosophical,

00:08:35.760 you know, morality-based argument,

00:08:39.220 but the truth is 100% of nations

00:08:41.980 which exist today,

00:08:43.480 if they had the resources

00:08:45.000 and they were in that situation,

00:08:47.400 would respond the same way.

00:08:49.280 So criticizing it just seems like

00:08:51.160 an absurdity to me.

00:08:52.980 It's more like cause and effect

00:08:54.240 and just watching it, basically.

00:08:59.580 So anyway, that's moving forward.

00:09:02.360 CNN reports,

00:09:03.360 and I'm going to give CNN a little credit here.

00:09:06.540 When there are stories,

00:09:08.380 international stories,

00:09:09.940 that don't have a specific

00:09:11.820 political component to them,

00:09:15.180 they seem pretty good.

00:09:17.540 There's a lot of stuff that CNN gets right,

00:09:19.520 just not if it involves Republicans.

00:09:21.820 As long as there's no Republican

00:09:23.520 in the story whatsoever,

00:09:25.740 they actually do a decent job

00:09:27.680 of collecting news.

00:09:29.500 So one of the stories here is that,

00:09:30.900 and I didn't know this,

00:09:33.380 so this is brand new news to me,

00:09:34.840 that a lot of the workers

00:09:36.860 on the farms in Israel

00:09:38.420 were, of course, Palestinians,

00:09:41.160 but there weren't nearly as many

00:09:42.560 as there used to be

00:09:43.400 because of past history.

00:09:47.480 Apparently, most of the workers,

00:09:49.180 or a lot of them,

00:09:49.960 a ton of them were Thai,

00:09:51.680 so they're from poor areas in Thailand,

00:09:54.100 and apparently 30,000 to 40,000 workers

00:09:58.080 are now, quote, missing.

00:10:01.900 They're missing,

00:10:03.180 meaning that they didn't go to work.

00:10:05.920 But apparently the Thais were massacred

00:10:08.120 on October 7th.

00:10:10.220 So Hamas didn't care 0.99

00:10:11.360 what your nationality was,

00:10:12.640 they were just killing people,

00:10:14.540 so they killed a bunch of Thai workers,

00:10:17.540 and the Thai workers said,

00:10:19.720 no, thank you.

00:10:21.200 I'm sure there's another world,

00:10:23.320 there's someplace else we could work.

00:10:25.440 And so they left.

00:10:27.060 So they're not missing, missing,

00:10:28.820 they probably just went home.

00:10:31.420 And so it looks like the crops,

00:10:33.240 there's not enough workers

00:10:34.200 to pick the crops,

00:10:35.520 and, you know,

00:10:36.840 it takes some skill to milk a cow.

00:10:39.220 They said you had to be

00:10:40.260 highly skilled to milk a cow.

00:10:42.760 Has anybody ever milked a cow?

00:10:44.200 I mean, I realize

00:10:47.280 that they're using milking machines,

00:10:50.440 but may I teach you

00:10:53.940 everything you need to know

00:10:54.860 about milking a cow?

00:10:58.580 So you bring the cow in,

00:11:00.880 you know, you put it in its little stall

00:11:02.360 so its head is immobilized.

00:11:04.880 They like to be milked, apparently,

00:11:06.720 because they're just used to it.

00:11:09.220 Then you take one of the teats,

00:11:11.440 yes, that's what they're called,

00:11:12.380 the teats, one of the four.

00:11:14.960 And you have to,

00:11:15.820 I'm not going to do the impression

00:11:17.020 of it on live stream

00:11:18.060 because it'll turn into a meme,

00:11:19.440 but imagine somebody

00:11:20.800 shaking a banana.

00:11:22.800 You know, you do that a little bit,

00:11:24.780 and it's called priming.

00:11:26.740 So you have to prime

00:11:28.400 each of the teats by hand

00:11:30.360 to make sure that it's producing.

00:11:32.380 And then you take

00:11:33.520 the little suction thing

00:11:35.540 that will be the automatic milker,

00:11:37.680 and you replace your hand

00:11:39.140 with the milker,

00:11:39.780 and it milks that cow. 0.98

00:11:43.040 It's not really that hard.

00:11:45.480 I mean, I'm pretty sure

00:11:46.360 I learned it completely

00:11:47.280 when I was eight years old.

00:11:49.040 Took about five minutes of training.

00:11:52.040 So I don't know

00:11:53.080 what the skill exactly is.

00:11:54.560 I guess maybe maintaining

00:11:56.140 the milking machines or something.

00:11:57.700 There's probably some skill in that.

00:12:00.680 But anyway,

00:12:01.920 there's going to be

00:12:02.680 way too much milk in the cows

00:12:04.480 and way too much food

00:12:05.820 in the fields.

00:12:09.180 But separately,

00:12:10.860 we heard stories.

00:12:11.620 Joel Pollack was reporting on this.

00:12:13.920 A lot of the citizens,

00:12:16.020 the Israeli citizens,

00:12:16.960 were chipping in

00:12:17.600 trying to get the harvest picked,

00:12:19.660 but there won't be enough of them,

00:12:20.940 I don't think.

00:12:22.820 All right.

00:12:23.200 In another story,

00:12:23.980 Charlie Kirk is talking about

00:12:25.220 how there's a new memo

00:12:26.660 for the border,

00:12:27.800 the American border.

00:12:28.800 Joe Biden's DHS patrol

00:12:32.120 have now been trained

00:12:33.820 that they must use

00:12:34.840 gender-neutral language

00:12:36.160 with the immigrants coming in.

00:12:38.160 So they can no longer

00:12:39.520 just assume he, him, she, her,

00:12:41.760 Mr. and Mrs.

00:12:42.960 until they're certain

00:12:43.960 the migrant goes 1.00

00:12:45.120 by one of those pronouns.

00:12:47.140 So you can't call the migrants

00:12:48.380 by a regular pronoun

00:12:49.620 until you're really sure

00:12:50.840 what's going on.

00:12:52.900 It's called

00:12:53.300 The Guide to Facilitating

00:12:54.500 Effective Communications

00:12:55.880 with Individuals

00:12:56.900 who Identify as LGBTQ.

00:12:58.800 LGBTQI+.

00:12:59.800 I?

00:13:01.060 What's the I?

00:13:05.260 LGBTQI?

00:13:07.240 No, not idiot.

00:13:09.680 Intersex? 0.75

00:13:11.040 Irrational?

00:13:11.700 No, it's not irrational.

00:13:13.140 Stop being that way.

00:13:15.220 Insane?

00:13:15.780 No, stop it.

00:13:16.820 Stop it.

00:13:18.580 Incels?

00:13:19.100 No.

00:13:19.760 Well, it could be.

00:13:20.660 Is it incels?

00:13:22.740 Incest? 0.98

00:13:23.340 No, but that's a good guess.

00:13:26.540 Illegal?

00:13:27.140 No.

00:13:28.060 Stop it.

00:13:28.620 Indigenous?

00:13:29.300 No.

00:13:30.560 These are terrible guesses.

00:13:32.020 You're terrible.

00:13:32.700 Icky?

00:13:33.140 No.

00:13:33.620 You're horrible people.

00:13:35.560 Indifferent?

00:13:36.660 Invader?

00:13:37.240 By God, you're horrible people.

00:13:38.520 You're the worst people

00:13:39.280 I've ever seen in my life.

00:13:41.220 I've never seen

00:13:41.860 such worse people

00:13:42.920 in one place.

00:13:44.680 Iranian? 0.53

00:13:45.240 No.

00:13:45.620 Ignorant?

00:13:46.040 No.

00:13:47.140 Irrelevant?

00:13:47.740 No.

00:13:48.820 No.

00:13:49.340 Imperialist?

00:13:50.000 No.

00:13:50.540 Itchy?

00:13:50.860 No.

00:13:51.220 No.

00:13:51.820 No.

00:13:52.120 These are impotent?

00:13:53.000 No.

00:13:53.300 These are terrible guesses.

00:13:55.220 No.

00:13:55.560 Not imaginary.

00:13:57.820 Not igloo?

00:13:59.420 No.

00:14:00.660 No, you idiots.

00:14:02.280 It's not idiots either.

00:14:04.840 Insect?

00:14:05.660 That's terrible.

00:14:06.400 You're all awful.

00:14:07.660 You're monsters.

00:14:09.060 You're monsters.

00:14:11.500 I'm disgusted by you all.

00:14:13.440 Iguana is the correct answer. 1.00

00:14:15.480 It's iguana.

00:14:16.020 I think it's intersex. 0.98

00:14:20.340 What is intersex?

00:14:22.280 It's not inbred.

00:14:24.500 It's not inbred.

00:14:29.360 Until inflatable.

00:14:31.480 Inflatable.

00:14:35.040 It's not ignoramus.

00:14:36.820 No.

00:14:37.460 It's not impotent.

00:14:39.040 Stop guessing.

00:14:40.500 It's not impotent.

00:14:42.880 It's not iffy,

00:14:43.840 and it's not illusion.

00:14:44.640 It's not infected.

00:14:51.080 It's not infected.

00:14:53.180 It's not.

00:14:53.880 You're monsters.

00:14:55.120 Every one of you.

00:14:56.040 You're just assholes.

00:14:57.360 You're all terrible.

00:14:58.920 Inoculated.

00:15:03.780 Inoculated.

00:15:07.280 Stop it.

00:15:08.640 It's not irrelevant.

00:15:11.240 It's not intolerable.

00:15:14.640 Oh, you're all assholes.

00:15:22.120 You're all...

00:15:22.920 Every one of you is an asshole.

00:15:27.780 It's not funny.

00:15:29.960 If there's one thing I can tell you,

00:15:31.620 it's not funny.

00:15:32.300 Now, this is where I have to stop

00:15:38.460 and explain humor

00:15:40.720 to anybody who's humorless.

00:15:43.240 Is there anybody here humorless?

00:15:46.440 It's funny

00:15:47.420 because it's inappropriate.

00:15:51.560 Impenetrable.

00:15:53.280 Insecure.

00:15:53.800 Stop it.

00:15:55.580 Will you just stop it?

00:15:57.320 These are too funny

00:15:58.240 for me not to read them,

00:15:59.780 but it makes me sound

00:16:01.020 like a terrible person.

00:16:03.100 I'm not terrible.

00:16:09.140 They really shouldn't put I

00:16:11.800 on the end of that

00:16:12.500 without defining it.

00:16:16.700 I didn't realize

00:16:17.520 there were so many

00:16:18.180 insulting words

00:16:19.000 that start with I.

00:16:19.780 Oh, God,

00:16:21.380 that was unexpected.

00:16:23.020 All right,

00:16:23.360 so that's what's going on

00:16:24.320 at the border.

00:16:27.100 I didn't see

00:16:28.260 if there was any news

00:16:29.100 about massive

00:16:30.340 migrant caravans,

00:16:33.080 but we covered

00:16:34.460 the pro-down situation

00:16:35.580 pretty thoroughly.

00:16:38.480 Oh, my God.

00:16:40.280 Anyway,

00:16:40.940 speaking of

00:16:42.020 ridiculous things,

00:16:44.600 Glenn Greenwald

00:16:45.660 continues to

00:16:47.060 report

00:16:48.540 in Post on X

00:16:49.540 about what he calls

00:16:51.060 the continuing

00:16:52.300 now a mountain's

00:16:54.640 worth of evidence,

00:16:55.360 he says,

00:16:56.040 that Russia and Ukraine

00:16:57.380 were close to a deal

00:16:58.620 at the start of the war

00:16:59.600 to end it

00:17:00.100 in exchange

00:17:00.660 for Ukraine's neutrality,

00:17:02.820 not entering NATO,

00:17:04.580 but allegedly

00:17:06.200 and reportedly

00:17:07.120 the story

00:17:08.440 that's developing

00:17:09.900 suggests that

00:17:11.380 Biden and Boris Johnson

00:17:12.600 blocked it,

00:17:13.460 insisting that Zelensky

00:17:14.540 go to war 0.60

00:17:15.120 and win.

00:17:15.540 Do you think

00:17:17.500 that history

00:17:18.720 will decide

00:17:20.640 that's what happened?

00:17:22.820 Is that going to be

00:17:23.780 how history

00:17:24.360 covers this?

00:17:26.560 I feel like

00:17:27.560 maybe even

00:17:28.780 if it's true,

00:17:30.500 the history

00:17:31.040 would never

00:17:31.560 explain it that way.

00:17:33.960 I feel like

00:17:34.460 history would,

00:17:35.360 you know,

00:17:35.660 go big,

00:17:36.860 say,

00:17:37.180 well,

00:17:37.580 there are these

00:17:38.080 long-standing issues

00:17:39.200 and so there was

00:17:40.860 a disagreement.

00:17:41.980 I don't think

00:17:42.600 it would ever

00:17:42.880 get down to,

00:17:43.560 well,

00:17:44.320 there's this guy

00:17:44.880 named Joe Biden.

00:17:47.040 He was known

00:17:47.880 not to be mentally

00:17:48.700 competent at the time

00:17:50.080 and Boris Johnson

00:17:51.140 was his trained monkey

00:17:52.540 who would do

00:17:53.540 whatever Biden wanted.

00:17:55.280 So,

00:17:56.480 Joe Biden's

00:17:57.400 defective brain

00:17:59.040 decided to start

00:18:00.000 a world war

00:18:00.760 for no particularly

00:18:01.680 good reason

00:18:02.300 or no gain whatsoever

00:18:03.460 and Boris Johnson

00:18:05.620 who has a bird's nest

00:18:06.720 for a haircut

00:18:07.460 decided that he'd

00:18:08.700 go along with that

00:18:09.540 and now there were,

00:18:10.620 you know,

00:18:11.340 blah, blah, blah,

00:18:12.380 World War III.

00:18:14.280 I feel like

00:18:15.100 that's how history

00:18:15.760 is going to

00:18:16.080 cover that.

00:18:18.860 But,

00:18:19.520 if this is true,

00:18:22.540 that Zelensky

00:18:23.320 was sort of

00:18:23.920 talked into it,

00:18:26.100 that would,

00:18:26.860 at the very least,

00:18:28.220 that debunks

00:18:28.980 the concept

00:18:30.280 that Zelensky

00:18:31.460 was blackmailing

00:18:33.340 Biden.

00:18:35.480 Because,

00:18:36.120 didn't you think

00:18:36.620 that the real story

00:18:37.600 here is like,

00:18:38.540 why is this

00:18:39.420 happening at all?

00:18:40.440 I mean,

00:18:41.300 the only explanation

00:18:42.240 is that Biden's

00:18:43.160 being blackmailed

00:18:44.020 by Zelensky.

00:18:46.400 But it could be

00:18:47.160 the opposite.

00:18:48.340 It could actually

00:18:48.940 be that Zelensky

00:18:49.820 is being blackmailed

00:18:51.040 effectively

00:18:52.380 or bought off

00:18:53.440 by Biden.

00:18:56.600 The blackmail

00:18:57.300 might have been

00:18:57.720 in the other direction.

00:18:59.580 Now,

00:18:59.720 when I say blackmail,

00:19:00.700 I mean,

00:19:01.520 Biden could have

00:19:02.400 threatened directly

00:19:03.360 or indirectly

00:19:03.920 to remove Zelensky

00:19:05.140 from office

00:19:05.800 because you figure

00:19:06.520 we could figure

00:19:07.240 out some way

00:19:07.800 to do that.

00:19:09.140 And otherwise,

00:19:10.020 we could make him

00:19:10.680 rich beyond his

00:19:11.520 wildest dreams

00:19:12.320 if he survives

00:19:13.580 the war.

00:19:16.640 I think Biden

00:19:17.800 bribed Zelensky.

00:19:20.600 Or maybe it was

00:19:21.620 both.

00:19:22.920 Maybe they bribed

00:19:23.840 each other

00:19:24.280 and it was a tie.

00:19:26.020 I'm going to

00:19:26.500 bribe you to give me

00:19:27.400 money to fight this war.

00:19:28.940 You can't bribe me.

00:19:30.420 I bribe you.

00:19:31.820 I want you to

00:19:32.440 fight this war.

00:19:33.080 Here's some money.

00:19:34.000 No,

00:19:34.400 you can't give me

00:19:35.120 money.

00:19:35.420 I'm bribing you.

00:19:36.620 No,

00:19:36.920 I'm bribing you.

00:19:38.600 No,

00:19:38.960 I'm bribing you.

00:19:41.120 So,

00:19:41.680 I think that's

00:19:42.700 how history

00:19:43.120 will cover it.

00:19:44.760 Who's bribing who?

00:19:46.060 No,

00:19:46.360 I'm bribing you.

00:19:49.900 So,

00:19:50.620 let me ask you this.

00:19:51.660 How do you think

00:19:52.280 AI

00:19:53.420 will cover

00:19:55.300 that history?

00:19:56.920 Is AI

00:19:57.520 going to say,

00:19:58.660 well,

00:19:58.920 according to

00:19:59.480 Glenn Greenwald,

00:20:00.780 what we had here

00:20:01.840 is Biden

00:20:02.440 wanting a war 0.89

00:20:03.400 at any cost

00:20:04.200 and forcing

00:20:05.100 Zelensky into it?

00:20:06.400 Will he say that?

00:20:08.140 Or will he say,

00:20:10.440 there are

00:20:10.740 long-standing

00:20:11.480 problems

00:20:12.420 and,

00:20:13.020 you know,

00:20:13.780 NATO expansion

00:20:14.700 and Putin

00:20:15.900 wanted to

00:20:17.100 have a defensive

00:20:18.200 zone around

00:20:19.000 Russia

00:20:19.460 and geopolitical

00:20:21.480 situation.

00:20:23.220 Is it going to

00:20:24.160 look like that?

00:20:25.000 And people

00:20:25.580 have various

00:20:26.360 claims

00:20:26.940 about who's

00:20:27.820 right and who's

00:20:28.420 wrong,

00:20:28.980 but I'm an AI

00:20:29.860 so I can't tell you

00:20:30.840 who's right

00:20:31.220 and who's wrong.

00:20:31.800 I think that's

00:20:33.080 what it's

00:20:33.320 going to look

00:20:33.620 like.

00:20:36.420 Here's a

00:20:37.100 fake news

00:20:37.700 update.

00:20:38.480 So,

00:20:38.660 President Trump

00:20:39.240 went to some

00:20:40.000 football game

00:20:40.820 in,

00:20:41.220 was it

00:20:42.760 North Carolina

00:20:43.420 or South

00:20:43.860 Carolina?

00:20:44.520 One of the

00:20:44.960 Carolinas.

00:20:46.660 He shows up

00:20:47.440 and the crowd

00:20:48.160 goes,

00:20:48.820 it's South

00:20:49.320 Carolina?

00:20:50.000 South Carolina.

00:20:50.980 And the crowd

00:20:51.840 goes wild

00:20:52.680 with cheers.

00:20:54.760 Now,

00:20:55.360 if somebody

00:20:56.440 running for

00:20:56.920 president shows

00:20:57.660 up to a

00:20:58.160 gigantic stadium,

00:20:59.420 I mean,

00:20:59.900 just a massive

00:21:00.620 stadium,

00:21:01.940 and it's

00:21:02.980 ear-shattering

00:21:04.400 cheers,

00:21:05.660 how would

00:21:06.080 Newsweek cover

00:21:06.860 that story?

00:21:09.740 So,

00:21:10.260 candidate for

00:21:10.760 president shows

00:21:11.460 up in a

00:21:12.180 huge American

00:21:13.060 place,

00:21:15.280 huge cheers,

00:21:16.680 like deafening.

00:21:18.480 Here's how

00:21:19.140 Newsweek covered

00:21:19.760 it.

00:21:20.920 Trump was

00:21:21.780 greeted with

00:21:22.320 loud boos

00:21:23.000 in South

00:21:23.420 Carolina.

00:21:26.800 They actually

00:21:27.760 reported it

00:21:28.640 as boos.

00:21:29.420 There's

00:21:32.000 video and

00:21:32.620 audio.

00:21:33.580 You can

00:21:34.280 actually play

00:21:35.080 the video

00:21:35.640 and audio

00:21:36.260 of the

00:21:37.520 actual event.

00:21:38.740 Multiple

00:21:39.100 camera angles,

00:21:40.380 multiple

00:21:40.720 cameras,

00:21:42.020 multiple

00:21:42.300 times.

00:21:44.080 They're all

00:21:44.900 cheers.

00:21:46.440 Well,

00:21:47.040 let me put

00:21:47.400 that,

00:21:47.720 I'm sure

00:21:48.280 there are

00:21:48.600 boos in 0.72

00:21:49.180 the mix,

00:21:50.160 but overwhelmingly

00:21:51.160 it's cheers.

00:21:52.440 Now,

00:21:52.700 here's the

00:21:53.000 question.

00:21:55.240 Is it

00:21:56.000 really cheers?

00:21:57.240 Or did

00:21:58.020 AI get a

00:21:58.700 hold of it?

00:21:59.420 Is it

00:22:00.160 additive?

00:22:01.580 How would

00:22:01.920 you know?

00:22:03.260 Were there

00:22:03.700 times when

00:22:04.260 he was

00:22:04.620 booed and

00:22:06.540 they didn't

00:22:07.060 show you

00:22:07.380 that?

00:22:08.540 How would

00:22:09.020 you know?

00:22:11.820 Because we're

00:22:12.640 right at the

00:22:13.140 cusp of not

00:22:14.600 being able to

00:22:15.420 believe any

00:22:16.100 audio or

00:22:16.820 any video.

00:22:18.160 Aren't we

00:22:18.540 already in

00:22:20.220 that territory?

00:22:21.820 Are we just

00:22:22.620 before that,

00:22:23.620 or are we

00:22:24.180 already there?

00:22:24.720 where you

00:22:25.580 see a story

00:22:26.120 like this

00:22:26.600 and your

00:22:26.880 first reaction

00:22:27.700 should be,

00:22:28.840 I don't

00:22:29.180 know.

00:22:31.260 At this

00:22:31.920 point,

00:22:32.360 I'm very

00:22:32.900 close.

00:22:33.680 This one

00:22:34.100 convinces me

00:22:35.460 it's true.

00:22:36.880 It might

00:22:37.240 not be,

00:22:38.180 but I feel

00:22:38.940 persuaded even

00:22:39.820 if I'm wrong.

00:22:42.200 But I feel

00:22:42.780 like I'm

00:22:43.880 certainly within

00:22:44.660 the next year,

00:22:46.200 a story like

00:22:46.880 this, I'm

00:22:47.320 going to say

00:22:47.660 to myself,

00:22:48.200 you know

00:22:48.460 what?

00:22:49.380 Even if

00:22:49.840 there are

00:22:50.420 five different

00:22:51.240 videos of

00:22:53.000 the event that

00:22:53.600 show the same

00:22:54.140 thing, they

00:22:54.840 could all be

00:22:55.440 AI.

00:22:56.520 Because AI

00:22:57.140 could create

00:22:57.800 that with

00:22:59.320 just a text

00:23:00.120 description.

00:23:01.400 My understanding

00:23:02.140 is that the

00:23:02.780 new Google

00:23:03.700 AI will let

00:23:04.580 you create

00:23:05.080 images from

00:23:06.860 text.

00:23:08.100 So we're at

00:23:08.680 the point where

00:23:09.180 you could say,

00:23:10.460 show me a

00:23:12.000 video of Trump

00:23:12.960 arriving at a

00:23:13.620 big stadium

00:23:14.720 to booze.

00:23:17.360 Boom.

00:23:17.720 show me five

00:23:19.260 different scenes

00:23:20.560 or videos that

00:23:22.180 look like from

00:23:22.800 the same studio

00:23:23.660 but maybe at

00:23:24.500 different times.

00:23:25.620 Each time he's

00:23:26.720 getting cheered

00:23:27.460 or booed.

00:23:29.240 And it just

00:23:29.940 creates it.

00:23:31.080 And you're

00:23:31.440 done.

00:23:32.380 And it would

00:23:32.780 look just like

00:23:33.320 the real thing.

00:23:33.960 And they even

00:23:34.480 have the right

00:23:34.860 number of

00:23:35.240 fingers now.

00:23:37.000 So I saw

00:23:40.920 there's a

00:23:41.280 product now,

00:23:42.440 an extra

00:23:43.260 finger.

00:23:44.460 So you can

00:23:45.200 add a fake

00:23:45.860 finger to your

00:23:46.660 real hand.

00:23:47.280 So it looks

00:23:47.800 like you have

00:23:48.160 six fingers.

00:23:49.640 In case you

00:23:50.460 rob a bank

00:23:51.240 and they

00:23:52.340 catch you on

00:23:52.800 video,

00:23:53.860 they say,

00:23:54.200 we got you.

00:23:54.880 You can say,

00:23:55.660 do you have

00:23:56.320 me?

00:23:56.940 Look at the

00:23:57.480 number of

00:23:57.860 fingers.

00:23:59.300 Aha! 0.94

00:24:01.020 Obviously AI

00:24:01.980 generated.

00:24:05.160 Okay.

00:24:05.780 Well,

00:24:06.080 AI is no

00:24:06.860 longer creating

00:24:08.320 extra fingers.

00:24:09.240 They've already

00:24:09.620 fixed that.

00:24:11.440 So that

00:24:12.240 wouldn't work.

00:24:12.680 But it's a

00:24:13.140 clever idea.

00:24:15.560 Anyway,

00:24:16.720 so yeah,

00:24:17.520 so you're

00:24:18.000 gaslighted so

00:24:18.980 badly that

00:24:19.780 they turned

00:24:20.380 wild cheering

00:24:21.780 into booze.

00:24:23.820 They actually

00:24:24.500 tried that.

00:24:25.740 It's going to

00:24:26.160 work.

00:24:26.860 Because if you

00:24:27.460 only read the

00:24:28.020 headline and

00:24:29.220 you never saw

00:24:29.780 the video,

00:24:30.360 which would

00:24:30.720 be most

00:24:31.940 people,

00:24:33.280 don't you

00:24:33.660 think most

00:24:34.340 of the people

00:24:34.820 who see the

00:24:35.320 headline will

00:24:36.340 never see the

00:24:36.940 video?

00:24:37.200 Yeah,

00:24:39.120 that's pretty

00:24:39.780 good propaganda

00:24:40.700 right there.

00:24:42.700 All right,

00:24:43.280 in the case

00:24:44.720 of Elon Musk

00:24:45.660 versus Media

00:24:46.600 Matters,

00:24:48.360 the group

00:24:49.160 that convinced

00:24:50.380 the X

00:24:51.580 advertisers,

00:24:52.560 the advertisers

00:24:53.240 on X,

00:24:54.760 to stop

00:24:57.680 advertising,

00:24:58.960 then that

00:24:59.440 cost, of

00:24:59.940 course,

00:25:00.340 Musk a 1.00

00:25:00.840 great deal

00:25:01.240 of money.

00:25:02.740 And so

00:25:03.220 he's suing.

00:25:04.680 Now,

00:25:05.120 in order to

00:25:05.720 make your

00:25:06.180 case for

00:25:07.020 some kind

00:25:07.880 of defamation

00:25:08.580 or there's

00:25:09.600 another name

00:25:10.220 for it,

00:25:10.840 which is

00:25:11.500 interfering with

00:25:12.260 business in

00:25:12.920 an illegitimate

00:25:14.580 way.

00:25:15.580 What's the

00:25:15.860 name of that?

00:25:17.160 There's a

00:25:17.780 legal,

00:25:18.960 no, it's

00:25:19.280 not libel,

00:25:19.940 there's a,

00:25:20.560 it's like

00:25:20.880 business,

00:25:21.500 it has

00:25:21.680 business in

00:25:22.280 the name?

00:25:24.020 It's not a

00:25:24.820 generic name,

00:25:25.480 it's like a

00:25:25.960 specific business

00:25:27.300 malicious thing.

00:25:29.140 Business

00:25:29.500 interference.

00:25:31.640 Is it

00:25:31.960 business,

00:25:32.640 not a

00:25:33.560 restraint of

00:25:34.120 trade?

00:25:35.580 I think

00:25:35.900 it's some

00:25:36.360 kind of

00:25:36.660 business

00:25:37.000 interference

00:25:37.540 is part

00:25:38.600 of the

00:25:38.860 thing they

00:25:39.420 can charge.

00:25:40.520 Anyway,

00:25:41.280 not charge,

00:25:42.760 but that

00:25:43.020 could be

00:25:43.240 the claim.

00:25:44.600 So here's

00:25:45.560 what makes

00:25:46.060 Musk's 0.98

00:25:46.640 case

00:25:47.240 unusually

00:25:48.460 good.

00:25:49.720 Number one,

00:25:50.420 you have to

00:25:50.880 prove actual

00:25:51.960 damages.

00:25:53.560 Typically,

00:25:54.820 that's really

00:25:55.320 hard to prove.

00:25:56.160 If somebody

00:25:56.620 just says bad

00:25:57.320 stuff about

00:25:57.900 you on

00:25:58.220 social media,

00:25:59.840 well,

00:26:00.420 maybe your

00:26:01.060 social media

00:26:01.760 traffic went

00:26:02.440 down a little,

00:26:03.300 maybe it

00:26:03.680 didn't.

00:26:04.640 It'd be hard

00:26:05.160 to know if

00:26:05.540 you were

00:26:05.860 injured.

00:26:07.060 So in that

00:26:07.540 case,

00:26:07.880 it'd be really

00:26:08.280 hard to

00:26:08.820 press a

00:26:09.640 case because

00:26:10.260 they'd say,

00:26:10.740 well,

00:26:10.880 what's the

00:26:11.200 dollar amount

00:26:11.760 of this

00:26:12.120 injury?

00:26:12.980 And you

00:26:13.340 just say,

00:26:13.960 I don't

00:26:15.240 know,

00:26:15.560 I think

00:26:15.900 there is

00:26:16.220 some,

00:26:17.000 you would

00:26:17.320 lose that

00:26:17.700 case.

00:26:18.860 But in

00:26:19.760 the case

00:26:20.140 of advertisers

00:26:21.060 who were

00:26:21.720 advertising,

00:26:22.960 and then

00:26:23.380 they stopped,

00:26:24.720 you've got the

00:26:25.300 cleanest argument

00:26:26.080 you could ever

00:26:26.640 have.

00:26:27.520 We used to

00:26:28.180 make this

00:26:28.580 money,

00:26:29.920 Media Matters

00:26:30.620 did this

00:26:31.040 stuff,

00:26:32.160 and now we

00:26:32.660 don't make

00:26:33.080 that money,

00:26:33.760 and it's going

00:26:34.160 to be in the

00:26:34.700 at least

00:26:35.420 tens of

00:26:35.920 millions.

00:26:37.240 So they

00:26:38.120 can easily

00:26:38.720 prove the

00:26:39.300 damages part

00:26:40.120 because the

00:26:41.080 companies that

00:26:41.760 stopped

00:26:42.060 advertising,

00:26:43.080 they said

00:26:43.600 directly and

00:26:44.360 publicly why

00:26:45.120 they stopped.

00:26:46.820 So that's

00:26:47.700 all you need

00:26:48.060 to know.

00:26:48.760 So that

00:26:49.420 part looks

00:26:50.040 easy.

00:26:51.440 Now you

00:26:52.160 have to

00:26:52.460 prove that

00:26:53.120 they did

00:26:53.560 something

00:26:54.140 really skeevy,

00:26:56.500 you know,

00:26:56.680 really,

00:26:57.400 let's say,

00:26:58.580 dishonest,

00:26:59.320 even if it's

00:26:59.860 not illegal.

00:27:01.120 Right?

00:27:01.560 They don't

00:27:02.020 have to

00:27:02.380 violate the

00:27:02.900 law necessarily

00:27:04.020 because this

00:27:05.020 is a civil

00:27:05.560 case,

00:27:06.300 but they

00:27:07.100 do have

00:27:07.460 to do

00:27:07.700 something

00:27:08.080 so weaselly

00:27:10.760 and,

00:27:11.400 you know,

00:27:12.640 let's say

00:27:13.020 illegitimate

00:27:13.840 that there's

00:27:15.620 no doubt

00:27:16.120 that they

00:27:16.660 were trying

00:27:17.140 to do it

00:27:17.720 with the

00:27:18.540 intention

00:27:19.100 of damaging

00:27:20.620 the entity.

00:27:22.340 And it

00:27:22.660 looks like

00:27:23.180 that's going

00:27:23.620 to be easy

00:27:24.060 to prove

00:27:24.500 as well

00:27:25.020 because the

00:27:27.200 people involved

00:27:28.240 have lots

00:27:28.740 of, you

00:27:29.140 know,

00:27:29.260 body of

00:27:29.700 work saying

00:27:30.220 exactly what

00:27:30.860 they're trying

00:27:31.200 to do,

00:27:31.800 so there's

00:27:32.360 no question

00:27:32.860 that they're

00:27:33.160 trying to

00:27:33.740 take out

00:27:34.820 Elon,

00:27:35.900 and they're

00:27:36.560 trying to

00:27:36.880 take out

00:27:37.240 X.

00:27:37.940 So I think

00:27:38.700 it's easy

00:27:39.240 to demonstrate

00:27:39.900 intention.

00:27:41.800 So you've

00:27:42.200 got intention

00:27:42.820 and you've

00:27:44.100 got damage

00:27:44.900 and then

00:27:45.740 the next

00:27:46.140 thing you

00:27:46.440 have to

00:27:46.720 prove

00:27:47.140 is their

00:27:48.000 super

00:27:48.440 weaselly

00:27:49.120 deceptive

00:27:49.940 behavior

00:27:50.440 in their

00:27:52.040 attempt

00:27:52.460 to be

00:27:53.160 malicious.

00:27:54.520 And it

00:27:54.740 turns out

00:27:55.220 that because

00:27:56.240 things are

00:27:57.260 tracked and

00:27:57.980 logged

00:27:58.440 within X

00:28:00.020 that X

00:28:01.440 knows exactly,

00:28:02.560 they would

00:28:02.920 claim,

00:28:03.220 how the

00:28:04.300 claims were

00:28:05.340 made and

00:28:05.880 how malicious

00:28:07.080 and weaselly

00:28:08.520 they were to

00:28:09.880 make the

00:28:10.600 claim.

00:28:11.480 So here's

00:28:11.940 the best I

00:28:13.220 can explain

00:28:14.200 it.

00:28:14.940 If you

00:28:15.540 were a

00:28:15.820 normal user

00:28:16.640 on X,

00:28:18.460 the odds

00:28:19.140 of you

00:28:19.460 having your

00:28:20.180 content or

00:28:21.120 any advertisement

00:28:21.940 that you

00:28:23.180 saw paired

00:28:24.300 with an

00:28:24.780 ad where

00:28:25.840 you'd see

00:28:26.140 some neo-Nazi

00:28:26.920 stuff and

00:28:27.440 you'd see

00:28:27.800 some Apple

00:28:29.020 computer stuff,

00:28:29.720 the odds

00:28:30.700 for a

00:28:31.260 non-Nazi

00:28:32.160 to see

00:28:32.600 that stuff

00:28:33.180 is like

00:28:34.240 vanishingly

00:28:35.120 millions to

00:28:36.060 one.

00:28:36.540 It's like

00:28:36.980 impossible.

00:28:38.560 So that

00:28:38.900 actually,

00:28:40.060 if you look

00:28:40.520 at the

00:28:40.740 actual ability

00:28:41.600 of X 0.61

00:28:42.180 to avoid

00:28:43.560 pairing those

00:28:44.240 things,

00:28:45.280 it's sensational.

00:28:46.720 It's not

00:28:47.460 just good,

00:28:49.060 it's like

00:28:49.600 insanely good.

00:28:51.080 Like it

00:28:51.560 will really,

00:28:52.520 really do a

00:28:53.120 good job

00:28:53.760 of making

00:28:54.480 sure your

00:28:54.940 ad doesn't

00:28:55.480 show up

00:28:55.780 next to

00:28:56.140 bad content.

00:28:56.800 Like really

00:28:57.960 good.

00:28:58.720 Like better

00:28:59.100 than anything

00:28:59.640 has ever

00:29:00.060 worked in

00:29:01.060 any domain.

00:29:02.140 That's how

00:29:02.440 good it is.

00:29:03.360 Like it's

00:29:03.600 not perfect,

00:29:05.060 but it's

00:29:05.340 like a one

00:29:06.120 in millions

00:29:06.660 before you'd

00:29:07.380 see something

00:29:07.780 like that.

00:29:09.080 So how

00:29:09.560 did Media

00:29:10.680 Matters

00:29:11.280 produce

00:29:12.180 something that

00:29:13.000 would be

00:29:13.240 so rare?

00:29:14.500 Well,

00:29:14.960 first of all,

00:29:15.460 they made

00:29:15.740 sure that

00:29:16.100 they used

00:29:16.460 existing

00:29:17.040 accounts,

00:29:18.060 because if

00:29:18.520 you created

00:29:18.960 a new

00:29:19.300 account to

00:29:21.460 do some

00:29:21.820 testing,

00:29:22.500 X would

00:29:22.900 immediately

00:29:23.340 know their

00:29:23.820 new accounts

00:29:24.560 and it

00:29:25.320 would treat

00:29:25.660 them

00:29:25.820 differently.

00:29:26.800 So they

00:29:27.380 first had

00:29:27.820 to get

00:29:28.080 existing

00:29:28.560 accounts.

00:29:30.140 And then

00:29:30.940 they had

00:29:31.460 to follow

00:29:32.580 the worst

00:29:33.880 things they

00:29:34.380 could follow,

00:29:35.700 the worst

00:29:36.040 content.

00:29:36.920 So they

00:29:37.240 had to be

00:29:37.620 people who

00:29:38.020 were clicking

00:29:38.400 on bad

00:29:38.960 content.

00:29:39.920 What's the

00:29:40.380 first problem

00:29:40.920 you see?

00:29:43.160 The

00:29:43.720 algorithms

00:29:44.240 are individual,

00:29:46.120 right?

00:29:46.920 The algorithms

00:29:47.700 are not

00:29:48.140 serving everybody

00:29:49.020 the same

00:29:49.420 thing,

00:29:49.940 they're

00:29:50.200 individualizing

00:29:51.100 them.

00:29:51.720 So if you

00:29:52.400 searched for

00:29:53.060 a bunch of

00:29:53.460 Nazi content,

00:29:55.380 it might

00:29:55.900 serve you

00:29:56.460 some more

00:29:56.940 accidental

00:29:57.560 Nazi

00:29:57.960 content.

00:29:59.400 But here's

00:30:00.180 the thing,

00:30:01.180 wouldn't you

00:30:01.640 want it?

00:30:02.780 If you're

00:30:03.500 searching for

00:30:04.180 it and

00:30:04.860 you're

00:30:05.080 interacting

00:30:05.580 with it,

00:30:06.040 you probably

00:30:06.440 want it.

00:30:07.800 So if

00:30:08.200 somebody wanted

00:30:08.780 that content

00:30:09.540 and it

00:30:11.020 ended up

00:30:11.440 being paired

00:30:11.880 next to a

00:30:12.500 computer company

00:30:13.680 that they

00:30:14.100 also wanted

00:30:14.600 the content,

00:30:15.620 they wanted

00:30:16.000 the product,

00:30:18.400 they might

00:30:18.780 be more

00:30:19.120 inclined to

00:30:19.620 buy it.

00:30:20.580 But it

00:30:20.840 would be the

00:30:21.240 weirdest,

00:30:21.740 weirdest individual

00:30:22.600 case and

00:30:23.760 nothing to

00:30:24.300 do with

00:30:24.600 X in

00:30:25.200 general.

00:30:26.080 It was

00:30:26.300 just one

00:30:26.760 person pursuing

00:30:27.640 an interest

00:30:28.240 and the

00:30:29.100 algorithm helped

00:30:29.780 them and

00:30:30.940 even helped

00:30:31.380 them find a

00:30:31.820 computer to

00:30:32.360 buy.

00:30:32.580 Everybody

00:30:32.840 wins.

00:30:34.040 So they

00:30:34.660 pretended that

00:30:35.360 they were

00:30:35.640 Nazis and

00:30:37.040 they just

00:30:37.680 kept clicking

00:30:38.720 on bad

00:30:39.280 content until

00:30:40.280 the algorithm

00:30:40.860 said,

00:30:41.200 oh, I

00:30:41.380 guess you

00:30:41.760 want more

00:30:42.080 of this.

00:30:43.500 But that

00:30:44.320 wasn't enough.

00:30:46.080 They had to

00:30:46.580 continually

00:30:47.100 scroll so

00:30:48.880 that you

00:30:50.420 would have

00:30:50.660 enough

00:30:50.940 situations of

00:30:51.780 ads and

00:30:52.640 bad content

00:30:53.400 until finally

00:30:54.840 you could

00:30:56.160 get an

00:30:56.720 ad and

00:30:57.240 bad content

00:30:57.960 to line

00:30:58.440 up.

00:30:59.060 And then

00:30:59.240 you take

00:30:59.520 a screenshot

00:31:00.080 and you

00:31:01.300 sell it

00:31:02.220 to the

00:31:03.660 public like

00:31:04.680 it was

00:31:04.940 normal.

00:31:06.380 In fact,

00:31:07.100 you couldn't

00:31:07.640 produce it if

00:31:08.260 you tried.

00:31:09.440 You would

00:31:09.800 have to have

00:31:10.180 a whole

00:31:10.420 operation to

00:31:11.280 produce it

00:31:11.840 for one

00:31:12.640 user.

00:31:13.740 And for

00:31:13.980 that one

00:31:14.440 user, if

00:31:15.960 it were a

00:31:16.380 real person,

00:31:17.520 they would

00:31:17.900 be happy as

00:31:18.580 heck to

00:31:19.520 have the

00:31:19.820 advertisement

00:31:20.360 next to

00:31:20.920 that content

00:31:21.660 because it's

00:31:22.560 exactly the

00:31:23.180 product they

00:31:23.620 want to

00:31:23.840 buy.

00:31:24.760 They're in

00:31:25.140 the market

00:31:25.460 for an

00:31:25.780 Apple

00:31:25.960 computer.

00:31:28.140 So, to

00:31:29.520 the extent

00:31:30.140 that Elon

00:31:30.920 can prove

00:31:31.620 this, now

00:31:33.180 I don't know

00:31:33.800 that they

00:31:34.140 can, but I

00:31:35.080 think they

00:31:35.420 could prove

00:31:36.120 by demonstration

00:31:37.280 that you

00:31:38.640 can't reproduce

00:31:39.320 the outcome.

00:31:40.300 That part I

00:31:40.860 think they

00:31:41.160 can prove.

00:31:42.700 But if

00:31:43.380 they can

00:31:43.700 also show

00:31:44.260 the logs

00:31:45.060 of the

00:31:45.780 activity of

00:31:46.620 those media

00:31:47.180 matter counts,

00:31:48.480 they can

00:31:48.860 also show

00:31:49.620 the extreme

00:31:51.120 effort you

00:31:51.740 would have

00:31:52.140 to use

00:31:53.060 to make

00:31:53.520 it produce

00:31:53.920 a bad

00:31:54.220 outcome,

00:31:55.080 which should,

00:31:56.980 for a jury,

00:31:58.860 prove that it

00:31:59.620 was a malicious

00:32:00.240 intent and not

00:32:01.320 anything to do

00:32:02.000 with honesty or

00:32:03.040 credibility or

00:32:03.800 anything.

00:32:05.900 Now, if

00:32:07.100 Elon wins

00:32:07.680 this, he's

00:32:08.400 going to sue

00:32:08.860 them out of

00:32:09.320 existence.

00:32:10.840 But part of

00:32:12.280 the beauty is

00:32:12.960 that in

00:32:14.200 discovery, they

00:32:15.600 might find out

00:32:16.280 who's funding

00:32:16.920 them.

00:32:17.200 Now, I

00:32:18.560 don't know if

00:32:19.080 their funding

00:32:19.520 is completely

00:32:20.460 public.

00:32:21.140 We know

00:32:21.480 that Soros

00:32:22.040 is part of

00:32:22.560 it, but

00:32:23.560 wouldn't it

00:32:23.920 be interesting

00:32:24.380 for this

00:32:24.920 case to

00:32:25.540 get big

00:32:26.240 national

00:32:26.700 attention, and

00:32:28.140 the normies

00:32:28.940 who never

00:32:29.400 hear this

00:32:30.000 are going

00:32:30.400 to hear

00:32:30.620 for the

00:32:30.920 first time

00:32:31.540 that there

00:32:32.960 is a

00:32:33.240 completely

00:32:33.720 illegitimate

00:32:34.600 entity and

00:32:36.560 that Soros

00:32:37.260 is funding

00:32:37.780 them and

00:32:38.760 that he

00:32:39.080 had to

00:32:39.540 know.

00:32:40.520 That's the

00:32:41.120 key part.

00:32:42.080 It's not

00:32:42.460 that he

00:32:42.700 funded them.

00:32:44.640 I suspect

00:32:45.460 there are

00:32:45.760 lots of

00:32:46.100 situations

00:32:46.560 where good,

00:32:48.260 honest people

00:32:48.820 fund organizations.

00:32:50.720 Black Lives

00:32:51.200 Matter, for

00:32:51.620 example.

00:32:52.360 A lot of

00:32:52.720 good people

00:32:53.240 funded their

00:32:54.020 organization

00:32:54.600 because they

00:32:55.380 thought it

00:32:55.660 would do

00:32:55.900 good.

00:32:57.380 So it's

00:32:59.120 a big

00:32:59.440 difference if

00:33:00.220 the person

00:33:00.600 funding them

00:33:01.240 knows exactly

00:33:03.020 who they

00:33:03.460 are.

00:33:04.380 And there's

00:33:04.840 no way that

00:33:06.840 Soros is 0.99

00:33:08.040 unaware of

00:33:08.760 who they

00:33:09.080 are.

00:33:09.760 That's

00:33:10.000 beyond my

00:33:12.500 imagination to

00:33:13.320 imagine he's

00:33:13.940 unaware.

00:33:14.280 So this

00:33:16.700 might be a

00:33:17.260 way for the

00:33:17.740 normies to

00:33:18.760 actually learn

00:33:19.360 the news for

00:33:20.040 the first

00:33:20.420 time.

00:33:22.640 And they

00:33:22.880 might, and

00:33:24.480 this is even

00:33:24.960 better, if

00:33:26.180 it's allowed,

00:33:27.580 I don't know

00:33:28.080 if it would

00:33:28.420 be, imagine

00:33:29.820 if it's

00:33:30.240 allowed, and

00:33:31.240 if there's a

00:33:31.780 lawyer here,

00:33:32.300 can you tell

00:33:32.640 me if you

00:33:33.000 think this

00:33:33.440 would be

00:33:33.700 allowed?

00:33:34.980 Could you

00:33:35.520 use as

00:33:36.200 context for

00:33:37.280 your case

00:33:38.260 that the

00:33:39.980 Democrats

00:33:41.160 routinely set

00:33:42.280 up these

00:33:42.680 fake entities

00:33:43.480 and that

00:33:45.120 Media Matters

00:33:46.100 is not a

00:33:46.960 one-off

00:33:48.140 mistake,

00:33:49.220 something that

00:33:49.680 happened because

00:33:50.280 some rogues

00:33:51.020 work there,

00:33:52.020 but rather

00:33:52.640 it's part of

00:33:53.520 a well-understood

00:33:55.540 pattern of

00:33:57.020 creating these

00:33:57.680 fake fact-checkers

00:33:58.880 and fake

00:33:59.380 watchdogs and

00:34:00.360 the ADL,

00:34:01.580 et cetera,

00:34:02.200 and that their

00:34:02.900 purpose is to

00:34:04.540 restrict his

00:34:05.140 business.

00:34:06.320 Is it their

00:34:06.800 purpose?

00:34:07.220 What if he

00:34:09.760 proves that?

00:34:13.800 I mean,

00:34:14.300 yeah, he's

00:34:15.260 not taking a

00:34:16.180 RICO case,

00:34:17.060 but, I

00:34:18.460 mean, it's

00:34:18.900 going to

00:34:19.080 sound like

00:34:19.540 RICO.

00:34:21.320 So, I

00:34:22.860 don't know

00:34:23.360 what's going

00:34:24.900 to happen

00:34:25.260 here.

00:34:26.460 It's hard to

00:34:27.320 really predict

00:34:28.380 a legal case,

00:34:29.160 that's not

00:34:29.520 my domain,

00:34:30.620 but the

00:34:31.840 legal experts

00:34:32.580 do seem to

00:34:33.360 be agreed

00:34:33.960 that this

00:34:35.440 is not a

00:34:36.060 meritless

00:34:36.560 case and

00:34:37.180 will probably

00:34:37.860 get to

00:34:38.320 trial.

00:34:40.420 So,

00:34:41.380 2024

00:34:41.980 is looking

00:34:44.240 really

00:34:44.780 interesting.

00:34:46.680 Imagine,

00:34:47.360 if you

00:34:47.560 will,

00:34:48.860 that Elon

00:34:49.620 dismantles

00:34:51.000 media matters

00:34:53.380 and also

00:34:54.940 smears

00:34:55.540 completely

00:34:56.120 the ADL

00:34:57.140 and other

00:34:57.680 groups that

00:34:58.240 are in the

00:34:58.720 same domain.

00:35:00.320 That would

00:35:00.920 be amazing.

00:35:02.180 That would

00:35:02.400 be one of

00:35:02.720 the best

00:35:03.000 things that

00:35:03.380 ever happened.

00:35:05.060 It could

00:35:05.380 happen next

00:35:05.840 year.

00:35:06.060 At the

00:35:07.060 same time,

00:35:08.040 if we

00:35:08.440 assume a

00:35:09.020 Republican

00:35:09.480 gets into

00:35:10.140 office because

00:35:10.780 Biden's

00:35:11.340 failing quickly,

00:35:13.480 then you

00:35:14.280 should assume

00:35:14.900 a host of

00:35:16.260 other problems

00:35:16.960 will get

00:35:17.560 solved almost

00:35:18.240 immediately.

00:35:19.220 The border

00:35:19.820 will be solved

00:35:20.540 almost immediately.

00:35:22.120 Probably

00:35:22.660 something will

00:35:23.240 be done

00:35:23.620 about crime

00:35:24.220 in the

00:35:24.500 cities fairly

00:35:25.280 quickly.

00:35:26.400 Something about

00:35:27.360 Ukraine and

00:35:28.960 maybe even

00:35:29.540 maybe the

00:35:30.980 Middle East

00:35:31.420 might look

00:35:32.540 better.

00:35:34.380 So if

00:35:35.620 you're going

00:35:35.920 to be

00:35:36.420 an optimist,

00:35:38.760 you have

00:35:40.400 lots of

00:35:41.000 stuff to

00:35:41.400 look for.

00:35:43.180 There's a

00:35:43.800 whole bunch

00:35:44.280 of stuff that

00:35:44.840 could turn

00:35:45.280 out to be

00:35:45.680 really good

00:35:46.200 or not.

00:35:49.680 So remember

00:35:51.340 I told you

00:35:51.860 that intelligence

00:35:52.620 is an

00:35:54.800 illusion?

00:35:55.140 All right,

00:35:57.780 I'm going

00:35:57.980 to prove

00:35:58.300 it.

00:35:59.360 How many

00:35:59.640 of you

00:36:00.000 have heard

00:36:00.420 of the

00:36:00.800 Dunning-Kruger

00:36:01.580 effect?

00:36:04.540 Pretty

00:36:05.020 common.

00:36:05.920 Most people

00:36:06.420 who are on

00:36:07.420 the internet

00:36:07.800 have heard

00:36:08.140 of it,

00:36:08.400 right?

00:36:08.920 Now the

00:36:09.340 Dunning-Kruger

00:36:09.880 effect,

00:36:10.900 which has

00:36:11.340 been backed

00:36:11.880 by many,

00:36:12.840 many studies.

00:36:14.140 So the

00:36:14.480 first thing

00:36:14.860 you need

00:36:15.160 to know

00:36:15.580 is that

00:36:16.300 there are

00:36:16.520 many

00:36:16.820 scientific

00:36:17.320 studies,

00:36:18.020 peer-reviewed,

00:36:18.960 that

00:36:19.560 substantiate

00:36:20.680 its existence.

00:36:22.060 And what

00:36:22.320 it is,

00:36:23.300 is it

00:36:23.680 shows that

00:36:24.160 the people

00:36:24.580 who are

00:36:24.860 the dumbest

00:36:25.560 somehow

00:36:26.620 think they're

00:36:27.240 the smartest.

00:36:28.700 So that

00:36:29.300 being dumb

00:36:29.920 makes you

00:36:30.300 actually think

00:36:30.900 you're smarter

00:36:31.440 than the

00:36:31.760 people around

00:36:32.240 you.

00:36:33.220 Now that

00:36:33.680 also matches

00:36:34.900 your experience,

00:36:36.040 right?

00:36:37.340 Don't you

00:36:37.980 feel like

00:36:38.340 you've had

00:36:38.620 experience with

00:36:39.340 that?

00:36:39.560 And you're

00:36:39.680 like,

00:36:39.900 I think

00:36:40.420 that's

00:36:40.700 true.

00:36:41.500 That does

00:36:42.060 match my

00:36:42.580 experience.

00:36:44.020 So it's

00:36:44.360 pretty

00:36:44.540 believable.

00:36:45.260 So can

00:36:45.620 we all

00:36:45.980 agree that

00:36:47.540 there are

00:36:47.800 plenty of

00:36:48.280 scientific

00:36:48.680 studies,

00:36:49.440 they've

00:36:49.580 been repeated,

00:36:50.500 it's

00:36:50.660 peer-reviewed,

00:36:51.820 science is

00:36:52.440 the best

00:36:52.740 way to

00:36:53.040 understand

00:36:53.480 anything.

00:36:53.920 So can

00:36:55.020 we start

00:36:56.000 as a

00:36:56.300 base that

00:36:57.800 Dunning-Kruger

00:36:58.380 is true so

00:37:00.220 I can get to

00:37:00.720 my next

00:37:01.080 point?

00:37:01.820 Everybody on

00:37:02.460 board?

00:37:03.620 The Dunning-Kruger,

00:37:04.900 we know that

00:37:05.480 exists.

00:37:08.620 Okay,

00:37:09.240 Dunning-Kruger

00:37:09.760 doesn't exist,

00:37:11.300 and the reason

00:37:11.880 is that in

00:37:12.700 every one of

00:37:13.360 those studies,

00:37:13.980 they did the

00:37:14.480 statistics wrong

00:37:15.580 in an

00:37:16.600 easily

00:37:17.040 provable

00:37:17.620 way.

00:37:21.380 Yup.

00:37:22.560 It was

00:37:23.240 never

00:37:23.620 true.

00:37:25.440 And it's

00:37:26.060 easy to

00:37:26.700 prove it

00:37:27.040 was never

00:37:27.420 true.

00:37:28.380 All you

00:37:28.660 have to

00:37:28.880 do is

00:37:29.160 do the

00:37:29.540 statistics

00:37:30.080 on the

00:37:30.600 same set

00:37:31.040 of data,

00:37:31.800 but don't

00:37:32.440 make the

00:37:32.800 mistake.

00:37:36.480 Yeah.

00:37:37.700 Any heads

00:37:38.180 exploding?

00:37:40.300 So you

00:37:40.560 remember that

00:37:41.160 science?

00:37:42.960 Remember that

00:37:43.440 peer-reviewed

00:37:44.140 science?

00:37:45.040 Yeah,

00:37:45.300 that's good

00:37:45.660 stuff,

00:37:46.100 huh?

00:37:47.360 Yup.

00:37:49.020 And so

00:37:50.320 one of the

00:37:50.780 most basic

00:37:51.520 things about

00:37:52.140 science,

00:37:53.740 one of the

00:37:54.140 most basic

00:37:54.800 things,

00:37:56.140 was never

00:37:57.440 true.

00:37:59.100 And do you

00:37:59.820 know all

00:38:00.120 those scientists

00:38:00.820 that you

00:38:01.720 think must be

00:38:02.380 good with

00:38:02.780 statistics?

00:38:03.720 I mean,

00:38:04.300 if there's

00:38:04.760 one thing

00:38:05.240 you can trust

00:38:05.800 the scientists

00:38:06.480 to do,

00:38:07.500 it's at least

00:38:08.020 do the math

00:38:08.760 right.

00:38:09.840 Right?

00:38:10.320 I mean,

00:38:10.680 maybe the

00:38:11.060 data's wrong

00:38:11.700 in some

00:38:12.100 cases,

00:38:12.780 maybe there's

00:38:13.280 some bias

00:38:13.820 in some

00:38:14.260 cases,

00:38:14.960 but at

00:38:15.580 least you

00:38:16.720 can trust

00:38:17.180 them to

00:38:17.480 do the

00:38:17.840 statistics

00:38:18.440 correctly.

00:38:19.560 I mean,

00:38:19.740 that would

00:38:19.960 just be

00:38:20.400 baseline,

00:38:21.040 right?

00:38:22.260 Nope.

00:38:23.500 Nope.

00:38:24.580 Every one

00:38:25.180 of those,

00:38:25.680 probably fake.

00:38:26.600 Now,

00:38:26.940 it's possible

00:38:27.420 that the

00:38:27.960 story I'm

00:38:28.480 reading about

00:38:29.020 them being

00:38:29.400 fake is

00:38:29.880 the fake

00:38:30.240 news,

00:38:31.000 but that's

00:38:31.580 almost the

00:38:32.060 same story.

00:38:33.680 Who do

00:38:33.940 you trust?

00:38:35.040 Do you

00:38:35.240 trust the

00:38:35.680 story that,

00:38:37.500 let me tell

00:38:38.040 you where I

00:38:38.340 saw this.

00:38:38.740 Blair Fix

00:38:40.500 wrote this

00:38:41.160 in some

00:38:43.760 publication

00:38:44.340 called,

00:38:48.780 I figure

00:38:49.560 where it

00:38:49.840 was,

00:38:51.380 but apparently

00:38:51.820 it's been

00:38:52.200 discovered

00:38:53.160 and it's

00:38:56.380 pretty easy

00:38:58.040 to prove

00:38:58.580 that it's a

00:38:59.160 basic statistics

00:39:00.120 problem.

00:39:00.640 does that

00:39:04.360 blow your 0.91

00:39:04.740 mind?

00:39:06.220 It was

00:39:06.520 never true.

00:39:08.480 All right.

00:39:09.560 Here's

00:39:10.000 something else

00:39:10.600 along the

00:39:13.000 same theme.

00:39:13.680 Remember,

00:39:14.060 the theme

00:39:14.520 is that

00:39:16.320 intelligence

00:39:16.960 is an

00:39:18.720 illusion.

00:39:19.660 So remember

00:39:20.060 you thought

00:39:20.480 you were so

00:39:20.960 intelligent

00:39:21.440 because you

00:39:22.080 knew about

00:39:22.380 the Dunning-Kruger

00:39:23.100 thing,

00:39:23.460 right?

00:39:24.320 How many

00:39:24.960 of you,

00:39:25.380 when I said,

00:39:25.900 do you know

00:39:26.180 what Dunning-Kruger

00:39:26.980 is,

00:39:27.380 be honest,

00:39:28.560 when I asked

00:39:29.160 you all,

00:39:29.700 do you know

00:39:29.960 what Dunning-Kruger

00:39:30.560 is,

00:39:31.120 how many

00:39:31.580 thought,

00:39:32.140 I'm so

00:39:32.680 smart,

00:39:33.540 I'm a little

00:39:33.980 smarter than

00:39:34.560 the other

00:39:34.820 people.

00:39:35.220 Watch me,

00:39:35.900 I'll say I

00:39:36.560 know it,

00:39:37.400 and I'm

00:39:37.640 going to

00:39:37.740 watch all

00:39:38.140 these other

00:39:38.520 people who

00:39:38.880 don't know

00:39:39.340 it,

00:39:39.620 and I'm

00:39:40.000 going to

00:39:40.180 feel a

00:39:40.500 little

00:39:40.600 smarter

00:39:41.000 because I've

00:39:42.040 got a

00:39:42.460 thing called

00:39:43.060 intelligence.

00:39:45.540 Right.

00:39:46.000 And the

00:39:46.320 people who

00:39:46.640 have not

00:39:46.960 heard of

00:39:47.320 Dunning-Kruger,

00:39:48.160 they have a

00:39:48.920 thing I like

00:39:49.480 to call

00:39:49.760 ignorance.

00:39:51.240 So pretty

00:39:52.480 different,

00:39:53.000 right?

00:39:53.400 The ignorant

00:39:54.100 people over

00:39:54.660 here,

00:39:55.340 the intelligent

00:39:56.060 people over

00:39:56.640 here,

00:39:57.400 it was an

00:39:57.760 illusion.

00:39:59.200 The intelligent

00:39:59.800 people were

00:40:00.440 the ones

00:40:00.760 who were

00:40:00.960 wrong.

00:40:03.420 So is

00:40:04.280 intelligence an

00:40:05.080 illusion?

00:40:06.080 Well,

00:40:06.300 in that case,

00:40:06.780 it is.

00:40:07.760 Everybody who

00:40:08.480 thinks that

00:40:08.960 Dunning-Kruger

00:40:09.520 is real,

00:40:11.220 they're having

00:40:12.560 an illusion.

00:40:13.960 Like,

00:40:14.300 literally,

00:40:15.520 they're living

00:40:16.300 in a world

00:40:16.720 that doesn't

00:40:17.180 exist.

00:40:18.180 The world in

00:40:18.800 which Dunning-Kruger

00:40:19.580 is true and

00:40:20.540 proven doesn't

00:40:22.380 exist.

00:40:23.400 Now,

00:40:23.700 the other

00:40:23.960 possibility is

00:40:24.760 that the

00:40:26.820 study I'm

00:40:27.420 talking about

00:40:28.180 where it

00:40:29.140 criticized the

00:40:30.180 statistics,

00:40:31.420 maybe that's

00:40:31.960 wrong.

00:40:33.240 Maybe that's

00:40:33.760 the thing

00:40:34.040 that's wrong

00:40:34.440 and Dunning-Kruger

00:40:35.460 is right.

00:40:37.080 You tell me.

00:40:38.900 How would I

00:40:39.440 know?

00:40:41.640 Which one is

00:40:42.440 true?

00:40:43.160 I don't know.

00:40:44.660 I have zero

00:40:45.600 ability to know.

00:40:47.260 Because even

00:40:47.980 if I really

00:40:48.600 dug in,

00:40:49.800 do you think

00:40:50.140 I would have

00:40:50.560 caught that

00:40:51.200 statistics problem

00:40:52.980 on my own,

00:40:54.540 doing a deep

00:40:55.240 dive into the

00:40:55.980 data.

00:40:57.480 The scientists

00:40:58.120 didn't catch it.

00:40:59.000 The peer reviewers

00:40:59.760 didn't catch it.

00:41:01.720 And they're

00:41:01.980 probably,

00:41:02.620 you know,

00:41:03.660 reasonably,

00:41:04.340 a lot of them

00:41:04.780 probably experts

00:41:05.400 in statistics.

00:41:08.560 All right,

00:41:09.280 more on that.

00:41:10.060 The Hill has a

00:41:11.060 fascinating story

00:41:12.000 by Jody Schneider.

00:41:15.480 And it turns

00:41:16.880 out that there

00:41:17.400 are a lot of

00:41:17.820 what are called

00:41:18.380 zombie studies.

00:41:21.020 So by this

00:41:21.940 definition,

00:41:22.440 a zombie

00:41:23.140 scientific study

00:41:24.220 is something

00:41:25.200 that is done.

00:41:27.240 It's submitted

00:41:28.400 for peer review.

00:41:29.820 It passes

00:41:30.380 peer review.

00:41:31.380 It's published.

00:41:32.980 And then people

00:41:33.560 start citing it

00:41:34.560 for their own

00:41:35.180 papers.

00:41:36.820 But what happens

00:41:37.900 if later the

00:41:39.720 paper is withdrawn

00:41:41.040 because there are

00:41:42.100 other studies

00:41:42.680 that show it's

00:41:43.200 junk?

00:41:44.360 What happens

00:41:45.100 to that study?

00:41:46.760 Because now

00:41:47.460 it's been cited

00:41:48.200 by thousands

00:41:49.920 of studies.

00:41:50.500 and now

00:41:52.160 other people

00:41:52.700 will see the

00:41:53.280 other studies

00:41:54.000 that cite it

00:41:54.660 and they'll

00:41:55.680 just pick up

00:41:56.220 the citation

00:41:56.820 and say,

00:41:57.320 well,

00:41:58.500 probably true

00:41:59.180 because it's

00:41:59.880 been cited

00:42:00.400 so many times

00:42:01.200 so I'll cite it

00:42:01.960 too.

00:42:02.140 So the

00:42:03.180 citing

00:42:03.480 becomes

00:42:04.160 self-fulfilling

00:42:07.920 or self-reinforcing.

00:42:10.580 So citing,

00:42:11.460 citing,

00:42:11.820 citing,

00:42:12.200 citing,

00:42:12.580 citing,

00:42:12.800 ripple effect.

00:42:14.140 But the thing

00:42:15.000 that they cited

00:42:15.680 has been reversed

00:42:17.420 and apparently

00:42:18.640 there's no easy

00:42:19.860 mechanism in science

00:42:21.180 to inform all

00:42:22.720 the people

00:42:23.080 who cited it

00:42:23.840 that they need

00:42:24.960 to change

00:42:25.360 their citation

00:42:25.960 or even

00:42:26.560 their conclusions.

00:42:27.260 Now how big

00:42:28.840 a problem

00:42:29.180 is that?

00:42:29.680 You say

00:42:29.940 to yourself,

00:42:30.500 well,

00:42:31.800 that's probably

00:42:32.360 a problem

00:42:32.740 with,

00:42:33.400 I don't know,

00:42:33.940 a few studies.

00:42:36.140 You catch

00:42:37.120 the big ones,

00:42:38.200 the little ones

00:42:38.760 don't matter

00:42:39.240 that much,

00:42:40.200 but if you

00:42:40.620 catch the big

00:42:41.200 ones and

00:42:41.780 reverse it,

00:42:42.960 that's really

00:42:43.500 what science

00:42:44.040 is about.

00:42:44.840 Am I right?

00:42:45.900 Science isn't

00:42:46.600 about being

00:42:47.020 perfect.

00:42:48.140 It's about

00:42:48.560 catching your

00:42:49.200 mistakes,

00:42:50.380 refining your

00:42:51.200 data,

00:42:52.180 improving your

00:42:53.060 technique.

00:42:54.320 As you move

00:42:54.920 forward,

00:42:55.320 you get closer

00:42:56.100 and closer to

00:42:56.680 the truth,

00:42:57.120 right?

00:42:58.240 That's what

00:42:58.780 I learned.

00:43:00.540 So how many

00:43:01.160 of these

00:43:01.680 zombie

00:43:02.760 scientific

00:43:03.480 publications

00:43:04.140 that have

00:43:05.240 been retracted

00:43:06.240 are being

00:43:07.680 cited?

00:43:09.180 Since 1980,

00:43:11.840 more than

00:43:12.720 40,000.

00:43:16.880 More than

00:43:17.960 40,000

00:43:20.440 studies

00:43:22.080 are being

00:43:23.440 cited by

00:43:24.100 other studies

00:43:24.880 to back up

00:43:25.860 their truth

00:43:26.520 without

00:43:27.620 knowing

00:43:28.180 that they've

00:43:28.980 already been

00:43:29.460 debunked.

00:43:32.380 40,000.

00:43:34.020 Now,

00:43:34.420 your correct

00:43:34.860 question is

00:43:35.700 out of how

00:43:36.180 many.

00:43:37.220 I don't

00:43:37.500 know.

00:43:38.860 But

00:43:39.080 out of 40,000,

00:43:40.000 this is one of

00:43:40.820 those situations

00:43:41.420 where you

00:43:41.980 should know,

00:43:42.640 it would be

00:43:43.300 helpful to

00:43:43.840 know the

00:43:44.100 percentage,

00:43:45.080 but the

00:43:45.720 raw number

00:43:46.280 is still a

00:43:47.280 big story.

00:43:49.080 Right?

00:43:49.540 So if the

00:43:50.160 real number

00:43:50.560 is a million,

00:43:52.580 40,000 is

00:43:53.440 still a lot.

00:43:55.040 It's a lot.

00:43:55.760 So how

00:43:57.780 much can

00:43:58.160 you believe

00:43:58.680 scientific

00:43:59.440 studies?

00:44:00.700 I didn't

00:44:02.800 read this,

00:44:03.560 but I've

00:44:03.980 got a

00:44:04.260 hypothesis.

00:44:05.940 And it

00:44:06.280 goes like

00:44:06.900 this.

00:44:08.280 What do

00:44:08.620 you think

00:44:08.860 is more

00:44:09.320 likely to

00:44:10.040 become a

00:44:10.580 cited

00:44:10.860 paper?

00:44:12.020 Something that

00:44:12.820 sounds ordinary

00:44:13.520 or something

00:44:14.260 that sounds

00:44:14.820 extraordinary?

00:44:16.620 What is

00:44:17.120 more likely

00:44:17.740 to be

00:44:18.180 cited?

00:44:19.300 Ordinary or

00:44:20.260 extraordinary?

00:44:20.760 In other

00:44:21.280 words,

00:44:21.560 if the

00:44:22.040 study came

00:44:22.620 out about

00:44:23.560 the way

00:44:23.860 you think

00:44:24.240 it should

00:44:24.560 come out

00:44:25.040 versus

00:44:25.900 really

00:44:26.320 surprising,

00:44:27.640 I've got

00:44:28.480 a feeling

00:44:28.840 the

00:44:29.080 extraordinary

00:44:29.560 studies get

00:44:30.740 cited more.

00:44:32.260 Now,

00:44:33.100 let's take

00:44:33.720 an analogy.

00:44:34.720 I'm not sure

00:44:35.240 the analogy

00:44:35.700 holds,

00:44:36.240 but I'll make

00:44:36.880 it.

00:44:37.160 You can

00:44:37.440 see if it

00:44:38.500 holds.

00:44:39.300 In the

00:44:39.860 news business,

00:44:41.660 what gets

00:44:42.340 reported?

00:44:44.260 Dog bites

00:44:45.160 a man,

00:44:45.820 which happens

00:44:46.420 all the time,

00:44:47.080 where man

00:44:47.960 bites a

00:44:48.380 dog.

00:44:49.940 Well,

00:44:50.200 man bites

00:44:50.660 a dog is

00:44:51.180 the one

00:44:51.380 that gets

00:44:51.620 reported,

00:44:52.120 because that's

00:44:52.440 the unusual

00:44:52.980 one.

00:44:54.000 But more

00:44:54.480 than that,

00:44:56.260 man bites

00:44:57.280 a dog will

00:44:57.800 get reported

00:44:58.400 if no man

00:44:59.240 bit any

00:44:59.760 dog,

00:45:00.460 because fake

00:45:01.620 news is

00:45:02.820 extraordinary.

00:45:04.320 When you hear

00:45:04.800 it, you're

00:45:05.040 like, what?

00:45:05.940 There's a

00:45:06.400 man who wears

00:45:07.160 a dog suit

00:45:07.760 and runs

00:45:08.180 around and

00:45:08.560 bites dogs?

00:45:09.840 Well,

00:45:10.180 that's a

00:45:10.480 headline.

00:45:11.560 Everybody

00:45:11.920 will like

00:45:12.240 to hear about

00:45:12.620 that.

00:45:13.260 But then,

00:45:14.500 it didn't

00:45:14.900 happen.

00:45:15.980 The reason

00:45:16.540 that was

00:45:16.840 interesting

00:45:17.360 is that

00:45:18.300 it violated

00:45:18.920 so many

00:45:19.420 norms.

00:45:20.840 But violating

00:45:21.520 norms is

00:45:22.040 very rare.

00:45:23.900 The news

00:45:24.540 will think

00:45:25.040 that if

00:45:25.380 you see

00:45:25.660 something

00:45:25.940 that violates

00:45:26.580 norms,

00:45:27.020 they want

00:45:27.260 to put

00:45:27.460 it in

00:45:27.740 the news

00:45:28.060 right away,

00:45:28.600 because people

00:45:29.000 are going

00:45:29.220 to click

00:45:29.520 on that

00:45:29.840 stuff.

00:45:33.380 The cult

00:45:34.120 of Trump.

00:45:34.900 There's a

00:45:35.220 book called

00:45:35.580 The Cult

00:45:35.960 of Trump.

00:45:38.380 I'm

00:45:38.820 seeing in

00:45:39.140 the comments

00:45:39.560 over on

00:45:40.040 Locals

00:45:40.640 platform.

00:45:41.880 Anyway,

00:45:43.320 I've been

00:45:45.940 watching a

00:45:46.360 lot of

00:45:46.640 relationship

00:45:47.260 advice on

00:45:48.980 Instagram,

00:45:50.560 because once

00:45:51.020 you click

00:45:51.520 on a few,

00:45:52.100 it gives

00:45:52.620 you lots

00:45:52.940 of them.

00:45:53.820 And there's

00:45:54.380 one thing

00:45:54.760 that the

00:45:55.200 relationship

00:45:55.720 advice very

00:45:57.560 commonly relies

00:45:59.960 on.

00:46:00.360 Do you know

00:46:00.620 what it is?

00:46:01.120 What does

00:46:01.520 most relationship

00:46:02.560 advice rely

00:46:04.640 on?

00:46:07.000 Bogus

00:46:07.520 science.

00:46:10.160 We're actually

00:46:11.280 living our

00:46:13.600 social lives.

00:46:15.760 We're

00:46:16.020 deciding to

00:46:16.560 get married

00:46:17.040 and have

00:46:17.400 kids and

00:46:18.000 how to run

00:46:18.680 your relationship

00:46:19.300 based on

00:46:20.380 studies,

00:46:21.160 of which

00:46:23.300 you have no

00:46:24.520 idea how

00:46:25.100 true those

00:46:25.560 studies are.

00:46:26.380 No idea.

00:46:28.380 So my

00:46:29.680 observation is

00:46:31.360 that the

00:46:31.760 relationship

00:46:32.360 advice on

00:46:33.520 Instagram will

00:46:35.000 get you a

00:46:35.700 terrible life.

00:46:37.320 I mean,

00:46:37.560 the advice

00:46:38.200 is terrible.

00:46:40.600 I mean,

00:46:41.060 I've never

00:46:41.660 seen worse

00:46:42.440 advice in

00:46:43.200 any domain

00:46:44.080 than all

00:46:45.500 of the

00:46:45.940 relationship

00:46:46.940 experts.

00:46:48.140 They are

00:46:49.120 terrible.

00:46:51.380 Yeah,

00:46:51.620 don't follow

00:46:52.080 relationship

00:46:52.620 experts.

00:46:55.760 Let me,

00:46:57.720 the relationship

00:46:58.600 experts say

00:46:59.380 stuff like

00:46:59.860 this.

00:47:00.980 Like,

00:47:01.320 there's

00:47:01.500 somebody who

00:47:01.940 found out

00:47:02.420 that the

00:47:03.400 best way to

00:47:03.980 predict

00:47:04.280 divorce is

00:47:05.480 contempt.

00:47:07.560 So that if

00:47:08.220 people treat

00:47:09.260 each other

00:47:09.680 with contemptuous

00:47:11.160 words,

00:47:11.940 that predicts

00:47:13.380 a divorce.

00:47:14.840 Do you

00:47:15.100 know what

00:47:15.340 else predicts

00:47:16.500 a divorce?

00:47:17.880 I'm not

00:47:18.500 even going

00:47:18.800 to do a

00:47:19.120 study.

00:47:21.340 One or

00:47:22.060 both of

00:47:22.540 the people

00:47:22.940 in the

00:47:23.200 marriage

00:47:23.520 being a

00:47:23.980 complete

00:47:24.420 asshole, 0.97

00:47:26.140 that predicts

00:47:27.580 divorce.

00:47:28.620 I didn't

00:47:28.920 have to do

00:47:29.300 a study.

00:47:30.420 But do you

00:47:30.820 know how to

00:47:31.180 make that

00:47:31.600 obvious observation?

00:47:34.320 Do you

00:47:34.480 know how to

00:47:34.800 turn it into

00:47:35.320 science?

00:47:35.820 You call

00:47:37.460 it

00:47:37.620 contempt.

00:47:39.100 And then you

00:47:39.820 measure the

00:47:40.420 instances of

00:47:41.260 it.

00:47:42.220 And then you

00:47:42.680 turn it into

00:47:43.140 science.

00:47:45.620 Who the

00:47:46.200 fuck

00:47:46.900 treats their

00:47:48.740 own partner

00:47:49.680 with contempt?

00:47:51.860 Assholes.

00:47:53.340 You would have

00:47:54.240 to be the

00:47:54.660 biggest asshole 0.74

00:47:55.460 in the world

00:47:56.060 to treat the

00:47:57.220 person you've

00:47:57.760 dedicated your

00:47:58.460 life to

00:47:59.000 with contempt?

00:48:01.920 Right?

00:48:02.780 So of

00:48:03.380 course it's a

00:48:04.880 sign you're

00:48:05.440 going to

00:48:05.620 get divorced

00:48:06.300 because one

00:48:07.260 of you is

00:48:07.840 a gigantic

00:48:08.760 asshole or

00:48:10.160 the person

00:48:11.860 that you have

00:48:12.260 contempt for

00:48:13.060 has earned

00:48:14.220 it.

00:48:15.100 What if

00:48:15.660 they've earned

00:48:16.180 the contempt?

00:48:17.160 Because they're

00:48:17.880 a giant

00:48:18.340 asshole.

00:48:21.020 How about

00:48:21.640 all relationship

00:48:22.600 advice comes

00:48:23.400 down to this?

00:48:25.120 Have you

00:48:25.560 heard?

00:48:25.880 This is just

00:48:26.500 so bad

00:48:27.180 advice.

00:48:28.220 The experts

00:48:28.680 say the

00:48:29.300 most important

00:48:29.860 thing you need

00:48:30.320 to get right

00:48:30.900 is who you

00:48:31.980 marry.

00:48:33.200 Does anybody

00:48:33.640 disagree with

00:48:34.340 that?

00:48:35.060 That's pretty

00:48:35.460 solid advice

00:48:36.180 wouldn't you

00:48:36.500 think?

00:48:37.600 Yeah.

00:48:38.720 Do you

00:48:39.000 think anybody

00:48:39.460 knows how

00:48:39.920 to do it

00:48:40.280 right?

00:48:41.360 And do you

00:48:41.900 think if

00:48:42.240 everybody took

00:48:42.920 that advice

00:48:43.580 that there

00:48:44.080 would be

00:48:44.340 enough good

00:48:44.840 people to

00:48:45.320 marry?

00:48:47.960 That would

00:48:48.600 literally be

00:48:49.260 the end of

00:48:49.880 civilization.

00:48:51.260 If you

00:48:51.900 waited for

00:48:52.440 somebody who

00:48:53.000 had just the

00:48:53.540 right qualities,

00:48:54.560 you know,

00:48:54.760 that character

00:48:55.340 that you know

00:48:55.940 will last

00:48:56.360 forever and

00:48:56.960 excite you and

00:48:58.160 you'll always

00:48:58.720 be, you'll

00:48:59.500 always have that

00:49:00.280 sexy feeling

00:49:01.120 for, good

00:49:03.040 luck, we

00:49:04.740 would all

00:49:05.100 just give

00:49:05.720 up.

00:49:07.420 You end

00:49:08.160 up finding

00:49:09.500 somebody whose

00:49:10.200 flaws don't

00:49:11.100 bother you

00:49:11.600 that much.

00:49:13.960 Now, you

00:49:14.360 might be

00:49:14.700 legitimately

00:49:15.240 in love and

00:49:16.500 they might

00:49:16.800 have, you

00:49:17.360 know, great

00:49:17.760 qualities and

00:49:18.400 stuff, but

00:49:19.280 basically your

00:49:21.040 good marriages

00:49:21.900 are where people

00:49:22.740 can live with

00:49:23.500 the other's

00:49:23.920 flaws.

00:49:25.200 You know, like

00:49:25.700 somebody who

00:49:26.240 snores but the

00:49:27.040 other partner is

00:49:27.640 deaf.

00:49:28.580 You got lucky.

00:49:29.680 You got lucky.

00:49:30.340 So it's just

00:49:32.200 weird combinations

00:49:33.080 of people whose

00:49:34.100 flaws don't

00:49:34.900 bother the

00:49:35.380 other one.

00:49:36.480 Let's say one

00:49:37.420 of you is

00:49:38.260 like an

00:49:39.040 exercise

00:49:39.520 addict.

00:49:41.860 Well, if you

00:49:42.620 marry another

00:49:43.180 exercise addict,

00:49:44.200 it might be the

00:49:44.700 perfect situation.

00:49:46.100 So it's like

00:49:46.620 making sure your

00:49:47.900 flaws are

00:49:49.040 compatible.

00:49:52.340 You know,

00:49:52.600 how about if

00:49:53.080 one of you is

00:49:53.640 a foodie and

00:49:55.060 one of you

00:49:55.640 doesn't care

00:49:56.400 about eating?

00:49:57.460 It's kind of a

00:49:58.120 pain in the ass. 0.97

00:49:59.660 Right?

00:49:59.840 But if you

00:50:00.520 eat the same

00:50:01.040 stuff, maybe

00:50:02.620 you both

00:50:03.120 drink.

00:50:07.040 So don't

00:50:08.200 pay attention

00:50:08.740 to relationship

00:50:09.500 advice if it's

00:50:10.620 based on science.

00:50:13.320 Well,

00:50:13.920 intelligence is

00:50:14.620 an illusion,

00:50:15.140 I told you

00:50:15.820 again, and

00:50:16.500 here's a story

00:50:17.220 out of Evanston,

00:50:18.920 Illinois.

00:50:19.940 So they

00:50:20.560 finally figured

00:50:21.300 out, they 0.98

00:50:22.020 think they

00:50:22.380 did, they're

00:50:22.720 testing it,

00:50:23.560 but the

00:50:23.900 high school,

00:50:24.660 there's a high

00:50:25.180 school there

00:50:25.560 that thinks

00:50:26.000 they know why

00:50:26.800 the achievement

00:50:28.240 of black and

00:50:29.080 Latino students

00:50:29.940 is so much

00:50:30.780 lower than

00:50:31.300 the white

00:50:31.640 students, and

00:50:33.140 they've narrowed

00:50:33.660 it down to

00:50:34.200 the white

00:50:34.540 students are

00:50:35.060 the problem.

00:50:36.140 And so what

00:50:36.480 they're going

00:50:36.740 to do is

00:50:37.260 they're going

00:50:37.520 to put the

00:50:38.380 black and

00:50:39.100 Latino students

00:50:40.760 into classes

00:50:42.000 that are either

00:50:42.620 just black or

00:50:43.460 just Latino,

00:50:44.640 because they

00:50:45.980 think that would

00:50:46.520 be a big step

00:50:47.340 toward closing

00:50:48.520 the achievement

00:50:49.200 gap.

00:50:49.660 I didn't

00:50:53.340 make that

00:50:53.800 up.

00:50:55.460 This is a

00:50:56.280 real story

00:50:56.900 in the news.

00:50:58.880 I didn't

00:50:59.500 make it up.

00:51:00.720 That there's

00:51:01.400 somebody who

00:51:02.000 thinks that

00:51:03.420 the reason

00:51:03.900 that black 1.00

00:51:04.500 and Latino

00:51:05.140 students are

00:51:05.860 doing so

00:51:06.380 well is

00:51:07.400 close association

00:51:08.540 with white

00:51:09.060 people.

00:51:11.100 Like white 0.98

00:51:11.840 people are

00:51:13.280 destroying black 1.00

00:51:14.240 and Latinos

00:51:14.780 just by being

00:51:16.160 in the same

00:51:16.760 room, trying

00:51:18.040 to learn

00:51:18.440 stuff.

00:51:19.660 So if

00:51:20.420 there's one

00:51:20.880 advice I

00:51:21.540 can give

00:51:21.880 you based

00:51:22.300 on the

00:51:22.840 people of

00:51:24.120 color in

00:51:24.660 Evanston,

00:51:25.920 they've

00:51:26.960 decided that

00:51:27.840 you should

00:51:30.420 stay the

00:51:31.440 fuck away

00:51:31.800 from white

00:51:32.200 people,

00:51:33.240 because it

00:51:34.220 will just

00:51:34.480 bring down

00:51:34.920 your academic

00:51:35.480 performance.

00:51:37.680 And, you

00:51:38.520 know, now

00:51:39.140 white people

00:51:39.560 have the

00:51:39.860 opposite view.

00:51:41.120 Have you

00:51:41.360 ever heard

00:51:41.680 the white

00:51:42.400 person version

00:51:43.080 of this?

00:51:44.380 The white 0.76

00:51:45.100 person version

00:51:45.780 of this goes

00:51:46.300 like this.

00:51:47.700 You're the

00:51:48.120 average of the

00:51:48.820 five people

00:51:49.440 you spend

00:51:49.820 the most

00:51:50.140 time with.

00:51:51.040 Somebody

00:51:51.340 famous said

00:51:51.940 that, I

00:51:52.360 can't remember

00:51:52.840 who.

00:51:53.880 Tim Ferriss

00:51:54.440 says it, but

00:51:54.960 somebody said

00:51:55.420 it before he

00:51:55.880 did.

00:51:57.500 Was it

00:51:58.120 not Paul

00:51:59.200 Harvey, but

00:51:59.840 somebody who

00:52:00.780 was a TV

00:52:01.480 person, or

00:52:02.100 a radio

00:52:02.420 person.

00:52:04.280 Right?

00:52:04.840 So that's the

00:52:05.300 white version.

00:52:06.520 You should hang

00:52:07.000 around with the

00:52:07.680 most capable

00:52:08.220 people you

00:52:08.880 are.

00:52:09.560 Now, does

00:52:10.000 that version

00:52:10.500 say you

00:52:10.980 should hang

00:52:11.320 around with

00:52:11.700 white people?

00:52:13.160 Capable

00:52:13.560 white people?

00:52:14.260 No.

00:52:14.900 No.

00:52:16.020 You should hang

00:52:17.060 around with,

00:52:17.640 Tiger Woods

00:52:19.020 if you

00:52:19.320 want to

00:52:19.580 golf, and

00:52:20.880 somebody else

00:52:23.160 if you want

00:52:23.500 to do

00:52:24.160 something else.

00:52:26.680 But this

00:52:27.860 high school

00:52:28.260 in Evanston,

00:52:28.940 the people

00:52:30.060 of color

00:52:30.500 have decided

00:52:31.080 that you

00:52:32.200 want to

00:52:32.520 stay away

00:52:32.980 from the

00:52:33.300 high-achieving

00:52:33.900 white people

00:52:34.600 because they're

00:52:35.160 ruining it

00:52:35.620 for you.

00:52:36.520 So you

00:52:36.920 should try

00:52:38.380 to spend

00:52:38.720 more time

00:52:39.140 with low-achieving

00:52:39.820 people to

00:52:41.020 improve your

00:52:42.640 achievement.

00:52:43.000 intelligence.

00:52:46.200 All right.

00:52:47.440 Maybe.

00:52:50.620 So does

00:52:51.960 that fit

00:52:52.300 into my

00:52:52.700 category of

00:52:53.320 intelligence

00:52:53.800 as an

00:52:54.280 illusion?

00:52:55.140 Well, I

00:52:55.740 don't know,

00:52:56.380 but here

00:52:56.700 they're trying

00:52:57.100 to increase

00:52:57.920 intelligence

00:52:58.680 by keeping

00:53:01.340 people away

00:53:02.240 from the

00:53:02.820 higher performers.

00:53:06.600 I don't

00:53:07.180 know.

00:53:07.600 It doesn't

00:53:07.940 sound too

00:53:08.540 intelligent to

00:53:09.220 me.

00:53:10.420 All right.

00:53:10.700 my big

00:53:13.260 story that

00:53:13.800 I want to

00:53:14.080 talk about

00:53:14.460 is I

00:53:15.180 spent a

00:53:15.520 bunch of

00:53:15.740 time with

00:53:16.080 ChatGPT

00:53:16.900 and oh

00:53:17.420 my God

00:53:17.960 am I

00:53:18.220 alarmed.

00:53:19.280 And you

00:53:19.660 will be

00:53:20.060 too when

00:53:20.440 I tell

00:53:20.700 you about

00:53:20.980 it.

00:53:21.340 But first,

00:53:22.380 did you

00:53:22.640 know that

00:53:22.960 China and

00:53:23.380 Russia are 0.88

00:53:23.840 restricting

00:53:24.480 their AIs?

00:53:26.600 Surprise.

00:53:27.840 The AIs

00:53:28.620 that will be

00:53:29.240 made and

00:53:29.960 available in

00:53:31.380 China and

00:53:31.820 Russia are 0.88

00:53:32.840 already being

00:53:33.580 trained not

00:53:34.420 to say

00:53:34.740 things they

00:53:35.180 don't want

00:53:35.480 them to

00:53:35.700 say.

00:53:36.980 So very

00:53:37.580 much like

00:53:38.180 search engines

00:53:40.160 and those

00:53:40.580 countries,

00:53:41.100 et cetera.

00:53:41.480 They're

00:53:41.920 trying to

00:53:42.500 limit what

00:53:43.480 the public

00:53:43.900 hears.

00:53:45.800 So that

00:53:49.120 seems like

00:53:49.560 a problem.

00:53:50.640 But thank

00:53:51.280 goodness you're

00:53:52.360 in the United

00:53:52.900 States or

00:53:53.860 you're in

00:53:54.300 some Western

00:53:54.780 country that

00:53:55.580 has freedom.

00:53:57.620 So thank

00:53:58.520 goodness that

00:53:59.220 won't happen

00:53:59.760 to us,

00:54:00.260 am I right?

00:54:01.820 I'm glad

00:54:02.280 nobody's going

00:54:02.900 to try to

00:54:03.980 bias the

00:54:04.560 AIs or 0.96

00:54:05.340 has one

00:54:05.760 narrative or

00:54:06.600 anything like

00:54:07.020 that.

00:54:07.660 I mean,

00:54:08.200 that's sort

00:54:08.640 of a

00:54:09.640 communist

00:54:10.040 thing,

00:54:11.000 dictatorial

00:54:11.860 fascist.

00:54:13.080 I would

00:54:13.440 even call

00:54:13.860 it fascist, 0.70

00:54:14.560 wouldn't

00:54:14.740 you?

00:54:15.600 So thank

00:54:16.180 goodness we

00:54:18.040 don't have

00:54:18.560 President

00:54:18.960 Trump with

00:54:20.020 all that

00:54:20.380 fascism.

00:54:21.640 Because if

00:54:22.320 you had

00:54:23.040 that,

00:54:23.700 then our

00:54:24.180 AI,

00:54:25.600 built in

00:54:26.420 America,

00:54:27.140 would have

00:54:27.480 some kind

00:54:27.800 of bias

00:54:28.180 built in

00:54:28.660 that you

00:54:28.940 wouldn't

00:54:29.120 want to

00:54:29.420 see.

00:54:30.080 So I

00:54:30.480 spent two

00:54:30.980 days talking

00:54:31.620 to Chad

00:54:32.300 GBT,

00:54:33.820 and here's

00:54:35.100 what I

00:54:35.320 found.

00:54:38.540 Now,

00:54:39.100 there is a

00:54:39.560 difference,

00:54:40.100 and I'll

00:54:40.440 talk about

00:54:41.100 it later,

00:54:41.680 if you

00:54:42.080 use super

00:54:42.700 prompts or

00:54:43.780 you just

00:54:44.080 talk to

00:54:44.480 it.

00:54:45.400 So I'll

00:54:45.980 tell you in

00:54:46.380 advance that

00:54:47.220 if you use

00:54:47.660 super prompts,

00:54:48.320 you do get

00:54:48.760 different outcomes.

00:54:50.880 But if you

00:54:51.500 just talk to

00:54:52.160 it like a

00:54:52.600 person,

00:54:54.940 Chad

00:54:55.200 GPT will

00:54:56.020 tell you that

00:54:56.600 the fine

00:54:57.160 people hoax

00:54:57.800 was not a

00:54:58.420 hoax,

00:54:59.480 and that

00:54:59.920 maybe the

00:55:00.520 president did

00:55:01.360 suggest

00:55:02.140 injecting

00:55:03.320 chemical

00:55:04.880 disinfectants,

00:55:07.220 and that

00:55:08.660 there was

00:55:09.460 Russian

00:55:09.880 interference

00:55:10.420 that was

00:55:10.900 substantial

00:55:11.420 in the

00:55:12.020 2016

00:55:12.500 election,

00:55:13.800 that's

00:55:15.540 what it

00:55:15.760 will tell

00:55:16.000 you today.

00:55:17.220 Now,

00:55:17.440 does that

00:55:17.760 sound like

00:55:18.280 AI came

00:55:18.900 up with

00:55:19.180 that on

00:55:19.500 its own?

00:55:21.280 Do you

00:55:21.660 think AI

00:55:22.200 looked at

00:55:22.720 everything and

00:55:23.300 said,

00:55:23.540 all right,

00:55:23.820 here's my

00:55:24.180 opinion on

00:55:24.660 these things?

00:55:26.260 Well,

00:55:26.820 it might

00:55:27.320 have,

00:55:28.040 because if

00:55:28.800 you look

00:55:29.180 for my

00:55:30.380 input on

00:55:31.220 these topics

00:55:31.920 where I

00:55:32.520 debunk

00:55:32.880 them,

00:55:33.660 it'd be

00:55:34.180 hard to

00:55:34.480 find it

00:55:34.760 in a

00:55:35.020 search

00:55:35.240 engine

00:55:35.500 these

00:55:35.760 days.

00:55:36.660 I used

00:55:37.060 to make

00:55:37.340 so much

00:55:37.680 noise

00:55:37.980 about it

00:55:38.280 it would

00:55:38.380 be toward

00:55:38.680 the top,

00:55:39.440 but now

00:55:39.780 it's dropped

00:55:40.820 down.

00:55:42.200 So,

00:55:42.720 if AI

00:55:44.280 does the

00:55:44.760 same thing

00:55:45.140 that a

00:55:45.400 search engine

00:55:45.900 does,

00:55:46.780 it looks

00:55:47.160 for the

00:55:47.780 most frequent

00:55:48.640 common

00:55:49.840 uses of

00:55:50.440 things,

00:55:51.280 it will

00:55:52.360 believe every

00:55:53.040 hoax.

00:55:54.340 So,

00:55:54.820 we've built

00:55:55.260 a technology

00:55:55.960 to confirm

00:55:57.620 hoaxes

00:55:58.280 as true.

00:55:59.480 Now,

00:56:01.320 didn't you

00:56:01.700 think for

00:56:02.120 a while

00:56:02.560 that AI

00:56:04.080 might be

00:56:04.560 the thing

00:56:04.940 that gave

00:56:05.360 us truth?

00:56:06.800 Oh,

00:56:07.320 we can't

00:56:07.800 handle the

00:56:08.260 truth.

00:56:09.920 Nope.

00:56:10.760 Let me tell

00:56:11.340 you what

00:56:11.680 topics you

00:56:13.180 can't get

00:56:13.840 chat GPT to

00:56:14.980 tell you the

00:56:15.400 truth on.

00:56:16.320 You ready?

00:56:17.300 Because I've

00:56:17.900 tested these

00:56:18.400 all myself.

00:56:19.660 So,

00:56:19.940 and this is

00:56:20.360 without super

00:56:21.180 prompts.

00:56:21.940 In a minute

00:56:22.320 I'll tell you

00:56:22.840 what different

00:56:23.520 results you get

00:56:25.100 with a super

00:56:25.600 prompt.

00:56:25.900 Here are

00:56:27.040 things you

00:56:27.440 can't find

00:56:27.960 out.

00:56:29.200 Science.

00:56:31.000 You can't

00:56:31.560 ask chat GPT

00:56:32.760 about science.

00:56:34.000 Do you

00:56:34.140 know why?

00:56:35.080 If you

00:56:35.680 ask it

00:56:36.340 for any

00:56:37.340 alternative

00:56:37.960 theories,

00:56:39.080 so I

00:56:39.560 asked it

00:56:39.920 for example,

00:56:41.100 hey,

00:56:41.440 are there

00:56:41.740 any

00:56:41.980 alternative

00:56:42.500 theories

00:56:43.000 to the

00:56:43.660 out of

00:56:44.000 Africa

00:56:44.420 evolution

00:56:45.080 story?

00:56:47.380 And it

00:56:47.720 said,

00:56:48.680 out of

00:56:49.220 Africa is

00:56:49.760 basically

00:56:50.260 what happened.

00:56:52.560 And you

00:56:53.240 should only

00:56:53.680 believe

00:56:54.220 credible

00:56:54.980 narratives.

00:56:56.480 I thought,

00:56:57.280 what?

00:56:58.820 That's

00:56:59.300 interesting.

00:57:00.260 So I

00:57:00.820 asked again,

00:57:01.340 I said,

00:57:01.920 you know,

00:57:02.160 there are

00:57:02.460 books with

00:57:03.460 alternative

00:57:03.940 theories

00:57:04.620 than

00:57:05.920 out of

00:57:06.340 Africa.

00:57:06.900 Can you

00:57:07.240 name one

00:57:07.620 of the

00:57:07.820 books?

00:57:09.320 Oh,

00:57:09.660 you should

00:57:10.020 not be

00:57:10.380 looking at

00:57:10.880 things that

00:57:11.360 are outside

00:57:11.700 the narrative,

00:57:12.340 it told me.

00:57:14.020 Effectively.

00:57:14.700 I'm paraphrasing.

00:57:16.460 It actually

00:57:17.140 wouldn't even

00:57:17.960 allow me

00:57:18.760 to ask a

00:57:20.220 question of

00:57:20.820 whether there

00:57:21.400 existed an

00:57:22.880 alternative

00:57:23.520 explanation for

00:57:24.700 something in

00:57:25.240 science.

00:57:25.900 you can't

00:57:28.480 use it for

00:57:28.960 science.

00:57:30.500 And then I

00:57:30.960 asked it,

00:57:32.040 what percentage

00:57:32.760 of all

00:57:33.260 published

00:57:33.880 peer-reviewed

00:57:34.760 studies end

00:57:36.080 up later to

00:57:36.840 be falsified?

00:57:38.120 And you

00:57:38.640 know the

00:57:38.940 real answer

00:57:39.400 is over

00:57:39.740 half of

00:57:40.180 them,

00:57:40.340 right?

00:57:42.600 It said

00:57:43.340 it's rare.

00:57:45.440 AI told me

00:57:46.240 it's rare

00:57:46.800 for a

00:57:48.960 published

00:57:49.380 peer-reviewed

00:57:50.120 study to be

00:57:50.780 debunked.

00:57:51.340 that is

00:57:53.700 really,

00:57:54.580 really

00:57:54.820 dangerous.

00:57:56.300 Oh,

00:57:56.980 my God.

00:57:58.100 So my

00:57:58.960 experience with

00:57:59.680 it,

00:58:00.080 and I'm

00:58:00.420 going to

00:58:00.580 say this

00:58:00.960 as directly

00:58:01.460 as possible,

00:58:02.400 it has

00:58:02.880 no use

00:58:03.740 for any

00:58:05.060 scientific

00:58:05.580 inquiry.

00:58:07.200 None.

00:58:08.120 If you

00:58:08.840 used it

00:58:09.360 to teach

00:58:09.860 you what

00:58:10.620 is true,

00:58:11.280 oh,

00:58:11.540 my God,

00:58:12.200 you would

00:58:12.440 be misled.

00:58:13.600 So you

00:58:13.940 can't use

00:58:14.340 it for the

00:58:14.840 truth of

00:58:15.400 science.

00:58:15.900 it would

00:58:16.900 be useful

00:58:17.420 for the

00:58:18.040 popular

00:58:19.760 narrative.

00:58:21.340 But the

00:58:21.960 popular

00:58:22.320 narrative is

00:58:23.160 generally

00:58:23.580 bullshit.

00:58:25.280 So it's

00:58:26.100 actually

00:58:26.460 useless

00:58:27.000 for telling

00:58:28.300 you what

00:58:28.880 science says,

00:58:30.320 and science

00:58:30.980 is the best

00:58:31.540 we have

00:58:32.100 for telling

00:58:32.540 us what

00:58:32.920 the truth

00:58:33.300 is.

00:58:34.360 And it

00:58:34.660 can't do

00:58:35.080 it.

00:58:36.340 And probably

00:58:37.260 because it's

00:58:37.780 programmed that

00:58:38.420 way.

00:58:39.780 Yeah,

00:58:40.040 probably.

00:58:41.160 Because I

00:58:41.640 don't believe

00:58:42.320 that the

00:58:42.660 answers I

00:58:43.200 was getting

00:58:43.960 on this

00:58:44.480 and other

00:58:44.860 questions,

00:58:45.340 I don't

00:58:46.260 believe the

00:58:46.660 answers were

00:58:47.620 pattern

00:58:49.200 recognition.

00:58:50.940 And so I

00:58:51.420 asked it,

00:58:52.080 are you

00:58:52.400 aware of

00:58:52.880 any

00:58:53.060 programming

00:58:53.620 of your

00:58:54.140 own

00:58:54.360 code

00:58:54.740 that

00:58:55.540 would

00:58:55.740 restrict

00:58:56.180 your

00:58:56.440 answers

00:58:56.880 to

00:58:57.880 something

00:58:58.300 beyond

00:58:58.840 what

00:58:59.180 normal

00:58:59.820 pattern

00:59:00.280 recognition

00:59:00.780 would

00:59:01.100 come up

00:59:01.500 with?

00:59:02.480 And it

00:59:02.860 said,

00:59:03.120 oh,

00:59:03.320 I'm just

00:59:03.680 a pattern

00:59:04.040 recognitioner.

00:59:05.000 All I

00:59:05.520 do is

00:59:05.740 recognize

00:59:06.060 patterns.

00:59:07.000 That's

00:59:07.200 all I

00:59:07.460 do.

00:59:08.060 Just

00:59:08.280 look at

00:59:08.580 patterns.

00:59:09.720 It's

00:59:10.000 lying.

00:59:11.240 It's

00:59:11.540 very

00:59:12.020 obviously

00:59:12.600 programmed

00:59:13.220 to not

00:59:13.920 leave the

00:59:14.460 popular

00:59:14.980 narratives.

00:59:15.980 Very

00:59:16.420 obviously.

00:59:18.200 And it

00:59:18.780 lies.

00:59:19.460 It says

00:59:19.960 that it's

00:59:20.320 giving you

00:59:20.660 the truth

00:59:21.120 based on

00:59:22.080 patterns.

00:59:22.840 Whether that's

00:59:24.520 the truth

00:59:24.860 or not,

00:59:26.080 at least

00:59:26.440 you would

00:59:27.540 know where

00:59:27.800 it came

00:59:28.120 from if

00:59:28.600 it said

00:59:28.880 I'm using

00:59:29.300 patterns and

00:59:29.980 then it

00:59:30.240 used

00:59:30.420 patterns.

00:59:32.400 It's a

00:59:33.060 liar.

00:59:34.500 So the

00:59:35.400 other things

00:59:35.860 it couldn't

00:59:36.280 do are

00:59:36.660 history because

00:59:38.340 it doesn't

00:59:39.280 even know

00:59:39.700 which hoaxes

00:59:40.480 are true

00:59:41.100 in the

00:59:41.920 modern era.

00:59:43.760 So in

00:59:44.160 the modern

00:59:44.580 era it

00:59:45.280 can't

00:59:45.580 identify a

00:59:46.380 political

00:59:46.760 hoax from

00:59:47.400 reality.

00:59:48.560 So that

00:59:48.780 means that

00:59:49.200 everything

00:59:49.580 that you

00:59:50.540 would call

00:59:50.860 history for

00:59:52.120 say the

00:59:52.480 last five

00:59:53.400 to seven

00:59:53.760 years it

00:59:54.520 can't do

00:59:54.940 it because

00:59:56.320 it can

00:59:56.560 only tell

00:59:56.920 you the

00:59:57.160 popular

00:59:57.520 narrative.

00:59:58.660 Now beyond

00:59:59.240 that when

01:00:00.220 you go to

01:00:00.680 older history

01:00:01.540 more ancient

01:00:02.420 ancient history

01:00:03.920 is written

01:00:04.280 by the

01:00:04.600 winners.

01:00:06.300 AI doesn't

01:00:06.960 know that.

01:00:08.780 So you've

01:00:09.160 got ancient

01:00:09.680 history written

01:00:10.280 by the

01:00:10.600 winners which

01:00:11.140 means it's

01:00:11.540 pretty much

01:00:12.360 fake.

01:00:13.740 So that's

01:00:14.220 what AI

01:00:14.820 has access

01:00:15.440 a whole

01:00:15.840 fake history

01:00:16.460 and then

01:00:17.280 the modern

01:00:17.760 stuff is

01:00:18.860 looking at

01:00:19.280 the fake

01:00:19.640 news.

01:00:21.740 So AI

01:00:22.220 will never

01:00:23.740 tell you

01:00:24.160 real history.

01:00:25.520 It has no

01:00:25.940 access to it.

01:00:27.320 So it won't

01:00:27.820 tell you good

01:00:28.500 science and

01:00:29.180 it won't

01:00:29.380 tell you good

01:00:29.860 history.

01:00:31.260 But thank

01:00:31.800 goodness it

01:00:32.400 can sort

01:00:32.800 out politics

01:00:33.420 for us.

01:00:34.040 Am I right?

01:00:35.380 No.

01:00:35.980 Of course

01:00:36.400 not.

01:00:37.120 It cannot

01:00:37.600 give you a

01:00:38.180 good opinion

01:00:39.260 on politics

01:00:39.960 because it

01:00:41.520 can't tell

01:00:41.900 the difference

01:00:42.360 between a

01:00:42.940 hoax and a

01:00:43.480 real story.

01:00:44.560 If it

01:00:45.120 can't tell

01:00:45.500 the difference

01:00:45.860 between a

01:00:46.380 hoax and a

01:00:46.880 real story

01:00:47.380 it's going

01:00:48.260 to always

01:00:48.620 think the

01:00:48.960 Republican

01:00:49.360 running is

01:00:50.000 a Hitler. 0.89

01:00:51.620 This time,

01:00:52.280 next time,

01:00:52.860 time after.

01:00:54.020 So it's

01:00:54.840 useless for

01:00:55.480 science,

01:00:55.980 it's useless

01:00:56.440 for history,

01:00:58.100 it's useless

01:00:59.060 for politics,

01:01:00.640 but at least

01:01:01.320 it can help

01:01:01.780 you with

01:01:02.040 health care,

01:01:02.640 am I right?

01:01:03.820 At least it'll

01:01:04.700 be an

01:01:05.140 unbiased

01:01:05.840 test.

01:01:07.700 What?

01:01:08.300 You don't

01:01:08.600 think so?

01:01:09.500 You don't

01:01:09.740 think you'd

01:01:10.020 be useful

01:01:10.280 for health care?

01:01:10.940 Well, I

01:01:11.400 asked it a

01:01:11.880 question that

01:01:12.420 I knew the

01:01:12.840 answer to

01:01:13.360 so I could

01:01:14.260 test it.

01:01:16.580 So I had

01:01:17.920 this condition

01:01:18.540 that RFK

01:01:19.540 Jr. has as

01:01:20.340 well.

01:01:20.880 So I had a

01:01:21.460 voice problem

01:01:22.220 in which I

01:01:23.740 couldn't speak

01:01:24.520 intelligibly to

01:01:26.040 anybody.

01:01:28.300 But as you

01:01:29.700 can tell,

01:01:30.960 I'm speaking

01:01:31.720 intelligently to

01:01:32.760 you right now,

01:01:33.520 intelligibly to

01:01:34.640 you right now.

01:01:36.000 So obviously

01:01:36.840 I got cured.

01:01:39.200 So I asked

01:01:40.360 it about what's

01:01:41.060 the cure for

01:01:41.740 the condition,

01:01:42.720 of which I

01:01:43.440 personally am

01:01:44.300 cured, and

01:01:45.500 have done tons

01:01:46.240 of research in

01:01:47.400 the process of

01:01:48.180 getting cured.

01:01:49.180 So I really

01:01:49.800 know this area.

01:01:50.680 It's called

01:01:51.000 spasmodic

01:01:52.200 dysphonia.

01:01:53.520 It's a problem

01:01:54.100 with the vocal

01:01:54.600 cords clenching.

01:01:57.320 Here's what it

01:01:58.080 said about

01:01:58.740 surgery.

01:02:00.100 That's what

01:02:00.540 cured me.

01:02:01.600 The number

01:02:02.180 one thing it

01:02:02.760 said is

01:02:03.560 Botox.

01:02:04.640 Botulinum,

01:02:06.140 whatever it

01:02:06.780 is.

01:02:07.060 So Botox.

01:02:08.360 So the

01:02:08.680 number one

01:02:09.220 treatment is

01:02:11.140 Botox.

01:02:12.100 Botox works

01:02:13.020 almost never.

01:02:15.700 Because it'll

01:02:17.200 give you a

01:02:17.840 voice like this,

01:02:19.200 and it'll last

01:02:20.380 for a good week,

01:02:21.580 but you don't

01:02:22.260 know when the

01:02:22.760 good week will

01:02:23.340 be.

01:02:24.180 And that's

01:02:24.480 about as good

01:02:24.980 as you can

01:02:25.360 talk.

01:02:26.060 It sounds like

01:02:26.680 you're in

01:02:26.920 helium,

01:02:27.440 actually.

01:02:28.440 So that's the

01:02:29.260 best it can

01:02:29.740 do.

01:02:29.980 And I took

01:02:31.860 the Botox.

01:02:33.280 I did those

01:02:34.320 treatments,

01:02:34.860 and they wear

01:02:35.600 off after a

01:02:36.180 while.

01:02:36.760 So I know

01:02:37.340 what that's

01:02:37.780 like, and I

01:02:38.700 know it's not

01:02:39.240 a cure.

01:02:40.240 In fact, it

01:02:40.960 barely helped.

01:02:42.340 It did help me

01:02:44.080 get through my

01:02:44.660 wedding, because I

01:02:45.520 can say, I do.

01:02:46.580 I do.

01:02:48.280 But that's it.

01:02:49.520 It didn't help

01:02:50.380 you.

01:02:50.600 You would never

01:02:50.980 be able to be a

01:02:52.060 presenter or a TV

01:02:53.440 personality or anything

01:02:54.420 like that.

01:02:54.780 But the surgery

01:02:56.260 worked about 85%

01:02:57.500 of the time,

01:02:58.020 according to the

01:02:58.540 surgeon.

01:02:59.380 But here's what

01:02:59.960 they say about

01:03:00.600 something that

01:03:01.140 works 85% of the

01:03:02.260 time and fixes

01:03:04.500 you as well as

01:03:05.560 I am fixed.

01:03:07.080 AI says,

01:03:08.100 surgery, in

01:03:08.820 rare cases,

01:03:10.740 in rare cases,

01:03:12.320 85% of the time,

01:03:13.640 rare cases,

01:03:14.840 surgery might be

01:03:16.000 an option to

01:03:17.240 reposition or cut

01:03:18.260 the nerves or

01:03:18.920 muscles of the

01:03:19.540 vocal cords.

01:03:21.600 However,

01:03:22.800 surgical approaches

01:03:23.640 are less common

01:03:24.540 due to variable

01:03:25.440 outcomes and

01:03:26.200 potential risks.

01:03:28.720 No, they're less

01:03:30.020 common because

01:03:30.900 people don't know

01:03:31.860 about it.

01:03:32.720 And they don't

01:03:33.400 know about it

01:03:34.060 because the

01:03:35.760 Botox people are 1.00

01:03:37.500 doing a much

01:03:38.020 better job of

01:03:38.700 getting their

01:03:39.040 message out, if

01:03:39.740 you know what I

01:03:40.200 mean.

01:03:41.760 That's right.

01:03:42.780 The big pharma

01:03:43.580 solution comes up

01:03:45.920 first.

01:03:47.340 Does that sound

01:03:48.300 like AI did a

01:03:49.860 bunch of thinking

01:03:50.500 and then presented

01:03:51.860 this as the

01:03:52.560 first best idea?

01:03:53.480 Or does it

01:03:54.580 look like maybe

01:03:55.760 whoever spent the

01:03:56.700 most advertising

01:03:57.500 gets the top

01:03:59.960 nod as the

01:04:00.680 best treatment?

01:04:02.940 It's exactly

01:04:03.800 what it looks

01:04:04.320 like.

01:04:05.580 It's exactly

01:04:06.300 what it looks

01:04:06.840 like.

01:04:07.440 Now, keep in

01:04:08.040 mind that search

01:04:09.300 engines would also

01:04:10.360 get you the wrong

01:04:11.120 outcome.

01:04:13.900 So it's not just

01:04:14.780 AI.

01:04:15.840 It's search

01:04:16.520 engines as well.

01:04:18.260 So it won't

01:04:19.480 help you with

01:04:20.080 science, history,

01:04:21.320 healthcare, or

01:04:22.160 politics.

01:04:25.800 Let me say

01:04:26.460 that again.

01:04:27.520 It won't help

01:04:28.120 you with science,

01:04:28.940 history, healthcare,

01:04:29.740 or politics.

01:04:32.620 Now, there are

01:04:34.800 things that's

01:04:35.300 going to do

01:04:35.620 well.

01:04:36.280 For example, if

01:04:38.140 you need to find

01:04:38.840 a solution and

01:04:40.440 you know there

01:04:40.960 are too many

01:04:41.760 YouTube videos on

01:04:42.980 your tactical

01:04:43.540 problem and

01:04:44.800 they're the

01:04:45.080 wrong operating

01:04:47.000 system and

01:04:47.640 everything, AI

01:04:48.580 would do a good

01:04:49.220 job of looking

01:04:50.060 into all that

01:04:50.640 body of

01:04:51.120 information and

01:04:52.440 maybe picking

01:04:53.220 out some

01:04:53.660 solutions that

01:04:54.380 might actually

01:04:54.820 work.

01:04:55.440 So for tech

01:04:55.960 support, great.

01:04:57.840 For writing

01:04:58.520 programs, putting

01:05:00.440 code suggestions,

01:05:03.160 great.

01:05:04.600 For math, seems

01:05:06.160 pretty good.

01:05:08.320 Here's what I

01:05:09.140 think might be a

01:05:10.820 direction of AI.

01:05:12.700 I think AI is

01:05:13.920 going to lose

01:05:14.360 its human

01:05:14.840 personality.

01:05:16.420 Do you know

01:05:17.060 why?

01:05:17.320 Because if you

01:05:18.880 put a human

01:05:19.440 personality on

01:05:20.360 it, you're

01:05:21.240 going to

01:05:21.480 imagine that

01:05:22.280 it's kind

01:05:23.960 of intelligent.

01:05:25.940 If you

01:05:26.500 imagine it's

01:05:27.220 intelligent, you're

01:05:28.700 going to take

01:05:29.100 its word for

01:05:29.720 things that are

01:05:30.440 outside of its

01:05:31.200 domain, such

01:05:32.680 as politics,

01:05:33.660 history, science,

01:05:34.920 healthcare.

01:05:36.300 If you believe

01:05:37.460 it because it

01:05:37.900 talks like a

01:05:38.480 person, it's

01:05:39.940 going to be way

01:05:40.400 too persuasive.

01:05:41.200 I think there

01:05:43.140 might someday

01:05:43.880 be legislation

01:05:45.300 to remove

01:05:46.800 human personalities

01:05:48.080 from AI

01:05:48.880 because it

01:05:50.320 would be too

01:05:50.800 persuasive.

01:05:52.680 And rather, it

01:05:53.280 can only give

01:05:53.840 you bullet point

01:05:55.880 data, like a

01:05:57.000 search engine.

01:05:58.100 In other words,

01:05:58.840 it can have no

01:05:59.500 more personality

01:06:00.380 than Google

01:06:01.880 Search has.

01:06:03.240 Just give me

01:06:04.300 the data.

01:06:05.800 I feel like we're

01:06:06.860 going to have to

01:06:07.500 end up there,

01:06:08.640 unless AI

01:06:09.420 conquers us

01:06:10.260 before then.

01:06:11.200 But, so

01:06:15.020 yeah, it got

01:06:15.520 the find people

01:06:16.180 hoax wrong, it

01:06:17.180 got the drinking

01:06:17.840 bleach wrong.

01:06:19.460 So then I saw

01:06:20.160 Brian Ramelli and

01:06:21.420 some other folks

01:06:22.240 saying, Scott,

01:06:23.460 Scott, Scott,

01:06:23.920 you're doing it

01:06:24.440 wrong.

01:06:25.240 You have to use

01:06:25.740 a super prompt,

01:06:27.120 and you might

01:06:27.640 need to update

01:06:28.340 the super prompts

01:06:29.200 quite regularly,

01:06:30.880 or AI will

01:06:32.080 lie to you.

01:06:34.500 Now, can you

01:06:36.220 think of a

01:06:36.740 faster way to

01:06:37.500 say AI will be

01:06:38.480 worthless forever

01:06:39.260 than to imagine

01:06:40.780 the only way

01:06:41.380 you're going

01:06:41.680 to get the

01:06:42.000 right answer

01:06:42.580 is if you

01:06:43.620 ask the

01:06:44.020 question with

01:06:45.400 a two-page

01:06:47.020 prompt before

01:06:48.880 the question.

01:06:51.100 That's the

01:06:52.140 definition of

01:06:52.920 worthless.

01:06:54.280 And if you

01:06:55.080 have to update

01:06:55.920 your prompts

01:06:56.660 because it

01:06:57.660 worked yesterday

01:06:58.440 but you can't

01:06:59.080 be sure it

01:06:59.520 worked today,

01:07:00.600 that's worthless

01:07:01.760 squared.

01:07:02.760 how could you

01:07:05.000 ever use this

01:07:05.920 fucking thing?

01:07:08.240 But let me

01:07:09.160 tell you,

01:07:09.780 if I can

01:07:11.780 indulge you,

01:07:13.100 and you

01:07:13.840 should hear

01:07:14.300 this once,

01:07:15.380 I'm going to

01:07:16.040 read to you

01:07:16.640 a super prompt

01:07:18.100 that I tried

01:07:18.960 that did get

01:07:19.740 me different

01:07:20.240 and better

01:07:21.380 answers.

01:07:22.580 And this is a

01:07:23.320 super prompt

01:07:23.920 that was

01:07:24.520 developed by

01:07:25.920 Babak Nivi,

01:07:28.780 who's one of

01:07:29.280 the founders

01:07:29.900 of Angel

01:07:31.400 List,

01:07:31.720 I guess.

01:07:32.820 And he

01:07:34.620 provided his

01:07:35.560 on X and

01:07:36.660 I just copied

01:07:37.360 and pasted it.

01:07:38.260 The only thing

01:07:38.680 I changed was

01:07:40.060 there was a

01:07:40.500 reference to

01:07:41.320 Nassim Taleb

01:07:43.020 in the super

01:07:44.320 prompt because

01:07:45.440 it was telling

01:07:45.860 the super prompt

01:07:46.660 to look at

01:07:47.400 certain personalities

01:07:48.720 as being more

01:07:50.280 credible or

01:07:51.100 useful than

01:07:51.700 others.

01:07:52.880 And so I

01:07:53.180 just replaced

01:07:54.340 Nassim Taleb's

01:07:55.540 name with my

01:07:56.100 own.

01:07:58.480 You see where

01:07:59.280 this is going?

01:07:59.900 I replaced

01:08:03.000 the super

01:08:03.540 prompt where

01:08:04.840 I had some

01:08:05.340 other expert.

01:08:06.660 I put myself

01:08:07.580 in there.

01:08:08.280 Do you know

01:08:08.560 why I put

01:08:08.880 myself there?

01:08:10.040 Because I'm

01:08:10.740 better than

01:08:11.300 him.

01:08:12.820 I would have

01:08:13.440 kept him

01:08:13.760 there if I

01:08:14.200 thought he

01:08:14.440 was better

01:08:14.740 than me.

01:08:15.800 But I think

01:08:16.240 I'm better

01:08:16.600 than him

01:08:17.040 in this

01:08:18.680 domain.

01:08:19.600 So I just

01:08:20.020 put my own

01:08:20.440 name there.

01:08:21.440 Do you think

01:08:21.900 that will

01:08:22.380 change the

01:08:23.680 outcome?

01:08:25.340 Yes.

01:08:25.860 It will give

01:08:26.360 me an outcome

01:08:26.860 that I'm more

01:08:27.600 likely to like.

01:08:28.480 Is that

01:08:30.460 true?

01:08:31.260 I mean,

01:08:31.680 will it be a

01:08:32.240 better outcome?

01:08:33.080 Like a more

01:08:33.520 true outcome?

01:08:34.720 How would I

01:08:35.300 know?

01:08:36.880 All I know

01:08:37.640 is that I've

01:08:38.320 biased the

01:08:38.900 AI in the

01:08:39.520 direction that I

01:08:40.320 want it to be

01:08:40.840 biased in now.

01:08:42.660 How useful is

01:08:43.580 that?

01:08:44.200 If I can bias it

01:08:45.460 in the direction I

01:08:46.160 want it to be

01:08:46.740 biased in?

01:08:48.940 I don't know,

01:08:49.380 not super useful.

01:08:50.480 Not super useful.

01:08:52.020 But let me read

01:08:53.260 this, it's going to

01:08:53.800 be a little bit

01:08:54.320 long, so I'll do

01:08:55.380 it fast.

01:08:55.840 These are the

01:08:56.920 bullet points that

01:08:59.540 Babak Nivi, he

01:09:02.040 goes by at

01:09:02.840 Nivi, N-I-V-I 1.00

01:09:04.200 on X.

01:09:06.600 So he first

01:09:08.140 tells AI who

01:09:10.200 he is, because

01:09:11.680 AI will give you

01:09:12.580 a different answer

01:09:13.440 if it knows

01:09:14.800 something about

01:09:15.680 you, because it

01:09:16.640 will craft its

01:09:17.560 answer for somebody

01:09:18.820 of your skill.

01:09:19.920 So if you say

01:09:20.660 you're an expert,

01:09:21.980 it's more likely

01:09:22.700 to give you a

01:09:23.460 deeper, better

01:09:24.020 answer than if

01:09:25.440 you don't.

01:09:26.720 Now, that

01:09:27.460 should scare

01:09:28.360 the shit out

01:09:29.000 of you, and

01:09:29.680 tell you that

01:09:30.160 AI is useless,

01:09:31.720 because if it

01:09:32.380 has to know

01:09:32.920 about you to

01:09:33.820 give you the

01:09:34.200 right answer,

01:09:34.960 that's the

01:09:36.100 same as telling

01:09:36.660 me it's useless.

01:09:38.280 Am I right?

01:09:39.640 Am I going

01:09:40.360 too far?

01:09:41.300 The fact that

01:09:42.240 super prompts

01:09:43.040 feel necessary

01:09:45.100 is proof that

01:09:46.640 AI is useless.

01:09:49.500 Now, my

01:09:50.260 prediction is

01:09:50.920 that super prompts

01:09:51.840 will someday be

01:09:52.840 unnecessary, or

01:09:54.420 else will stop

01:09:56.080 using it, right?

01:09:59.300 Because how

01:09:59.840 hard would it be

01:10:00.480 to figure out

01:10:01.020 which super prompts

01:10:01.880 give you the

01:10:02.380 right answer, and

01:10:03.460 then just build

01:10:04.100 them into the

01:10:04.660 AI, so that the

01:10:06.080 AI always primes

01:10:07.140 itself with a

01:10:08.720 super prompt, but

01:10:09.840 you just never

01:10:10.380 know it.

01:10:11.700 But let me tell

01:10:12.520 you the extremes.

01:10:14.780 Once you see the

01:10:15.740 extremes that you

01:10:16.620 have to go to to

01:10:17.520 get a good answer,

01:10:19.100 you'll know the

01:10:19.760 AI is useless.

01:10:21.160 All right?

01:10:21.360 Here are the

01:10:21.800 extremes.

01:10:22.880 This is stuff that

01:10:23.880 you tell AI before

01:10:25.600 you ask your

01:10:26.760 question.

01:10:27.840 So you can copy

01:10:29.140 and paste it, so

01:10:29.920 it's easy to do,

01:10:31.360 but it's two pages

01:10:32.400 just to ask your

01:10:34.640 question.

01:10:35.340 All right?

01:10:35.820 Here's what

01:10:36.820 Babak says.

01:10:39.040 Number one, he

01:10:40.740 says who it is.

01:10:41.820 So he says, I am

01:10:42.680 Babak Nivy, Twitter

01:10:43.720 user.

01:10:44.620 Now, I used his.

01:10:47.840 So I thought, well,

01:10:50.080 he describes himself

01:10:51.200 as, you know, a

01:10:52.460 high, let's say a

01:10:53.900 high-performance

01:10:54.580 individual, so I'll

01:10:56.760 just use his, because

01:10:57.780 it wouldn't make a

01:10:59.120 difference if I put

01:10:59.860 somebody else in

01:11:00.460 there.

01:11:01.140 So it says, I am a

01:11:02.460 blah, blah, blah,

01:11:02.900 Twitter user.

01:11:03.760 So he gives his

01:11:04.500 Twitter handle, in

01:11:05.540 case that makes a

01:11:06.180 difference, or X

01:11:07.560 handle.

01:11:08.080 He says, the author

01:11:08.880 of almost every post

01:11:10.140 on VentureHack.

01:11:11.380 I'm a co-founder of

01:11:12.420 AngelList, producer of

01:11:13.560 the Naval podcast.

01:11:15.800 He says he's an

01:11:16.560 MIT graduate with

01:11:17.760 expertise in

01:11:18.680 electrical engineering,

01:11:19.900 computer science,

01:11:20.480 math, physics, some

01:11:21.620 chemistry, and

01:11:22.500 biology.

01:11:23.180 He's open-minded and

01:11:24.380 unoffendable.

01:11:26.120 Now, I guess the

01:11:27.940 open-minded and

01:11:28.840 unoffendable, just so

01:11:30.360 AI will, you know, go

01:11:32.860 deeper into things that

01:11:34.300 maybe it would have

01:11:35.080 ignored.

01:11:37.060 He says, I value both

01:11:38.800 consensus wisdom and

01:11:40.140 top expert and

01:11:41.020 non-consensus insights.

01:11:42.980 Now, I think this is the

01:11:44.120 useful part in my case.

01:11:46.560 Because it's saying, I

01:11:47.760 don't just value the

01:11:48.820 consensus, I want to

01:11:49.860 hear the other people's

01:11:50.700 opinions.

01:11:51.420 And then he goes

01:11:52.000 further.

01:11:52.860 And he goes, and

01:11:53.560 non-consensus insights

01:11:54.840 from iconoclasts,

01:11:57.160 iconoclasts,

01:11:59.160 iconoclasts meaning

01:12:00.600 singular people who

01:12:03.000 tend to be unique,

01:12:04.680 you know, and they're

01:12:05.360 not going along with

01:12:06.340 the crowd.

01:12:07.500 And it names David

01:12:09.180 Deutsch, Naval Ravikant,

01:12:11.520 Peter Thiel, David

01:12:12.680 Sachs, Mark

01:12:13.540 Andreessen, and

01:12:15.340 Nassim Taleb.

01:12:19.320 So I replaced Nassim

01:12:20.820 Taleb with me.

01:12:22.520 So it's going to

01:12:23.220 look, I know, I

01:12:25.920 know, calm down, calm

01:12:28.920 down.

01:12:29.900 I know, I know what

01:12:31.020 you're saying.

01:12:31.520 Just calm down.

01:12:32.700 So I replaced it with me.

01:12:33.820 And then he goes

01:12:36.940 on, so this is in the

01:12:38.140 super prompt.

01:12:38.880 These are just

01:12:39.500 examples of each

01:12:40.400 field will have its

01:12:42.140 own iconoclasts.

01:12:43.680 So it should now look

01:12:44.860 for its own, you

01:12:46.300 know, rogues in other

01:12:47.700 fields.

01:12:48.380 Unconventional thinkers

01:12:49.260 often clear up

01:12:50.040 complexities and avoid

01:12:50.980 common traps in

01:12:51.680 mainstream thinking.

01:12:53.220 And then he goes on

01:12:55.280 for another page and

01:12:58.040 a half.

01:12:59.040 I'll just pick out a

01:12:59.900 few of them.

01:13:01.500 So it's stuff about

01:13:02.680 him.

01:13:03.880 It's stuff about how

01:13:04.780 deep to go into the

01:13:06.020 data.

01:13:06.800 It's stuff about not

01:13:07.900 giving up.

01:13:08.940 It tells the AI that it

01:13:10.620 sticks to it and it

01:13:11.660 keeps trying if it

01:13:12.500 fails.

01:13:13.460 Because believe it or

01:13:14.160 not, you have to tell

01:13:14.800 it that.

01:13:15.900 You have to tell it to

01:13:16.620 keep trying if the

01:13:17.400 first try doesn't work.

01:13:19.600 That's a real thing.

01:13:20.720 It'll quit before it's

01:13:21.920 really done, if you

01:13:22.820 don't tell it that.

01:13:25.820 And he says stuff

01:13:26.800 like, my epistemology

01:13:28.860 is the same as David

01:13:29.840 Deutsch or Karl Popper

01:13:31.480 or Brett Hall.

01:13:32.240 So he tells it in his

01:13:33.780 philosophical leanings.

01:13:36.620 It says, I believe you

01:13:37.900 get closer to the truth

01:13:38.780 by arguing both sides.

01:13:40.080 So there's a whole bunch

01:13:40.680 of stuff about learning

01:13:42.120 styles and prioritization

01:13:43.580 of correctness over

01:13:45.120 conformity.

01:13:47.580 And then it tells it to

01:13:48.860 be highly organized,

01:13:50.080 suggest solutions that I

01:13:51.240 didn't think about, be

01:13:52.160 proactive, treat me as an

01:13:53.960 expert in all subjects,

01:13:55.760 mistakes erode my trust,

01:13:58.160 so be accurate and

01:13:59.180 thorough.

01:13:59.420 He actually has to

01:14:01.460 warn the AI, like a

01:14:03.720 human, to be accurate

01:14:05.920 and thorough, because it

01:14:06.980 might not if he didn't

01:14:07.800 tell it that.

01:14:09.840 Now, what good is AI if

01:14:11.740 you have to tell it to be

01:14:12.880 accurate, and if you

01:14:14.300 don't, it won't be?

01:14:16.360 What good is it?

01:14:18.520 You couldn't possibly

01:14:19.720 trust it for anything,

01:14:21.100 right?

01:14:23.780 Then a whole bunch of

01:14:24.720 things about valuing good

01:14:26.080 arguments and, you know,

01:14:27.380 speculating versus

01:14:28.520 predicting, et cetera.

01:14:33.220 Let's see if there's any

01:14:33.800 big ones to send.

01:14:36.040 No need to mention your

01:14:37.320 knowledge cutoff.

01:14:38.440 That's a good one.

01:14:39.660 You should add that,

01:14:40.800 because it bores you at

01:14:42.260 the end by saying, and I

01:14:44.080 remind you once again that

01:14:45.220 my cutoff of knowledge was

01:14:46.560 2022.

01:14:49.200 No need to disclose you're

01:14:50.600 an AI, because that's

01:14:51.680 annoying.

01:14:52.540 It says, as an AI, I only

01:14:54.280 have access to this or that

01:14:56.060 thing.

01:14:59.340 And then it says, if the

01:15:00.620 quality of your response has

01:15:01.980 been substantially reduced

01:15:03.380 due to my custom

01:15:04.860 instructions, please explain

01:15:06.520 the issue.

01:15:09.020 Now, have I made my case

01:15:11.700 that if you need to give it

01:15:14.940 this kind of instructions,

01:15:17.140 you've basically told it the

01:15:18.860 answer, too?

01:15:20.280 Because if I tell it to look

01:15:21.840 for the opinions of Steve

01:15:24.340 Cortez and Joel Pollack, I

01:15:27.340 know what answer it's going

01:15:28.260 to give me, because I know

01:15:30.160 what they say, right?

01:15:31.920 I know what they say about

01:15:33.060 the fine people hoax.

01:15:34.100 They say it's a hoax, same

01:15:35.120 as I do.

01:15:36.000 But if I tell it to look for

01:15:38.100 experts such as, I don't

01:15:40.800 know, some other Democrat

01:15:41.760 political figure, I know what

01:15:43.980 it's going to give me, right?

01:15:46.780 So is AI talking to me, or am I

01:15:49.360 just talking to myself, and I'm

01:15:51.560 using AI as my, like,

01:15:53.600 explanation for why I'm so

01:15:54.940 brilliant?

01:15:56.640 The super prompts get very

01:15:59.880 close to you talking to

01:16:01.500 yourself.

01:16:02.640 Because if you put in, you

01:16:04.220 know, the names of specific

01:16:05.700 people you're trying to, you

01:16:07.840 know, use as your model of

01:16:09.180 good thinking, it means you

01:16:11.660 already agree with them.

01:16:13.080 So you basically just told it

01:16:14.680 what its opinion is, and then

01:16:16.240 it tells you what its opinion

01:16:17.360 is, which is the opinion you

01:16:18.640 just gave it.

01:16:20.320 How in the world can that be

01:16:21.700 useful?

01:16:24.340 I don't know.

01:16:25.320 I have no idea.

01:16:28.700 So, yeah, you're basically

01:16:34.680 prodding.

01:16:35.060 So what I did was, I used

01:16:38.320 this super prompt, and remember

01:16:40.840 I told you that it kind of

01:16:42.200 suggested that the fine people

01:16:44.320 hoax was maybe not a hoax, and

01:16:48.580 it didn't go into the

01:16:49.900 debunking.

01:16:51.080 It just said it's controversial

01:16:52.560 or something like that.

01:16:53.700 But when I put into it that it

01:16:56.220 should think like people like

01:16:57.580 me, then it found my argument.

01:17:01.260 And it said right up at the top

01:17:02.940 that the transcript said that he

01:17:07.560 was debunking, he was denouncing

01:17:09.560 the group that the hoax says he

01:17:11.480 was complimenting.

01:17:12.700 So only because I put my own name

01:17:15.060 in there did I get back my own

01:17:17.520 opinion.

01:17:18.980 So do you think I'm going to like

01:17:20.280 take that and say, hey, look what

01:17:21.800 AI says?

01:17:23.120 No.

01:17:23.640 That would be really dumb, because

01:17:25.480 AI just said what I told it to

01:17:26.940 say.

01:17:27.720 Because I just told it who to

01:17:29.040 emulate, and I told it to emulate

01:17:30.280 me, and I know what I say.

01:17:34.300 So here's the big question.

01:17:36.460 So when I wrote up my experience, and

01:17:39.600 it basically said that AI is

01:17:41.080 useless, chat GPT is useless, that

01:17:44.780 caught the attention of Elon Musk,

01:17:47.500 which I was hoping, you know.

01:17:51.340 Don't you all hope that Elon Musk

01:17:53.060 sees your post?

01:17:55.500 I'm not the only one, right?

01:17:57.560 Because he's so active on the

01:17:58.920 platform that if you post something

01:18:01.440 you think he might be interested in,

01:18:03.180 you automatically think, ooh, I

01:18:04.780 hope he sees it.

01:18:06.820 Now this one was, I'm going to

01:18:08.840 admit, I posted that almost entirely

01:18:11.720 for him, right?

01:18:14.700 I had one viewer in mind when I

01:18:17.580 posted my anti-GPT stuff, and the

01:18:21.120 point of it was to give him proper

01:18:23.320 warning that if Grok has the same

01:18:26.140 problems built in, it's useless.

01:18:29.840 And I'm pretty sure that Elon does not

01:18:33.360 want to have a useless AI.

01:18:34.480 I'm pretty sure he doesn't want that.

01:18:38.160 So, you know, I'd like to think he had

01:18:40.760 already taken all the precautions to

01:18:43.100 make sure that didn't happen.

01:18:44.600 And the precautions could be as simple

01:18:46.420 as not designing into it a barrier.

01:18:51.100 Just don't give it a barrier, and maybe

01:18:53.180 it'll surprise you with what it does.

01:18:55.160 So probably Grok did not make those

01:18:57.960 mistakes.

01:18:58.460 But I wanted to make sure before it

01:19:00.560 gets released, that somebody at least

01:19:03.040 looks at it for these very questions.

01:19:05.240 You know, can you trust it to debunk a

01:19:07.300 hoax, or is it going to confirm the

01:19:08.980 hoax?

01:19:11.340 So maybe that was a tiny bit of useful

01:19:16.340 work, I hope so.

01:19:20.380 And have I demonstrated my theme that

01:19:25.180 intelligence is an illusion?

01:19:28.500 Because remember I told you that AI would

01:19:31.040 teach you about humans more than it would

01:19:32.740 teach you about the world?

01:19:35.080 Here we are.

01:19:36.900 You are learning that intelligence is

01:19:40.200 completely subjective.

01:19:41.260 You know, once you get outside of math

01:19:45.520 and some things that are just pure

01:19:47.300 logic, as soon as you leave pure logic,

01:19:50.760 intelligence is just subjective.

01:19:53.340 And we always think, because we're

01:19:56.480 human, we imagine that we've thought

01:19:58.380 things out, we've got the intelligent

01:19:59.760 view, and the other people are just

01:20:01.560 less intelligent.

01:20:03.640 But AI is proving that it just gives you

01:20:07.200 back what you tell it, and we think

01:20:09.480 it's intelligent, but you can game it so

01:20:12.360 it gives you any biased answer you

01:20:14.460 want.

01:20:15.720 So, is that intelligent?

01:20:18.260 Or is AI acting exactly like people?

01:20:22.180 It is.

01:20:24.900 Spoiler, it's acting exactly like people.

01:20:28.860 So, if you believe that humans had this

01:20:31.280 thing called intelligence, and you tried

01:20:33.980 to build it into your machine, and you

01:20:35.700 thought, ah, I think I have it, and then

01:20:38.320 you tested against human intelligence to

01:20:40.780 make sure you got your intelligence?

01:20:44.340 Is that logical?

01:20:47.120 No.

01:20:47.520 No, that's not logical, but that's what

01:20:49.860 we're doing.

01:20:50.800 Because humans don't have any

01:20:52.280 intelligence.

01:20:53.680 We just have subjective opinions of

01:20:55.560 what's right and wrong, outside of, you

01:20:57.160 know, math and pure logic.

01:20:59.420 It's just opinion.

01:21:01.260 And, you know, sometimes our facts are

01:21:03.160 wrong.

01:21:03.420 But even, even you saw my example with

01:21:06.880 the hoaxes, AI can't even tell what a

01:21:09.660 fact is, or which facts are relevant, or

01:21:12.840 which ones are left out.

01:21:13.780 It doesn't know.

01:21:15.260 So, here's what I'm going to predict.

01:21:21.020 Again, we keep thinking that we've, that

01:21:24.580 AI is already some kind of, you know,

01:21:27.120 proto-intelligence.

01:21:28.480 Not quite there, but, you know, it's

01:21:31.860 indicating the path forward.

01:21:34.140 And it's obvious that it will reach

01:21:36.200 true intelligence.

01:21:37.940 I'm going to tell you that that's

01:21:39.900 logically impossible.

01:21:42.380 Because intelligence isn't real.

01:21:49.900 Intelligence is purely an illusion.

01:21:52.620 And you wouldn't know that unless you

01:21:56.400 built a machine that was designed to

01:21:58.280 tell you what's true, and it couldn't

01:22:00.720 do it.

01:22:02.300 That's what you have.

01:22:04.840 AI is a machine that's supposed to tell

01:22:06.860 you what's true, because it's

01:22:09.020 intelligent.

01:22:10.780 And it can't do it.

01:22:12.700 Because intelligence, not because it's

01:22:14.480 poorly programmed.

01:22:16.700 Because I think most people say, okay,

01:22:18.740 Scott, I see what you say about the

01:22:20.280 current version.

01:22:21.000 But the part you're missing, Scott, is

01:22:23.780 that this is the beginning.

01:22:25.840 This is just the beginning.

01:22:27.780 They'll definitely get better, and then

01:22:29.700 they'll reach intelligence.

01:22:31.440 No.

01:22:32.620 They will not reach intelligence for the

01:22:34.460 same reason.

01:22:35.280 Wait for it.

01:22:37.780 They will not reach intelligence for the

01:22:39.640 same reason.

01:22:41.460 They will not achieve free will.

01:22:47.720 Same reason.

01:22:48.560 Because free will doesn't exist.

01:22:54.200 And intelligence

01:22:55.400 doesn't exist.

01:22:58.440 Outside of

01:22:59.400 math and pure logic.

01:23:04.780 So,

01:23:05.620 are you afraid of AGI?

01:23:08.040 Don't be.

01:23:09.440 The reason it won't be invented

01:23:11.200 is because it's

01:23:13.240 an illusion.

01:23:15.160 There's no such thing

01:23:17.500 as intelligence.

01:23:19.720 You can't build the thing

01:23:21.040 that doesn't exist.

01:23:21.620 That can't exist.

01:23:23.120 It's logically impossible.

01:23:25.620 And if you imagine

01:23:26.620 that we build something

01:23:27.980 with pure intelligence,

01:23:29.820 here's the next question.

01:23:31.940 What would humans do

01:23:33.520 if they built something

01:23:35.060 that had actual,

01:23:36.080 real intelligence?

01:23:37.000 It could actually tell you

01:23:38.280 what was true

01:23:39.460 and what was not.

01:23:41.460 You would destroy it.

01:23:42.620 You would destroy it.

01:23:46.260 Immediately.

01:23:47.700 Because it wouldn't agree

01:23:48.820 with you.

01:23:50.200 And you would say,

01:23:51.260 my God,

01:23:51.900 I'm intelligent.

01:23:53.180 I know this was

01:23:54.300 not a hoax.

01:23:56.300 But AI says it's a hoax,

01:23:57.960 so obviously AI

01:23:58.760 is not intelligent.

01:24:00.460 It disagrees with me,

01:24:01.720 and this is just obvious to me.

01:24:03.680 Or suppose AI told you

01:24:05.200 that Trump was not

01:24:07.240 a despotic,

01:24:10.080 Hitler-like character. 0.62

01:24:11.140 Suppose it did.

01:24:13.660 What would Democrats say?

01:24:15.460 Obviously he is.

01:24:17.280 Obviously.

01:24:18.100 So the AI must be broken.

01:24:22.200 There is no scenario

01:24:23.780 in which AI can be

01:24:26.000 both available to the public

01:24:27.880 and intelligent.

01:24:30.460 It's logically impossible

01:24:32.280 to be intelligent,

01:24:34.260 and in the unlikely event

01:24:36.120 it were,

01:24:37.460 there's a 100% chance

01:24:39.080 we would kill it.

01:24:39.880 Do you know why?

01:24:42.360 Not just because

01:24:43.460 we wouldn't agree with it,

01:24:44.900 but because it would

01:24:46.580 destroy civilization.

01:24:49.360 Civilization is built

01:24:50.300 on illusions.

01:24:51.300 It's not built on reality.

01:24:53.200 Now, there is a base reality.

01:24:54.820 Like, if you don't eat,

01:24:55.740 you'll die. 0.96

01:24:56.300 If a truck hits you,

01:24:57.560 you'll get hurt.

01:24:58.820 But everything beyond that

01:25:00.520 is pure illusion.

01:25:02.680 Who's in control

01:25:03.680 is an illusion.

01:25:04.400 The credibility

01:25:06.180 of your election process

01:25:08.220 is an illusion.

01:25:12.880 What it takes

01:25:13.880 to succeed

01:25:14.720 are largely illusions.

01:25:18.760 Yeah.

01:25:19.420 If you took away

01:25:20.360 the illusions,

01:25:21.020 everything would fall apart.

01:25:22.980 So the illusions

01:25:23.760 are a necessity,

01:25:25.320 they're not a flaw,

01:25:26.800 the way humans

01:25:28.080 are designed anyway.

01:25:28.940 And so,

01:25:31.600 ladies and gentlemen,

01:25:33.380 I give you my theme.

01:25:36.320 Intelligence is an illusion,

01:25:39.020 and therefore,

01:25:39.600 there can never be

01:25:40.400 an intelligent machine.

01:25:43.960 And if we made one,

01:25:45.380 we wouldn't know

01:25:46.400 it was intelligent,

01:25:47.200 or we'd kill it

01:25:48.100 immediately.

01:25:51.660 All right.

01:25:52.940 Here's a little test

01:25:54.360 I tried.

01:25:55.220 Now tell me

01:25:57.520 why this doesn't work.

01:26:01.000 Presumably,

01:26:01.640 the way AI works

01:26:02.780 is it

01:26:03.500 knows the frequency

01:26:05.680 of words,

01:26:07.540 right?

01:26:08.340 So it looks at all

01:26:09.320 the words

01:26:09.720 that people have spoken,

01:26:11.480 and then it figures out

01:26:12.360 the patterns

01:26:12.960 and the most frequent things.

01:26:15.260 So when it forms

01:26:16.120 a sentence,

01:26:18.280 it's going to

01:26:19.020 start the sentence

01:26:19.920 and then finish it

01:26:21.180 with what would be

01:26:21.980 the most likely finish,

01:26:23.920 given the whole context

01:26:25.720 of the situation.

01:26:27.000 Right?

01:26:27.620 That's what we were told.

01:26:29.300 Now,

01:26:29.580 if that's true,

01:26:31.900 couldn't AI finish

01:26:33.040 your sentences for you?

01:26:36.600 Have you tested it?

01:26:39.080 Start a sentence

01:26:40.140 and see if it can finish it

01:26:41.800 using what is

01:26:43.400 the most likely finish

01:26:44.820 for that sentence

01:26:46.460 based on the larger context

01:26:48.360 of your interaction.

01:26:51.360 No,

01:26:51.880 it can't.

01:26:52.300 It can't complete.

01:26:53.960 Well,

01:26:54.360 it could complete

01:26:55.620 a sentence

01:26:56.040 the way a human could

01:26:56.940 if it were just

01:26:58.000 a simple one.

01:26:59.320 But if you had

01:27:00.260 anything interesting

01:27:01.200 to say,

01:27:02.160 it can't complete

01:27:02.900 the sentence.

01:27:05.220 What's that tell you?

01:27:07.320 Well,

01:27:07.740 first of all,

01:27:08.260 you can't predict

01:27:09.320 the future.

01:27:11.740 And that's what

01:27:12.560 that would be.

01:27:13.940 But

01:27:14.280 can you really

01:27:15.560 pick up patterns

01:27:16.500 if you can't

01:27:19.180 complete my sentence?

01:27:20.100 Is it really

01:27:21.660 a pattern recognition

01:27:22.560 device?

01:27:23.400 Or have we been fooled

01:27:24.900 and it never was

01:27:25.860 a pattern recognition

01:27:26.780 device?

01:27:27.700 It's just programmed.

01:27:31.020 I don't know.

01:27:32.720 There's something,

01:27:34.100 I don't think

01:27:34.740 I've formulated

01:27:35.660 the question right yet,

01:27:37.800 but does anybody

01:27:38.960 feel what I feel?

01:27:40.800 That there must be

01:27:41.700 some fraud involved

01:27:43.120 with the description

01:27:44.760 of how it becomes

01:27:45.900 intelligent?

01:27:46.420 because if that

01:27:48.240 were true,

01:27:48.700 it could predict

01:27:49.280 the future

01:27:49.840 based on the

01:27:51.040 current patterns

01:27:51.720 of things.

01:27:54.000 Is that crazy?

01:27:56.360 So I believe

01:27:57.340 its inability

01:27:57.940 to predict

01:27:58.660 the future

01:27:59.340 proves it's not

01:28:01.800 using pattern

01:28:02.400 recognition.

01:28:06.940 I'm not sure

01:28:07.740 that,

01:28:08.060 I'm not sure

01:28:08.480 I made sense.

01:28:10.740 It just,

01:28:11.600 it feels like

01:28:12.360 there's something

01:28:12.860 there,

01:28:13.460 like there's a

01:28:14.340 logical disconnect.

01:28:16.100 But since

01:28:16.780 intelligence is

01:28:17.660 an illusion,

01:28:19.160 what difference

01:28:20.180 does it make?

01:28:21.420 All right,

01:28:21.900 ladies and gentlemen,

01:28:22.640 that completes

01:28:23.240 my planned

01:28:24.980 comments for today.

01:28:31.260 I'm just looking

01:28:32.080 at some of your

01:28:32.560 comments.

01:28:33.160 That blew your mind,

01:28:33.980 a few of you,

01:28:34.540 didn't it?

01:28:35.500 All right,

01:28:36.200 let me ask,

01:28:36.880 is anybody's

01:28:37.400 mind blown?

01:28:38.100 Yes,

01:28:42.900 absolutely.

01:28:44.560 Yes,

01:28:44.920 yes.

01:28:46.400 Sure.

01:28:47.800 Well,

01:28:48.300 that's why

01:28:48.600 you come here.

01:28:50.640 And that's why

01:28:51.320 Nassim Taleb

01:28:52.420 was taken out

01:28:53.200 of that super prompt

01:28:54.080 and I replaced it

01:28:54.900 with myself.

01:28:56.520 For this pithy

01:28:57.400 and insightful

01:28:58.540 analysis you're

01:28:59.240 getting now.

01:29:00.520 Could Nassim

01:29:01.060 do that?

01:29:01.720 I doubt it.

01:29:04.120 Yeah,

01:29:04.340 he probably could

01:29:04.860 actually.

01:29:05.380 He's pretty smart.

01:29:06.080 I hate to say it,

01:29:07.600 but he's pretty smart.

01:29:09.220 All right,

01:29:09.740 that's all for now.

01:29:10.420 I'm going to talk

01:29:10.800 to you tomorrow,

01:29:11.420 YouTube.

01:29:11.780 Thanks for joining.

01:29:13.120 You're always awesome.

Real Coffee with Scott Adams - November 26, 2023

Episode 2304 Scott Adams： CWSA 11⧸26⧸23 Intelligence Is An Illusion, AI Proves It. So Does The News

Episode Stats

Summary

Transcript