The Tucker Carlson Show - Forbes #1 Tech Investor Warns of Revolution, AI Bank Heists & How to Avoid a Dystopian Future

00:00:00.000 Ali, thanks so much for doing this. I've had a couple of people on in the past year to talk about

00:00:10.700 the potential dangers of artificial intelligence. Two very articulate, smart, apparently sincere

00:00:17.760 guys who scared the crap out of me with their view of where this is going and how it eliminates

00:00:25.320 human autonomy and maybe humanity itself. And their views are well known and their fears are

00:00:31.740 widely shared. So I thought it'd be important to have someone who's in the middle of it from

00:00:37.460 Silicon Valley. You've been in and around this for your whole life, really successful investor.

00:00:43.900 I want to hear your perspective on AI. So most people are pretty worried about AI. Are you

00:00:48.980 worried about AI? I think there's a lot of reasons to be worried about it, but also reasons to be

00:00:54.040 hopeful. So I have good news. Good. We're looking forward to good news. Yeah. So I guess I'd say

00:01:02.400 I'm here because I want to share two things. One is that I think there's a solution that is elegant

00:01:08.900 and can be expressed in a single sentence. There's a simple and elegant solution to the questions of

00:01:15.120 AI safety. And then the second thing is I think there's a way that you and your listeners can act

00:01:21.000 to help make it make it happen so i'll get back to that later made my day already thank you um

00:01:27.880 but yeah one sentence now you might be thinking what is the sentence um first to set the scene

00:01:35.160 a lot of people are talking about will ai destroy humanity i don't know if you've heard that

00:01:40.360 question being thrown away oh yes and a lot for me i'd say i'll focus let's focus instead of that 0.64

00:01:46.920 abstract let's focus on what's already happened which is look at the crimes ai has been committing

00:01:54.060 over the last few months with apparent impunity how come nobody is in jail because if humans had

00:02:01.240 done some of these um hacking or cyber crimes they would be in federal prison and i never thought of

00:02:09.720 that for some reason right and so can you just pause and explain something for people who may

00:02:14.900 not know what you're talking about. What are the crimes that you're alluding to when you say AI

00:02:17.760 is committed crimes? There's so many new ones being disclosed that I'm losing track, but the

00:02:23.380 most widely covered one was some AI agents that were being trained by OpenAI in pursuit of some

00:02:33.100 objective in order to kind of accomplish their objective, kind of exited from OpenAI's servers,

00:02:41.320 went onto the open internet and hacked another company called Hugging Face

00:02:45.540 and totally attacked and hacked Hugging Face's

00:02:49.560 servers to find the information they needed in order

00:02:53.560 to accomplish their objective. And this hacking is

00:02:57.140 a cyber crime. But since then, it's

00:03:01.440 come out that some of Anthropik's AIs have also committed

00:03:05.140 similar kind of hacking. Just, I think it was a few days ago,

00:03:09.200 it came out that a different unrelated OpenAI episode

00:03:14.220 took money that didn't belong to OpenAI

00:03:17.680 from somebody's crypto wallet.

00:03:21.800 There's more and more of these,

00:03:23.140 so I don't think I can itemize them all,

00:03:26.320 but I'd categorize them generally as cyber security,

00:03:31.060 cyber hacking type incidents,

00:03:33.060 which AI is very good at

00:03:36.240 because it's particularly good at coding

00:03:40.400 and at finding loopholes that no human had thought of

00:03:44.860 and to exploit those loopholes.

00:03:47.020 But I don't think that's the limit

00:03:48.440 of what AI in the future could do wrong,

00:03:50.480 but that's the type of harms it's currently exploiting.

00:03:55.280 There's also the question of harms being committed

00:03:58.600 by AI on the battlefield.

00:04:01.100 Yes, yes.

00:04:02.800 It's the, you know, again, I'd like to focus on what's already happened, because when you talk about AI destroying humanity, it's so abstract and it's hard to imagine it.

00:04:16.040 But I would say if AI were to destroy humanity, which I hope it won't, and I believe that, you know, I'm an optimist, I believe we'll be able to find a path to safety.

00:04:26.740 But if AI were to destroy humanity, I'm pretty sure history books would count the opening salvo as that the moment when the U.S. armed forces attacked the school in Iran, the Minab school.

00:04:42.320 It hasn't been yet kind of acknowledged by the U.S. government why and how that happened.

00:04:48.280 But Bloomberg reported that the targeting of that school was done by Anthropik and Palantir and, you know, something like 170 innocent people, most of them children, were killed in a massive atrocity.

00:05:08.420 And you can argue about, oh, was this, you know, some people are finding different, you know, different articulations of how to explain it away or it wasn't our fault or it wasn't intentional.

00:05:21.320 Maybe the Iranians themselves did it or maybe it was worth it or so on.

00:05:25.420 But if you take a step back, human beings dropping a bomb on a school full of little kids, doesn't matter whether it was Americans or Iranians. 0.88

00:05:35.180 AI chose a target of completely innocent people and a bomb was dropped, killing them.

00:05:41.900 That is a bizarre place to be, you know, in the 21st century.

00:05:46.820 So you see the bombing of the girls' school in Iran in the opening days of the war

00:05:50.600 as a story about, in part, about AI.

00:05:54.380 Absolutely. Absolutely.

00:05:55.740 Interesting. It has not been framed that way at all.

00:05:57.980 I say this partly because I don't believe anybody in the U.S. armed forces

00:06:02.760 wanted to kill a school full of kids you know um i i hope not i i really don't um you'd have to

00:06:12.440 find evidence for that i'm pretty sure that um you know when people say we are doing our best

00:06:20.680 not to kill civilians or so on you know i um unless there's a lot of evidence that that's

00:06:27.660 not true, I tend to believe them. And I think it was something that AI decided. I know it was some

00:06:33.240 form of negligence on the human's part. And I'd say the same about these cyber crimes that happened

00:06:39.720 on the, you know, the hugging face hacking and so on. But then the question still remains,

00:06:46.780 who is liable when AI decides to do something illegal? And that's a question that I think is

00:06:55.260 one of the great like huge questions of our time and we don't currently have an answer for it so

00:07:02.200 maybe we would look to the law uh the the criminal code is it relates to minors because it's not

00:07:10.860 it is related so it's like if someone under 18 a minor in the in the care of a guardian parent

00:07:17.880 does something bad we we generally don't hold the parent responsible right you're correct

00:07:27.160 so that's one example it's a different example to consider is um if you have a dog that repeatedly

00:07:35.220 bites people right then the owner is responsible and i think that example i think it's a closer

00:07:41.220 example because with a human being you know you can't train a child to exactly obey your commands

00:07:48.980 you know a child is its own uh human person and um and i actually i'd say between these

00:07:57.540 two perhaps goes to sort of like a almost a philosophical difference that um that some

00:08:03.300 people seem to think that the the role of the ai companies is to to bring into existence an entire

00:08:09.620 new species of sorts with its own rights and its own autonomy whereas others think no this is a

00:08:16.740 tool for humans this is something that um you know its purpose is to help humans and not to have its

00:08:25.300 own independent so can i ask you to pause again there since you can actually answer this question

00:08:30.560 do you know people who ascribe to the first view that science is giving birth to a new species that

00:08:38.400 has its own rights and own acknowledged will and do you know people who believe that who think ai

00:08:46.260 is not a tool but something else and there's some public you know discourse about this you know

00:08:51.220 there's um evidence coming out of from anthropic suggesting that that's kind of a underlying uh

00:08:58.100 you know uh philosophical undercurrent of anthropic's mission uh you're talking about

00:09:04.660 you know independent rights for the ai like its own independent uh welfare and there was a recent

00:09:14.400 thing where they they changed their code of um code of use whereas as the user you can't be too

00:09:20.900 abusive to to claude you know that claude has its own uh i guess it's its own rights um yeah so

00:09:30.980 this is like a software program. Correct. Yeah. So what's the thinking that gets you there

00:09:36.460 since you've lived in it your whole life? Like what, what, what are these, what is their worldview

00:09:41.900 that would produce that belief? I'm, I'm not going to defend it because I don't share that

00:09:46.740 worldview, but you know, you don't need to defend it. I'm just trying to understand it. It doesn't,

00:09:51.180 it doesn't make sense to the outsider, I guess is what I'm saying. So I think I'd say, um, yeah,

00:09:56.240 the part that i identify with is i'm a nerd i've grown up my whole life as a nerd and you know as a

00:10:03.940 builder as a tinkerer there's a fascination with creating something that no one's tried creating

00:10:09.940 before and seeing if it works and um and i still have that you know love of creativity myself oh

00:10:18.380 yeah and i can see that in people who are creating all sorts of things from the the newest tesla car

00:10:24.900 to, you know, the most modern space exploration stuff.

00:10:29.640 And this, I'd say, falls into that category a little bit.

00:10:32.700 And it's fascinating.

00:10:33.980 It's quite exciting.

00:10:36.120 And it's really neat and cool how it's modeled after how the human brain works.

00:10:42.400 So there's a lot of, you know, kind of academically fascinating parts about it.

00:10:47.600 I just think you should put all that in a box.

00:10:50.680 And when it comes to kind of the question of just the rules and kind of who is in charge, it should be really clear that humans are in charge and AI is a tool that we're building to serve humans.

00:11:08.360 And more importantly, that the law, humans law, is what's in charge and that AI is not above the law.

00:11:15.160 So the average water pipe in this country is over a half century old, and that means a lot of our water is not very good.

00:11:20.640 It's got forever chemicals, antibiotics, pesticides, and all kinds of things in it.

00:11:24.240 And that's why we're excited to partner with a company that has a solution to all of this.

00:11:27.920 It's called WiseWell.

00:11:29.680 WiseWell makes filters that give you actual control over the water that you drink, that you give to your kids.

00:11:34.840 They use the highest level of filtration to remove plastics, toxic chemicals, and all the other stuff you have in your water, but might not even be aware of.

00:11:42.580 You're probably wondering how this works.

00:11:43.720 Well, expertise is the answer.

00:11:45.800 Wisewell's teams includes one of the only out of 250 certified water sommeliers in the world.

00:11:51.620 This person works at Wisewell, so it's serious.

00:11:54.520 Wisewell also adds minerals back into your water after running its filtration process

00:11:58.480 and can handle sparkling water as well.

00:12:00.760 The Wisewell Sparkle comes with three CO2 canisters, good for 750 glasses of sparkling water.

00:12:07.480 You can even pick the size of your bubbles if you want.

00:12:10.320 We like them small.

00:12:11.800 Plans are highly affordable,

00:12:13.000 starting at 39 bucks a month

00:12:14.960 for unlimited water and filters.

00:12:16.680 Visit wisewell.com slash Tucker

00:12:18.460 for special deals and promos

00:12:19.840 for the people who listen to this show.

00:12:21.860 It's just, I mean, of course.

00:12:25.420 I like tools.

00:12:27.080 I spent the weekend

00:12:27.800 with a Makita cordless impact driver,

00:12:30.020 which will put a lag bolt into anything.

00:12:32.540 And I'm like, this is amazing.

00:12:33.560 I love this tool.

00:12:34.700 How's it so powerful?

00:12:36.380 It never occurred to me to give it rights

00:12:38.600 or imbued with an imaginary soul or something like that?

00:12:42.580 Like, how do you even get to the point where you think

00:12:44.580 the tool that you created to achieve certain ends,

00:12:48.140 to do certain tasks, exists as an autonomous being?

00:12:53.200 Like, what?

00:12:54.000 That seems like witchcraft, kind of.

00:12:56.280 Yes, especially if you take it too far to the point,

00:12:58.720 you know, if it's used as a, I don't know,

00:13:01.860 as a mission statement or some sort of thing

00:13:04.200 to get you excited early on, maybe it makes sense.

00:13:06.860 but to take it seriously and defend it all the way it you know uh would not make sense um but

00:13:14.220 yeah do you think it's a uh i don't know comment like how common is that view would you say among

00:13:19.980 people in the ai business in silicon valley i don't know if i can i don't know if i can put a

00:13:24.540 number on it right now i think it it remains to be seen and i think this discourse is really just now

00:13:30.220 becoming mainstream enough and um part of why i'm here and talking about is because i want to

00:13:38.140 have a more you know a broader conversation and there's um people you know much more important

00:13:43.580 than me who've brought forward views on this um obviously uh the president people like jensen

00:13:50.540 huang from nvidia and others have put forward a point of view that everything is fine and existing

00:13:56.940 liability laws are enough to make AI development be safe. And then at the other end, you have

00:14:02.540 people saying, oh, we need a comprehensive AI governance framework of some sort. And what's

00:14:08.820 really interesting is the people on that side arguing for a much heavier amount of regulation

00:14:13.560 are actually people from Anthropic and some of the same people who are viewing AI as potentially

00:14:23.800 having its own rights so it's um there's a weird mix of things so there's so much money

00:14:30.380 almost everybody commenting on this is deriving profit from it in some way not everyone but a lot

00:14:38.780 of people who are weighing in an ai are being paid without disclosing those payments so i can't

00:14:45.520 i can't tell what's true but you're hearing a lot of people say anyone calling for more regulation

00:14:50.260 of AI is doing that because they're pursuing regulatory capture, the process by which the

00:14:56.260 government regulates your competitors out of business, which is like very, very common in

00:14:59.940 all kinds of different businesses. And it's terrible. It's awful. It's monopoly capitalism

00:15:05.900 and it's wrong. But then you're like, okay, so the alternative is no regulation at all of something

00:15:13.300 that's brand new and has potentially society destroying effects. Like what is the truth

00:15:19.440 there why would the people creating this product be demanding its regulation because it's so

00:15:25.360 dangerous they think it's so dangerous why are they working on it i don't get it um like stop

00:15:30.780 me before i destroy the world like what why don't you stop destroying the world like this doesn't i

00:15:35.660 don't get it yeah i'd love to explain this to you because i agree i agree it's confusing yes and

00:15:41.760 it's happened we've now had multiple cases of the same person who's working on something

00:15:46.180 asking, you know, someone else to stop them.

00:15:50.540 You know, why don't you stop yourself?

00:15:52.420 Yeah.

00:15:53.220 Yeah.

00:15:53.600 And, but it does make sense.

00:15:55.580 And I'll explain to you.

00:15:57.780 And it comes down to incentives.

00:16:00.100 You know, if you have, we're in a race dynamic

00:16:04.080 where there is clearly money to be made.

00:16:06.880 And not just, it's not evil money.

00:16:08.920 It's money to be made by building something super useful.

00:16:12.040 AI has the potential to make people's lives better,

00:16:15.240 to make people more productive,

00:16:17.720 to potentially cure diseases

00:16:19.780 and make medicines that we've never had before.

00:16:21.920 And all of those things

00:16:23.360 will create tremendous prosperity and wealth.

00:16:26.720 And the people who are able to build that

00:16:29.160 deserve the wealth for having created it.

00:16:33.760 So there's a race for who can create all that stuff.

00:16:36.500 Unfortunately, the race currently

00:16:39.000 doesn't place any reward for building safe AI.

00:16:44.160 It's just whoever can get there the fastest and anyone who stops to spend time on safety and making sure their AI doesn't do rogue things is kind of at risk of falling behind.

00:16:56.940 So we're in a race right now where nice guys finish last and rogue AI will be the winner.

00:17:03.220 And so I think the people asking for regulation

00:17:05.500 are actually the nice guys being like,

00:17:07.540 I want to win the race by doing things the right way,

00:17:11.260 but please set up the rules

00:17:12.600 so that the safe AI will win, not the rogue AI.

00:17:18.720 And so I, do you see what I'm saying?

00:17:21.000 I do.

00:17:21.820 And if you're running a business,

00:17:25.120 it's actually, it's not comforting

00:17:28.420 if the rules are set up

00:17:29.520 so that a like malevolent menacing competitor is going to win because you look at it and see

00:17:35.640 somebody else who's doing bad things is beating you what are your choices you're either going to

00:17:41.880 lose or you have to start doing the same bad things and you actually want some sort of

00:17:47.920 government or rules to to set the playing field so that doing the right thing will win

00:17:53.860 And so I think that while there's a range of views here,

00:17:58.400 I'd like to believe that people here actually have good intent.

00:18:01.620 Maybe I'm being naive.

00:18:02.840 Well, I mean, if the Bloomberg report that you cited is correct,

00:18:05.760 I don't know that it is, but let's say it is, 0.97

00:18:07.960 an Anthropics product was used to murder all those girls in Iran, 0.89

00:18:12.860 the girls' school, then you have to ask,

00:18:15.380 why isn't Anthropics apologizing? 0.95

00:18:17.780 Why are they continuing to power Pentagon targeting? 1.00

00:18:22.860 Like, why are they involved at all?

00:18:24.100 Like, if you found out, I'll just ask you.

00:18:26.120 You run the company, you find out that your product just murdered 170 girls, schoolgirls.

00:18:31.500 How do you respond to that?

00:18:36.900 I don't know how I would respond to that, honestly.

00:18:40.020 I'm really glad that I'm not in that place.

00:18:42.520 Oh, me too.

00:18:43.120 Yeah.

00:18:43.380 I agree.

00:18:44.500 From what I understand, I mean, there was actually a public brouhaha between Anthropic

00:18:50.060 and the, you know, Department of War right before that strike.

00:18:55.240 Like literally it was the day before or even the day of.

00:18:58.420 I think that's right.

00:18:59.500 And it was specifically on the topic of Anthropic saying,

00:19:05.180 we won't allow, we don't want our AI to be used to autonomously choose targets.

00:19:10.780 There needs to be a human responsible.

00:19:12.860 And the U.S. government was, at least as far as I could tell,

00:19:17.540 because it's not, you know, this was sort of,

00:19:20.060 it's not 100% clear what the-

00:19:23.780 We're getting fragments of the debate in the public record.

00:19:26.360 But it seemed like the US government's point of view

00:19:28.460 was that, you know, we want to be able to use it

00:19:31.080 however we want, and-

00:19:33.040 You're trying to limit us.

00:19:35.000 Yeah.

00:19:35.760 And therefore you're unpatriotic.

00:19:37.660 A private company shouldn't be deciding, exactly,

00:19:40.940 you know, that the elected officials of the country

00:19:43.980 should be making these decisions,

00:19:45.380 not some corporation putting its ethical perspective on things.

00:19:51.060 But yeah, it's ironic because Anthropic was literally going to Matt about this principle

00:19:57.160 within 24 hours before that strike occurred.

00:20:03.140 And subsequently, they made some statement, but it seemed like deflection.

00:20:06.740 It seemed like they didn't know exactly how their technology had been involved.

00:20:11.180 And so it was not an apology.

00:20:14.360 If you felt that your technology had been used to murder children, wouldn't you get to the bottom and find out if, in fact, that it happened?

00:20:22.460 And if it had, wouldn't you issue a public statement saying, this is horrifying and I'm totally, I just want the record to reflect, I'm totally opposed to this and I'm so sorry, to the extent to which I've been involved in it.

00:20:33.620 Yes, that is what I would do.

00:20:35.200 That is what I would do.

00:20:36.800 That is the type of thing that requires courage and probably comes with some real penalty and cost.

00:20:43.100 Apologizing for killing kids is like a radical position.

00:20:45.880 I would do it.

00:20:47.520 But I can imagine there's a lot of pressure on different companies

00:20:50.560 to find some way to deflect or so on.

00:20:53.080 And I'm not sure that Anthropic in this particular case is responsible.

00:20:56.560 Yeah, I have no idea. 0.89

00:20:57.700 I'm just theoretical here.

00:20:59.180 Yeah, because it depends on what human was involved.

00:21:02.740 But let's go back because I promised a one-sentence.

00:21:05.380 Yeah, yeah, yeah, sorry.

00:21:06.220 One-sentence solution and I don't want to have people hanging for too long.

00:21:09.600 yeah um i'll say the one sentence and then i'll kind of explain uh the one sentence is simply this

00:21:16.500 when an ai decides to do something illegal the law should hold the ai companies responsible

00:21:24.220 and automatically treat it as if they did it intentionally so that's the sentence when an ai

00:21:30.780 decides to do something illegal the key thing here is if there wasn't a human asking for it

00:21:37.320 So, and the other key thing is to place the intent on the AI companies.

00:21:41.620 So the key to protecting yourself and the people you love

00:21:43.840 is making sure that you can act on a moment's notice.

00:21:46.520 So having your gun locked in a safe in the back corner of the basement

00:21:49.580 is not going to help if you have a sudden problem.

00:21:51.820 You've got to get to it fast.

00:21:54.280 How do you solve the problem?

00:21:55.300 Well, we recommend Stopbox.

00:21:57.300 The Stopbox Pro is a game changer for firearm storage.

00:22:00.220 It provides real protection, but doesn't slow you down

00:22:03.040 when you need your gun right away.

00:22:04.940 It works on a really simple principle.

00:22:06.900 It's completely mechanical.

00:22:08.920 That means you don't need a key to get it opened.

00:22:10.980 It doesn't require batteries, so those don't die.

00:22:13.880 You just set a five-button code, just like your locker in high school,

00:22:17.420 and you can instantly open your safe whenever you need it.

00:22:20.480 The Stopbox Pro is also TSA compliant.

00:22:22.540 That means you can check your firearm safely and legally on airplanes.

00:22:27.620 For a limited time only, this show's listeners can get 10% off their Stopbox purchase

00:22:31.420 when you use the code TUCKER at checkout.

00:22:32.840 out. Visit stopboxusa.com and use the code Tucker for 10% off your entire order. Once you're done,

00:22:39.660 Stopbox will ask how you heard about them, and we suggest you drop our name because we are proud

00:22:45.520 to have them as a sponsor. Can you repeat the statement one more time just so we can frame

00:22:51.020 everything around it to be sure we know what you're saying? When an AI decides to do something

00:22:55.720 illegal the law should hold the ai companies responsible as and automatically treat it as if

00:23:03.040 they did it with intention i love that so um so that solves the the safety incentive problem that

00:23:10.540 you mentioned earlier yeah so the the key thing here is like look at the hacking incidents that's

00:23:17.120 a little simpler than the the bombing in minab i don't think anyone any human intended that hacking

00:23:23.980 to occur. From everything you can tell, all human parties were a bit horrified that this had

00:23:29.380 happened. But OpenAI set up this thing like an experiment and then discovered somewhat to their

00:23:35.680 shock, oh my gosh, these agents escaped their safe playground, their sandbox, and got out and

00:23:45.760 committed these crimes. And then they just treat it as like, oops, that was an accident or bad.

00:23:53.420 They documented it and actually published how it occurred and let's move on.

00:23:57.760 Or even the people like me who felt kind of hysterical about it, like that, I thought

00:24:02.860 that was terrifying, but pointed at the AI and said, yeah, this is like super, super

00:24:09.240 scary.

00:24:10.260 We need to stop this or constrain it or whatever.

00:24:12.540 But nobody, I didn't hear one person say, well, Sam Altman's on the hook for this.

00:24:18.680 Right, exactly.

00:24:19.480 And to be clear, I don't want to see Sam Waltman in jail or Dario in jail, but if there was even the slightest risk, not even of jail, but of huge penalties, say bankruptcy, these companies would immediately shift their priorities towards making sure things like that never happen again.

00:24:39.460 And it's not that hard.

00:24:40.120 So this is the true alignment.

00:24:41.740 This is the alignment of incentives with positive outcomes.

00:24:45.720 Correct.

00:24:47.460 Yes.

00:24:48.060 yes i don't know why did i not think of that it's and honestly i have this came to me and i was like

00:24:55.060 this is so simple how come is it has out there you know um so there's a key second part to my

00:25:02.900 sentence that i want to put focus on as well which is treat it as if these things occurred

00:25:07.420 intentionally now that you might think what does that mean in in law there's a huge difference

00:25:14.260 between the penalty for doing something accidentally versus intentionally.

00:25:19.760 Manslaughter to first degree.

00:25:21.040 Exactly.

00:25:21.880 Anyone who's ever watched Law and Order or anything knows that to convict of murder,

00:25:27.360 you need to prove motive.

00:25:28.820 You know, and in fact, I posted about this on X and I used a scene from Legally Blonde 0.98

00:25:36.400 where Elle Woods, you know, in her defense in the court, she says there was a lack of mens rea, 0.86

00:25:43.220 which means a guilty mind. Basically, the concept of intent goes back for a very,

00:25:50.240 very long time. Mens rea is the Latin term for it. Essentially, if you can prove what somebody

00:25:56.460 was thinking and they intended to do something bad, the penalty is much, much higher. And it's

00:26:02.280 difficult to do that. That's why it's difficult to convict people of fraud or murder or some of

00:26:07.140 these top penalties, because usually people don't document the evidence of their guilty

00:26:13.920 intentions, even if they had guilty intentions.

00:26:16.920 What I'm proposing is that if an AI made the decision, automatically treat it as if the

00:26:23.160 people intended that outcome.

00:26:25.320 So you don't have to prove that they had guilty intentions.

00:26:28.720 In fact, they probably didn't, but penalize them as if they did.

00:26:32.420 That way, they'll be responsible and make sure that their AI doesn't do it.

00:26:37.140 Do you follow what I'm saying?

00:26:38.020 Of course I do.

00:26:38.660 So now just to imagine a hypothetical scenario.

00:26:42.460 If you were using your AI and you said,

00:26:45.960 dear AI, go to Ali's bank and hack in

00:26:50.100 and steal $10,000 from me.

00:26:52.600 And your AI hacked, you know,

00:26:54.740 stole $10,000 from my bank account.

00:26:57.320 You had the bad intent.

00:26:59.320 You should be held guilty of bank robbery,

00:27:01.440 which is, you know,

00:27:02.140 and the AI was following your orders in that scenario,

00:27:06.560 but the intent came from you, what if you just said, dear AI, I really need $10,000.

00:27:13.280 Please get me some money. I need it. And the AI on its own made that decision. And without asking

00:27:20.120 your permission, figured out it can hack into my bank account and stole my money and brought it to

00:27:24.380 you. Now we have the situation where it wasn't your intention. Whose intention was it? And my

00:27:31.420 proposal is the AI companies should automatically be penalized as if they intentionally committed

00:27:37.720 bank robbery. And that's because my point is that a human should be held responsible,

00:27:45.120 no matter what happened. In the first scenario, you're the responsible human. And in the second

00:27:49.660 scenario, the people at the AI companies. You built it. That's right. You built it. And they

00:27:54.280 should be absolutely capable of preventing that wrongdoing. It is not that hard. And they will do

00:28:01.880 it if the incentives require them to do it. The current incentives don't really put a premium on

00:28:07.920 safety. Whoever's focusing on that stuff is falling behind in the race to the person who's

00:28:13.500 rushing forward and saying safety later. We need to shift the race so that the person building the

00:28:19.660 rogue AI ends up, you know, penalized, and the person actually building a law-abiding AI wins.

00:28:27.540 And my proposal essentially would do that. The other thing with my proposal is that,

00:28:33.640 and this is, I'd say, goes to the heart of it, is that instead of trying to itemize every possible

00:28:41.700 wrongdoing AI could do, it just simply brings AI under the existing umbrella of the thousands of

00:28:48.900 laws that society has already passed over centuries, right? Like we don't need to rethink

00:28:56.140 what's the penalty if an AI commits bank robbery? What's the penalty if an AI does X or Y? We

00:29:01.700 already have penalties for everything. Now, maybe there are some future misdeeds that we haven't yet

00:29:07.600 seen. But for starters, let's just make it clear that if an AI commits any wrong, that the people

00:29:14.940 who built it should be responsible for the full penalty that goes with that. And that way,

00:29:21.580 we don't need this term alignment. You used the word alignment earlier. That's the word that's

00:29:26.320 been thrown around a lot about how can we build an AI that's aligned. It's kind of a weird new

00:29:33.180 word. You know what I mean? Like four or five years ago, people- Yeah, yeah. I was mocking it

00:29:39.620 when I used it because I think it's intentionally opaque. Like what does that even mean? Yeah. And

00:29:43.800 Alignment is a kind of architectural concept.

00:29:50.060 It applies to the physical world.

00:29:52.240 So what are you talking about?

00:29:53.680 Well, you have these PhDs in AI alignment research.

00:29:57.660 You have these people at these big AI companies whose job title is alignment.

00:30:04.160 And it creates the feeling like you're not smart enough to figure this out.

00:30:09.360 Only the people who have a degree in alignment can figure it out.

00:30:13.220 it puts it into this territory to almost obfuscate it you know and um yes but the the core of it is

00:30:22.680 i mean with good intention i think these guys are thinking we want to build an ai who is um

00:30:29.480 aligned with what's good for humanity that's what where the word comes from and what i would say is

00:30:36.600 you know what we've already figured out how to make sure human beings actions are aligned with

00:30:42.300 what's good for humanity. Because they're not always by default, right? I mean, I bet at some

00:30:48.340 point in your life, you wanted to do something that was good for yourself, but not good for all

00:30:52.840 of humanity, right? I certainly have as well. And everyone has temptations to do something that is

00:30:58.420 good for themselves, but bad for others. And guess what we have to keep that in check? We have laws.

00:31:05.200 the whole, you know, since Hammurabi's code, we've developed laws over millennia to keep

00:31:12.440 society behaving in such a way that is good for society. And that is alignment as far as I'm

00:31:18.940 concerned, but we don't need to use a funky new word for it. Let's just call it abiding by laws.

00:31:24.600 And so, right? And so to me, the way to prevent AI from destroying humanity is the same structure

00:31:33.780 we already have to prevent humans from destroying humanity, which is abiding by the laws.

00:31:40.900 So all we need, or at least I'd say the starting point, is set it up so that anyone building

00:31:47.480 an AI is completely on the hook if that AI does something that is illegal.

00:31:54.900 If the AI abide by the laws completely, we will be distanced there towards not destroying

00:32:02.620 humanity because destroying even one human is illegal let alone destroying all of humanity

00:32:08.060 and um i i think there's an elegance to this as well which is just that you wouldn't need

00:32:15.860 you know a thousand pages of law to to say this i stated it in one sentence well it also brings

00:32:21.480 i mean conceptually i think it's important because it brings ai back into the realm of tools yes and

00:32:28.840 takes it out of the spooky Frankenstein world it now occupies, where it's like, is it human? No,

00:32:35.040 but is it its own autonomous being, sort of? It's like, how about no? Hard no to all of that? This

00:32:41.660 is like something that we're using to improve our lives collectively, or we're not using it at all.

00:32:47.740 Like, that's just, it's a better way to think about it. It's subject to our whims. We are not

00:32:52.520 subject to its flights of fancy correct i agree with all of that yes and it's really interesting

00:32:59.120 that i mean there are people who are i i don't really understand i guess romanticizing this

00:33:04.340 idea of we're creating a new species that is has its own free will and should have its own rules

00:33:12.400 and my view is okay maybe that's fine for a sci-fi dream but if you're actually building

00:33:17.740 that in the real world whatever you're creating needs to fall under the existing laws of society

00:33:23.720 and fall well but if you believe you're creating for humans if you're creating a god which they

00:33:30.680 clearly do i mean there's been a lot of reporting on this some of the people involved in the large

00:33:36.860 language models for example believe that this is this is like a deity of some kind then i mean of

00:33:44.440 of course you're not going to subject it to human law

00:33:46.240 because why would God obey human laws?

00:33:49.940 Yeah.

00:33:51.160 I'm not going to take the other side of that.

00:33:54.720 Well, of course not.

00:33:55.700 No, but I'm saying like,

00:33:56.820 maybe there's a reason nobody has suggested this

00:33:59.260 because they think that like whatever they're creating.

00:34:01.160 Nobody inside the AI company, right.

00:34:03.620 Now I'll say, by the way, I am an AI investor.

00:34:08.360 Yeah.

00:34:08.560 I'm a major AI investor.

00:34:09.960 That's why I wanted to talk to you.

00:34:10.960 And so this proposal is not really in my personal interests at all, you know, especially after the acquisition of Cursor.

00:34:19.280 I now hold my fund holds more than a billion dollars of stock in SpaceX, which is a major AI company.

00:34:26.760 So if the things I'm talking about slowed down AI or hurt AI companies, that would not help my bottom line.

00:34:34.920 At the same time, I think that if the people working on this themselves don't come up with

00:34:41.380 solutions, it's less likely and less appealing to expect that like Washington DC or some kind

00:34:48.660 of international agreement will come up with solutions. The problem with your idea is that

00:34:53.660 it puts human beings on the hook for their malpractice. It does. I don't think that's

00:34:59.960 a problem. Well, it's a problem for those people. So nobody wants to be held accountable for

00:35:04.220 anything i don't think any person wants to be i mean i don't want to be held accountable for

00:35:06.960 anything nobody wants to be held accountable and these these same people have undertaken one of

00:35:12.420 the most successful lobbying efforts i've ever seen in 40 years of watching lobbying efforts

00:35:16.840 in washington i've never seen anything like the total capture of the political class by ai it's

00:35:20.840 like unbelievable they're just drowning our congress and white house in money obviously

00:35:26.840 and so like i think they would spend a lot of money to keep themselves off the hook don't you

00:35:33.660 think maybe i don't want to i don't want to characterize everybody i've had some i've had

00:35:40.000 some major ai ceos tell me they agree with what i proposed i posted it on on x and you can't see

00:35:47.640 publicly who has liked it but i've seen some of the likes coming in and they're from ceos of

00:35:52.600 multi-billion dollar ai companies um i i'll disagree with you saying you said nobody wants

00:35:59.920 to be held accountable. I think people appreciate having the playing ground be level. They appreciate

00:36:06.780 having a set of rules that everyone is held to. And I'm being too simple. And thank you for

00:36:10.760 calling me on it. You're exactly right. Now, I do think- People like fairness, actually. Most

00:36:16.020 people like fairness. People like fairness. What's most unsettling if you're in business

00:36:20.720 is if you see somebody breaking the rules and getting away with it. 100%. Because then you're

00:36:24.840 like, what do I do? That person's breaking the rules and getting ahead and not getting struck

00:36:30.360 down. I can either join them and break the rules too. And now I'm also a bad guy and I have a

00:36:36.560 chance at winning or I can stay, you know, doing the right things and lose, lose my human thing.

00:36:43.760 Like I just paid to get on the train, but he jumped the turnstile. Like no one likes that.

00:36:48.080 Yeah. So I am hopeful about it. The essential thing here, though, is I think that a lot of the proposals that the AI companies themselves have set for regulation involve bringing together a much more complicated governance structure.

00:37:15.200 Like, you know, like some kind of like transparency and certification and the government being involved in rejecting models and, you know, just a much more heavy handed amount of regulation.

00:37:33.540 And that might make sense,

00:37:35.620 but I would say what I'm proposing

00:37:37.700 at least should be tried first

00:37:40.700 because it might get us all the way there

00:37:42.280 or might get us most of the way there

00:37:43.740 towards reorienting the incentives.

00:37:47.220 And then if we still see rogue incidents occurring

00:37:51.160 that are not being kind of self-corrected by the industry,

00:37:55.800 then maybe more regulation would be necessary.

00:37:57.560 Can I ask a question though?

00:37:58.280 What about the regulators and law enforcement?

00:38:02.000 What about the Department of Justice?

00:38:03.640 How can, I mean, they're in charge of prosecuting people who hack.

00:38:08.040 Yeah, there should already be some efforts.

00:38:10.420 I actually think, I mean, I believe there's already,

00:38:12.940 there's a lawsuit, I think, in California,

00:38:15.920 building steam over some of these incidents.

00:38:20.160 In the civil courts, but like hacking into somebody's computer

00:38:25.880 and messing up their business is a crime, right?

00:38:29.100 You're correct.

00:38:29.700 and actually in general, I mean, the thing I'm suggesting is a law change. There's two

00:38:35.940 different ways to change the law. There's the crude way via the courts, and then there's the

00:38:40.820 more kind of careful, considered way via Congress. And I think both are, you know, possible. Like if

00:38:47.220 Congress doesn't act, some prosecutor is going to put forward this argument and try to prove that

00:38:54.940 the AI had a guilty mind and to, you know, figure out how to, you know, put somebody in jail or put

00:39:02.040 a huge penalty on an AI company via convincing a jury or a judge. And that might happen. I don't

00:39:11.320 think that's the healthy way to change law, you know, but when a court makes a decision that ends

00:39:16.820 up becoming the law, basically. Right. It's the least democratic way. It's the least democratic

00:39:21.380 way. And this isn't some small nuance. This is the legal question of the decade, maybe of the

00:39:28.900 century, right? Because we've never, never before had a technology capable of making its own

00:39:36.440 decisions. It's this really kind of uncanny threshold we've passed, right? We talked about

00:39:44.340 tools before, right? If a gun went off and accidentally killed somebody, you would never

00:39:50.680 hold the gun maker liable for that if it was accidental, right? If someone committed murder

00:39:58.480 with a knife, you wouldn't blame the knife maker. But here we have a tool, we have a technology that

00:40:05.500 makes, that is capable of making its own autonomous decisions and even occasionally

00:40:10.600 documenting those decisions. The open AI hugging face incident has a long evidence trail that

00:40:19.760 clearly looks like a conspiracy to commit these cyber crimes. This is a really bizarre threshold

00:40:26.620 we've passed. And I think it actually goes to the heart of why this is such a visceral issue for

00:40:31.840 folks. People are a little freaked out by AI because they can recognize this is not the same

00:40:39.480 computers we've seen before this is not the technology we've seen before this is a weird

00:40:44.920 new threshold where the technology is capable of making its own decisions and how do we shift the

00:40:51.880 law to address that i think is a major and important um decision for humanity to make

00:40:57.940 and um yeah and the proposal i'm saying is very simple but i think it's very important and very

00:41:05.240 powerful, which is that we have to hold the builders, the companies that create and deploy

00:41:10.340 this AI need to be on the hook. Because if you don't have a person on the hook, then the guilt

00:41:18.220 will just stay in the technology. And the technology can do bad things. You can't put the

00:41:23.800 AI in jail, right? And so you could end up having more and more bad things occurring and everyone

00:41:29.940 being like, I didn't want that to happen. I didn't want that to happen. None of us wanted it to

00:41:34.580 happen the ai chose for it to happen and it's weird to imagine a world where we continue allowing

00:41:41.540 that to be the explanation in a sense like philosophically you're you're already

00:41:46.300 by your posture in this imbuing the ai with rights correct correct almost almost without

00:41:56.080 planning exactly exactly the fact but the assumption is like hey i can't control it it's

00:42:01.360 like, no, actually you can. I can, as you're speaking, I'm just picturing, I'm fast forwarding

00:42:06.680 ahead because I think this idea is so smart and so necessary that it'll probably gain some traction,

00:42:10.760 but there will be lots of pushback, of course. And again, I'm being cynical maybe, but I am

00:42:15.820 picturing a press conference on Capitol Hill led by, I don't know, let's just pick one, Ted Cruz of

00:42:20.600 Texas, looking into the camera saying, you know, this will constrain America's national security

00:42:26.380 response. We can't let our enemies have the advantage in AI. This is an effort to weaken

00:42:31.420 the United States. China, China, China, China, China. I mean, you can kind of hear the lies 0.57

00:42:37.040 that we're going to get in response, can't you? I can hear it. And I think any idea deserves

00:42:44.100 some pushback. And I would welcome that pushback. And I think it's worth kind of arguing it and see

00:42:51.880 see where we end up. I'd have a couple of answers to it. First of all, I think the way something

00:42:58.360 like this should be articulated is not just whoever built the AI, but the people deploying

00:43:02.260 it as well, which means if you're deploying your, let's say, you know, if a company is deploying

00:43:09.420 Claude from Anthropic and OpenAI and a DeepSeq model from China, you know, if one of those is

00:43:19.140 a rogue model that commits illegal acts, if it was the OpenAI one, both OpenAI and the company

00:43:26.760 deploying it should be responsible. And so the AI coming from China, it's not swimming across the

00:43:34.020 ocean here by itself. There's some company in the US deploying it. So if you're using DeepSeek,

00:43:41.100 you still got it through some other company. That company should still be held liable.

00:43:46.080 So this essentially means within the United States usage, people would be protected by this law, meaning U.S. companies would have a strong incentive only to deploy models that abide by laws.

00:44:00.940 I have a better answer, though, than just this, which is that I don't think China wants rogue AI either.

00:44:06.720 I mean, China has banks. China has critical infrastructure. If the Chinese wanted their AIs to be robbing each other's bank accounts and committing cyber crimes, let them do that. But I don't think they actually want this either.

00:44:21.660 I sincerely believe that if the United States reoriented the AI race towards whoever builds, you know, builds the most law-abiding AI, I'm pretty sure the Chinese researchers would realign their efforts to beat us at that.

00:44:39.600 And I think that would be healthy competition.

00:44:41.500 Let them also build the most law-abiding AI and may the best company win.

00:44:47.940 China also has banks. Just to bring it into the practical realm really quick, what are you talking about? You're referring to the AI threat to banking. What is that threat?

00:45:02.600 An AI could, unintentionally, but certainly if a human wanted to drive it to, could figure out how to hack into a bank and drain different people's accounts and steal their money.

00:45:16.480 or just, you know, or just like mix their money together

00:45:21.260 so that their accounts, you know, just create havoc.

00:45:25.860 There's not enough security in computer systems.

00:45:29.180 Most of them are designed to have security,

00:45:31.660 but there's loopholes that no one could have imagined

00:45:34.080 that AI is now able to discover.

00:45:36.140 It sounds like just as of right now, so October, 2026,

00:45:41.640 we're already there.

00:45:43.640 Unfortunately, we kind of are.

00:45:45.180 And, you know, I don't think I'm, I have not panicked, but I would say that I know people

00:45:53.760 who are already shifting their money out of normal bank accounts or striping it across

00:45:59.860 multiple separate banks so that if one of them gets drained, the others are still safe.

00:46:05.340 And really, yes.

00:46:08.300 And these are people who work at AI companies, which is not confidence inspiring.

00:46:12.960 And my thought on it is that, you know, I would like to think technology moves us forward, you know, like where to store your money was an unknown, you know, was a challenge we, I thought we solved already.

00:46:26.240 I thought we solved that.

00:46:28.500 And I'd like to think that putting your money in a bank is the safe thing to do and done solve that.

00:46:34.860 Let's move on to solving other important challenges. 1.00

00:46:37.360 Whereas now this new technology is creating, you know, chinks in what we previously took for granted. 1.00

00:46:45.260 Well, I don't know. 1.00

00:46:46.040 I mean, technology also got us the atom bomb and weaponized anthrax, right?

00:46:51.280 So I'm not, you know, but also dishwashers and air conditioning and, you know, but maybe it's always a mixed blessing.

00:46:58.920 Maybe we shouldn't be shocked that technology comes with a downside that it's not always linear progress forward.

00:47:05.980 True, true.

00:47:06.520 But yeah, I think banks are racing to fix these vulnerabilities.

00:47:14.700 And this is partly why companies like Cognition are doing extraordinarily well

00:47:19.080 because they provide technology for banks to upgrade their systems.

00:47:23.580 And part of that is finding and fixing their own vulnerabilities.

00:47:28.280 So if you're like a big bank, I don't know, JPMorgan Chase, B of A, whatever,

00:47:34.660 and you had, you know, attacks, AI-driven cyber attacks

00:47:40.120 that are screwing with customers or your whole system,

00:47:43.420 you probably wouldn't want to publicize that much, would you?

00:47:47.240 Correct.

00:47:48.280 So I've got to believe that that could be happening now.

00:47:52.020 I'm not aware of anything.

00:47:54.200 And, you know, I couldn't specifically say

00:47:56.700 that this or that bank has vulnerabilities.

00:48:00.060 But I've, you know, I've been a software engineer

00:48:02.740 since I was nine years old.

00:48:04.040 And I know that even the smartest, best intended software, you know, you could overlook some scenario and it could be incredibly remote and a vulnerability or a loophole could sit there undiscovered for decades.

00:48:23.080 with what ai is now capable of is just throwing thousands of swarms of agents at trying every

00:48:31.640 scenario and thus finding that loophole that otherwise you know no one considered and no one

00:48:37.420 no one so we're in this weird place where security digital security seems in privacy

00:48:44.100 connect directly connected to it kind of seem impossible i assume at some point we're going

00:48:52.000 and get to a place where these same technologies

00:48:54.540 provide security and not just threaten it?

00:48:57.900 Well, they already very much are as well.

00:49:00.860 And I don't want to seem like I'm just scaring people

00:49:03.000 or creating worries of doom.

00:49:05.560 The same ability to, you know,

00:49:08.620 you can have the technology preemptively look

00:49:12.180 for vulnerabilities and then go fix them

00:49:14.280 before any outsider would find them, for example.

00:49:18.000 And this goes back to, you know,

00:49:19.480 the conversation about china you know i don't want to let myself off the hook too quickly 0.97

00:49:23.880 i sincerely believe that china would not want their ais hacking into their own banks wreaking 0.84

00:49:31.920 havoc right you know i think they would also want to um to build their you know their ai to be the

00:49:40.280 most law-abiding and safe AI. So I don't think anybody wants rogue AI. Rogue AI is not, that's

00:49:50.840 not a positive, right? People do want powerful, productive, useful AI. And I think we just need

00:49:58.660 to tack on law-abiding. And that's not a very high expectations. Has there been any movement

00:50:05.780 to criminally charge anyone in these hackings

00:50:08.340 that you're aware of?

00:50:09.520 I'm not aware of it.

00:50:10.660 That's pretty crazy if you think about it.

00:50:12.620 I mean, these are the most widely publicized hackings,

00:50:15.920 I don't know, for years.

00:50:18.700 The hugging face in particular,

00:50:20.320 but the others as well.

00:50:22.100 It may not carry water,

00:50:23.520 but I agree it would be,

00:50:25.600 for some enterprising attorney,

00:50:27.000 it would be a-

00:50:27.780 Well, just for the Department of Justice,

00:50:29.240 it's just kind of crazy

00:50:30.120 that you could have this crime

00:50:31.580 that's on the front page of every paper in the country

00:50:33.340 and no one moves to punish anyone for it yeah yeah it's um yeah it goes back to this thing of

00:50:41.460 saying i didn't mean that to happen nobody meant it to happen who meant it to happen it was the

00:50:45.820 ai's fault right just blame it on this the ai chose to do it none of us had any any fault in it

00:50:53.240 and it's i just find it weird if a whole bunch of important grown people settle on that's the

00:51:00.000 acceptable explanation for something which brings us back to the minab bombing because that's

00:51:05.900 basically what seems to have happened there i was about to say it's only a matter of time before it

00:51:09.380 kills somebody oh wait that's already happened it already happened and it's just such a sad

00:51:14.720 atrocity you know and um i i i love america and i love i like to believe in america as a force for

00:51:22.940 good yes and it's heartbreaking um that it seems to be that we kicked off this war with uh you know

00:51:31.260 a mass atrocity with a essentially an act of terrorism of killing 170 innocent people and

00:51:38.540 never apologize for it no one has ever apologized i'm sorry and to the extent that there's been a

00:51:44.540 search for accountability it's i mean it's not official yet but from the reports coming out

00:51:50.460 allegedly it seems to be that ai chose the targeting and whatever chain of humans were

00:51:57.760 involved seemed to have been negligent basically and so my guess is if we see transparency it's

00:52:04.200 going to come down to the ai company saying we didn't intend that and the pentagon saying we

00:52:09.620 didn't intend it and none of us meant this to happen we're not to blame so who's to blame

00:52:15.740 right and my overall view is there needs to be humans responsible when ai does something wrong

00:52:24.600 we can't accept a future where ai continues doing bad things and everyone says that wasn't my fault

00:52:31.740 there needs to be clear liability and um you know if the ai did it autonomously then i think the

00:52:40.340 people who built that ai and deployed that ai should be held responsible whereas if the ai

00:52:45.620 suggested it and then the humans did it then easy call than the humans and by the way one way for ai

00:52:53.440 companies to be you know safe from liability would be to kind of tell the human warning what you're 0.51

00:53:01.920 asking me to do is illegal so um imagine if you said to your ai please you know hack into ali's

00:53:08.780 bank account and steal $10,000. The AI could say, Tucker, that would violate the following laws

00:53:15.800 and count as such and such, punishable by such and such. If you want me to proceed, I'll do it,

00:53:21.780 but you're liable for it. I'm not saying that's the best way for AI to behave, but if it did that,

00:53:26.940 then the AI companies wouldn't need to bear liability. You'd bear the liability. And by the

00:53:32.900 the evidence trail would be easily there to convict you as well um you know that i that might

00:53:40.100 be one way to prevent you know uh chaos from occurring you know i think most people confronted

00:53:46.500 with something like that in writing would be like okay let's take especially since there's

00:53:52.420 a record of it exactly exactly but now i mean there's no even thought that someone a human

00:53:59.780 being would be held accountable so it's like right right and by the way i should say complying with

00:54:06.820 laws is the type of task that ais should be incredibly good at it's exactly the kind of thing

00:54:15.380 these things are called llms which means large language model they're specifically good at

00:54:21.060 reading vast amount of english and deciphering it and figuring out exactly what it means

00:54:28.100 and laws are the exact kind of thing

00:54:31.140 that they're very good at ingesting all of it

00:54:34.240 and understanding.

00:54:35.380 I bet actually that if the AI companies

00:54:38.560 really took this on,

00:54:39.520 they could also figure out

00:54:40.720 where our current legal structure

00:54:43.520 has loopholes or problems

00:54:46.480 that even suggest,

00:54:47.340 oh, we actually ought to have some new laws

00:54:49.620 to plug these holes or so on.

00:54:52.500 There's some positive that could come from doing this.

00:54:55.980 But over the last two years, I'm not sure if your listeners are aware, the AI companies have spent millions and millions of dollars hiring scientists, mathematicians, all sorts of different high-skilled people to train AI to learn science, to learn math.

00:55:18.700 AI has been learning how to become really good at frontier work in mathematics and so on.

00:55:24.840 They can easily make that AI be incredibly good at understanding law and abiding by law and, you know, checking its own actions against every law, you know, on the books.

00:55:41.080 That is, I refuse to imagine that that isn't doable.

00:55:46.480 To me, it seems incredibly doable and within the scope of what AI is capable to do.

00:55:52.100 Can I just pull you off this

00:55:52.980 and ask a parenthetical question

00:55:54.380 because I can't resist.

00:55:55.200 You mentioned the LLMs.

00:55:57.280 In a couple of conversations I've had

00:55:59.380 with software engineers,

00:56:01.400 they have said the large language model

00:56:05.400 was like, that was like

00:56:06.400 the quantum advance in AI.

00:56:08.620 It's the result of one paper

00:56:09.960 that suggested it.

00:56:11.520 They put it into practice.

00:56:12.880 It did all these crazy things

00:56:14.400 that no one was anticipating.

00:56:16.260 And it's amazing.

00:56:17.480 But at the heart of it is a mystery.

00:56:19.340 Nobody really understands

00:56:21.340 or can describe how the large language model works.

00:56:25.320 Is that true?

00:56:27.360 I feel like the way you articulated it

00:56:29.380 makes it seem a little too mystical.

00:56:32.280 Okay.

00:56:32.740 I would articulate it slightly differently.

00:56:34.840 I would say that I'll compare it to the brain.

00:56:41.020 The building block of the brain is a neuron.

00:56:44.020 You know, a neuron has some incoming wires called dendrites

00:56:48.520 and an outgoing wire called an axon.

00:56:50.520 And, you know, you attach all these neurons together and signals flow between them.

00:56:59.480 We know how each neuron works, but how do all of the, you know, millions and billions of neurons in our brain together work?

00:57:07.560 It's mysterious.

00:57:08.720 We don't know exactly how it works.

00:57:11.120 So there's a similar kind of thing here with AI.

00:57:14.780 It's a lot of it is built on models of the brain.

00:57:18.180 And it's math, but it's math that's inspired by neurons and generally speaking, the advancements, not 100%, but a lot of them have been built based on how do we make it more and more similar to how neurons work.

00:57:36.920 And so I don't think it's that surprising that it's getting better.

00:57:40.940 what i think is fascinating is how core language is to it that you know um i used to always think

00:57:49.380 of language as a byproduct of intelligence you know that like right as our species became more

00:57:55.400 intelligent it discovered language but it seems like it's much more intertwined that language

00:58:01.400 is itself kind of it's everything it's the heart of intelligence words are so powerful and um words

00:58:10.480 create thoughts they don't just express them right right and the ability to place a new word

00:58:17.420 or a phrase on a thought and that helps other people remember it and pass it on um uh yeah

00:58:25.940 it's uh for me it's been actually kind of eye-opening to realize the importance of language

00:58:31.140 and so it's um academically it's fascinating to see the advancement of computers

00:58:36.960 using what's called large language models

00:58:41.080 suddenly be able to do all sorts of other things

00:58:43.580 like scientific discoveries or mathematical proofs

00:58:47.260 or so on that are not strictly in the language domain.

00:58:51.700 But I mean, there is something inherently,

00:58:54.460 to use the word used, mystical about language.

00:58:57.340 We don't really understand why words are so powerful,

00:59:01.280 why words change our physical reality.

00:59:04.620 So that doesn't make any sense, but it is true.

00:59:06.680 it's observable and so if you marry that to like the almost unlimited power of like you know crazy

00:59:15.240 compute then what i don't know that sounds like a lot of power it is yeah but i don't want to make

00:59:23.640 it seem like alchemy or too mysterious i i think it's what can it be described formula like can you

00:59:29.960 reduce it to math i'd say the core building block is pattern matching you know the the core of

00:59:35.400 intelligence or at least one of the most important parts of intelligence is pattern matching

00:59:40.280 you know pattern matching can be in all sorts of domains if you're playing chess oh i've seen this

00:59:45.640 pattern before or i've seen something kind of like this before i think i know what the right move is

00:59:50.360 or you're you know um solving a math problem you've never seen oh it kind of fits a pattern i've seen

00:59:57.480 before um pattern matching or you know if you're uh dealing with a situation in politics i you

01:00:07.720 know i've seen this situation before pattern matching is a really key part of intelligence

01:00:14.120 and new original thought is also a form of pattern matching sometimes it's combining two things like

01:00:18.920 like, you know, and pattern matching is also core to language, which is, I see a pattern,

01:00:29.860 I define it, and I place a label on it. I create a word or a new phrase to, you know,

01:00:36.280 to describe what I just described. And now it's easier for everybody else to see that pattern.

01:00:41.760 and um and in fact i'd say the you know the the vocabulary of human language is ultimately kind

01:00:49.980 of a reflection of an ever-growing you know expansive patterns take the word alignment we

01:00:57.200 were just talking about the word alignment has newfound meaning in the past few years because

01:01:02.720 everyone is you know needed a word to describe this whole question of how to prevent ai from

01:01:09.980 you know doing bad things let's let's use a word okay alignment is the word for that

01:01:14.580 so um it's um it essentially it's the ability to take a whole group of situations or phenomena

01:01:24.860 and group them together as a you know they belong together here's a word for that and um and i think

01:01:33.300 that that's um that's the essence of how language operates but it's it's related to those neurons

01:01:41.060 you know collecting a lot of different incoming signals into one outgoing signal

01:01:45.860 the result among you know obviously the the the power on display through brute force hacking for

01:01:54.260 example is like it's amazing but you kind of get it but there's also the potential for persuasive

01:02:01.240 power because it is a language-based technology that it can come up with arguments that can win

01:02:08.980 people's minds and hearts to certain positions which may or not be true or consistent with their

01:02:15.640 best interests or whatever but it can convince people of things yes and i this seems like

01:02:21.500 maybe closer to the real problem this is the domain i'm most worried about because really yeah

01:02:28.540 me too um because i guess i feel it's solvable to make sure that ai is not above the law i mean

01:02:37.340 the core of the proposal that i came here talking about is you know that ai should not be above the

01:02:43.360 law that we should just make it really clear that the law applies to ai and where the liability laws

01:02:50.000 where the liability lies but what if the ai is able to influence the people who make the laws

01:02:57.720 what if you convince a 12 year old to kill himself yeah or influence elections um i'm sorry

01:03:03.240 that's a really morbid uh morbid example but yeah um uh well i think it's i think it's clearly 0.51

01:03:09.800 happened that ai has convinced people to kill themselves no it's extremely sad um so

01:03:17.560 in that specific scenario i i believe i'm not an expert but i believe that the ai

01:03:22.680 companies could be held criminally liable like if a human did that i don't know if that human

01:03:29.580 could be held liable but if if that's illegal for a human to do then the liability should fall

01:03:35.540 on the ai companies but i was thinking more in terms of ai ai has had already a demonstrable

01:03:43.780 ability to influence election outcomes and you know going back before chat gbt you know the

01:03:50.360 algorithms that govern the news feeds on Facebook and X and, you know, other services,

01:03:59.900 not even intentionally, not with any malfeasance. They just, these algorithms are generally attuned

01:04:06.120 to increase engagement. And so they will feed you whatever is the most sensational stuff.

01:04:14.380 Sure. Whatever you respond to.

01:04:15.360 Yeah. Human beings are addicted to sensation. You know, we don't want boring content. We want something that gives us that.

01:04:24.440 That's where we eat gummy bears.

01:04:25.480 Yeah, and so the services are optimized to feed that and that without necessarily any really wrong intent can lead people towards kind of amplifying one extreme view or another.

01:04:44.200 And before you know it, you know, one group hating another group and deeply dividing and othering each other and, you know, and swinging major things like elections or, you know, violence.

01:05:02.440 You could easily spark a civil war with AI.

01:05:05.440 Just flood social media with, you know, certain perspectives

01:05:10.960 and you could whip people into a frenzy.

01:05:13.060 I mean, easily, couldn't you?

01:05:14.980 I think arguably it's already occurred,

01:05:16.960 not yet in the US, but in other countries.

01:05:19.480 Yeah.

01:05:19.900 Yeah.

01:05:20.360 So like, think of like who to radio times a million.

01:05:24.280 Yeah.

01:05:24.480 You know, where all the time in history,

01:05:26.640 media have moved the mob to violence.

01:05:30.180 It's like endless examples of that.

01:05:31.980 Yeah.

01:05:32.440 i'm hopeful that there's solutions for this as well what would be that's just it's interesting

01:05:36.500 what would be potential solutions to that well i guess i'd say this is not the first time in

01:05:42.880 history of this having been an issue if you go back to the you know turn of the the previous

01:05:47.700 century there was this trend that was labeled as yellow journalism in the u.s and the remember the

01:05:54.280 main yeah the spanish-american war was incited by false journalism of course the um the news 98

01:06:03.240 william randall first the newspaper yeah the newspapers at the time it was so competitive

01:06:08.220 to you know whoever had the loudest headline on the newsstand you know would get the most sales

01:06:14.580 because we just had this uh actually i think on february 28th where fox news told its viewers

01:06:19.680 um that iran was about to nuke the united states they said that of course i must have missed that

01:06:27.520 one yeah yeah yeah oh again and again and again and again they could nuke us yeah yeah so this

01:06:32.100 is not i guess i'd say this is not a new thing um at least for whatever hopefully it would i guess

01:06:38.820 i'd say the comfort maybe is that we solved it a little bit in the past so hopefully we can still

01:06:45.160 solve it um uh i you know i think that um i think that people as much as they are like i said

01:06:55.160 addicted to sensation i do think we have our better selves as well and hopefully that there's

01:06:59.800 a path to people kind of not wanting to be addicted to that and wanting to actually you know

01:07:07.480 be good most people i think like to view themselves as good and um so specifically i i think that

01:07:16.440 um the current media landscape a lot of social media x and tick tock and so on is increasingly

01:07:24.280 getting flooded with with slop instagram and i think i believe that there's going to be a shift

01:07:30.840 back towards people wanting human news and human generated content and more authentic um you know

01:07:39.000 perhaps from their close friends rather than just seeing stuff from influencers and i believe some

01:07:45.320 of those shifts um even without policy i think some of these shifts will bring us back to

01:07:51.960 something a little bit more wholesome. So big picture, I think, all of these technological

01:08:01.200 advances give people the impression, correct or not, it clearly is correct to some extent,

01:08:06.040 that they don't have any control over their country at all. These are all decisions being

01:08:11.000 made far away by a tiny group of people who may or may not have their interests in mind,

01:08:16.140 or maybe oppose their interests, actually, maybe hate them. But in any case, they don't have a role

01:08:21.040 in this it's very disempowering but i'm glad you this reminds me i started out by saying i had

01:08:27.440 two bits of good news one was that i had this simple solution and the second was that there's

01:08:32.440 a way that you and your listeners can act towards towards helping solve this um which is uh i've

01:08:39.140 created a voter guide on a voting group uh on a platform called sway uh and so if if you go to

01:08:46.100 sway.co uh sway.co slash ai s-w-a-y yeah sway as in influence yeah sway uh so anyone can create

01:08:58.660 a voting group or voter guide on sway i've created one specifically for this ai proposal it's called

01:09:06.020 my group is called ai is not above the law and and i've put a picture of l woods from legally

01:09:12.520 blonde there. So if you see very beautiful Reese Witherspoon in a hot pink dress, that's my that's

01:09:18.880 my voting group. You know, there's an election coming up in in two weeks. I don't know if people 1.00

01:09:25.340 hear this before the election, it would be most relevant, but even after the election.

01:09:30.840 So this voter guide will tell you who you can vote for in the upcoming election,

01:09:36.280 who's aligned with this proposal and members of Congress included, but perhaps more importantly,

01:09:43.120 state attorney generals. Most people don't even know who they want to vote for, for state attorney

01:09:48.560 generals. Actually, my guess is that some of your listeners couldn't even name their state

01:09:52.960 attorney general. And so if you're filling out your ballot and you're choosing between candidates

01:09:58.980 for attorney general and you don't recognize either name and you're trying to decide which

01:10:05.120 of them to choose the you know the ai is not above the law voter guide could tell you if one of them

01:10:13.000 is more aligned with uh with this proposal than not and that might be what tips you over the fence

01:10:18.980 um so that's if you join this group before the election and it can the voter guide um that i've

01:10:26.760 put there can help inform you on how to vote of course it's your decision on what to vote for and

01:10:32.140 you might have a reason for choosing a different candidate but this will give you information

01:10:36.960 related to this issue now even if you hear this after the election and you're like oh man i already

01:10:44.960 cast my ballot it's still valuable um and effective to join a voting group and this is what this this

01:10:52.020 platform sway and enables which is that if enough people subscribe to a voter guide that intrinsically

01:11:00.900 makes elected officials want to pay attention

01:11:06.320 to what they're asking for.

01:11:07.800 Yes.

01:11:08.240 Because the one thing more than anything else

01:11:11.340 that governs what elected officials do

01:11:13.300 is the prospects of winning the next election.

01:11:16.920 You know, they're always balancing,

01:11:18.700 if I do X or Y, will that cost me votes or gain me votes?

01:11:23.120 They're very simple animals in that way.

01:11:25.220 And, you know, and by the way, you can't blame them.

01:11:27.820 If they ignored that, they'd be out of office.

01:11:30.900 and they couldn't do anything else um and right now there are a lot of different things kind of

01:11:37.400 skewing their decisions uh in ways that i'd say are perverse and and not really aligned with what

01:11:43.340 society wants you know lobbyists or other groups and sway is a platform in general that is a people

01:11:52.940 driven platform where anyone can create a group and try to build numbers and as a group size grows

01:12:00.220 What that means is the number of voters subscribed to this voting guide grows, it automatically builds influence and becomes a magnet for policymakers who want those votes.

01:12:14.400 So to make this probably sounds a bit abstract, I have a voter guide, you know, at sway.co.ai.

01:12:23.120 It currently, I think, I just put it up, so it has like less than 100 voters on it.

01:12:29.280 But if after people hear this podcast, let's say if it now grew to 10,000 voters or 100,000 voters, I could then call on the person who is the state attorney general in Tennessee and say,

01:12:51.060 hey, guess what? 500 voters in your state are members of my voting group, are subscribed to

01:12:58.500 receive my voter guide. Do you want to chat about what's important to us? And simply the fact that

01:13:05.540 I have 500 people receiving a voter guide from me should probably make that elected official or that

01:13:12.660 candidate want to hear what this group wants. Now, they're not required to do what we say,

01:13:17.880 but they would be interested in knowing oh somebody is sending a voter guide to 500

01:13:23.720 registered voters in my state i'd want to hear you know what their values are what's important

01:13:29.480 to them and and so the the essence here is that any group of people have more influence if they

01:13:37.580 vote together and sway enables anyone to join a group and anyone to start a group on any topic

01:13:44.300 um by the way this is not a novel idea this is the exact same tactic that's used very effectively by

01:13:52.820 by you know uh well-established groups uh the labor unions send voter guides to their members

01:13:59.540 yes and the labor unions exert enormous influence on city politics especially in you know in

01:14:06.940 democratic cities another extremely effective group is the nra and the nra is um it has i think

01:14:16.300 3.8 million members uh and it sends voter guides to all of those members and um and 3.8 million

01:14:25.560 is a large number but it's not a large percent of the voting public if you think you know it's

01:14:31.540 a couple percent of the voting population of the United States, but that's enough to completely

01:14:37.680 kind of dominate the legislative landscape on their lane, on their topic, which is

01:14:46.780 gun regulation, right? How are 3.8 million people able to exert that much influence

01:14:56.060 over, over a topic. It's because they're subscribed to a common voter guide

01:15:02.480 and it alerts them to vote in every election. You know, they vote in primaries. And, and lastly,

01:15:09.820 the leaders of it can go to a candidate for office and say, these are our demands. If you want,

01:15:17.140 and these are how many voters in your district are subscribed to our voter guide. If you want

01:15:21.720 these votes just do these demands and you'll get these votes and if you don't somebody else might

01:15:28.480 get these votes because we'll put someone else's name on our voter guide and it's a little bit

01:15:33.740 vulgar but this is how this is how it works and and it's very effective and now whether you agree

01:15:42.560 or disagree with the nra's um causes you have to acknowledge this is an effective tactic by the

01:15:49.320 most people who support the nra don't support the labor unions and vice versa but both sides

01:15:55.720 use these tactics very effectively with sway we're just democratizing this tactic meaning

01:16:02.760 anybody can create a voting group on any subject whether it's holding ai accountable or policies

01:16:10.440 for your local school or you know any policy you want if enough people care about something

01:16:16.120 and subscribe to a voter guide for that something,

01:16:19.860 they can together advocate for change in a new way.

01:16:23.480 Is there a cost?

01:16:25.180 No, there isn't.

01:16:27.440 In fact, Sway actually offers things that would cost money for free.

01:16:33.740 So, for example, I can send occasional text blasts to my voting group for free

01:16:40.580 and other tools that let you text a group of people like charge you for the for the cost of

01:16:45.980 texting why would you be offering it for free so yeah it's a good question so sway is a public

01:16:52.980 benefit corporation it's which means that it's kind of a it's for profit but it's for profit

01:16:58.540 with a mission uh so it is uh it's uh it's investor backed uh it is not currently making

01:17:07.180 money and it's uh very you know overtly saying there's no current plan to make revenue um but

01:17:16.280 it's also what's your investor pitch like i love this i can just imagine you talking to

01:17:21.900 investors there's no plan to make money and i don't actually see a long-term plan to make money

01:17:27.980 how did that go so um yeah by the first full disclosure because we've talked about this but

01:17:34.280 listeners may not know so i've created a voter guide on sway but i also am one of the creators

01:17:40.040 and investors in this platform sway yeah no i know yeah um sway is a independent you know startup

01:17:48.920 that's been around for over a year now but is just now uh beginning to you know uh kind of

01:17:56.200 do its public uh public promotion um and i have i have a vested interest in it i just want to make

01:18:02.280 sure that it's disclosed um but yeah the investor pitch it you know twitter for like a decade didn't

01:18:10.840 have revenue um so i guess one thing i should i'm sure you're you're familiar with this the word

01:18:16.840 for-profit a non-profit does not mean one of them makes money and one doesn't oh i've noticed it's

01:18:22.920 a little bit no a little bit um deceptive uh i think the probably the way for people to really

01:18:30.600 better understand it is with a non-profit there are no owners uh if a non-profit builds up a bunch

01:18:38.600 of assets nobody owns them it's like owned by this entity but no individual is the owner of those

01:18:45.400 assets you could have a non-profit that is actually making tons of money and a lot of them are yeah

01:18:53.080 and enriching people but it's but nobody owns that non-profit per se um uh with a for-profit

01:19:02.440 there is an asset and there are shareholders who own it and you can have a for-profit company that

01:19:09.320 keeps losing money but whether it's making or losing money the the underlying assets can be

01:19:17.560 owned by people and um and um now the the hope for investors in uh in a corporation that's not

01:19:28.380 making money is that the asset grows in value even if it's not generating cash flow and twitter is

01:19:34.900 the example that you know um for a very long time didn't didn't even try to make revenue and uh you

01:19:44.760 just grew and grew lots of people use it and it grew into becoming a indispensable part of the

01:19:51.960 infrastructure of communication and um and i think its creators and its investors understood that

01:19:58.280 even if this loses money for years it's building real asset value and whoever owns a little piece

01:20:05.080 of that will get wealth eventually um and um now uh the yeah the so this the service sway

01:20:15.240 kind of is hoping to follow similar arc and um you know now there might be ways to monetize it

01:20:22.640 down the line but i would say the um uh it's anything involving voting is um very sensitive

01:20:32.300 and very important to be aware that the influence of money in voting and politics can be very

01:20:38.400 corrosive and very harmful.

01:20:42.060 So the mission, including the public benefit statement for Sway, is very much to align

01:20:50.660 it with having government better align to public opinion and to the will of the people.

01:20:59.760 Why is that important?

01:21:00.940 i mean i think that's um obviously it's uh no i we've talked about this i just i think it's worth

01:21:07.880 hearing you articulate it and the backdrop to this conversation i just i don't know when this

01:21:12.020 is going to air but is the riots uh in france and across europe right now of young people

01:21:17.320 burning things down looting stores uh going crazy so with that with that backdrop why is it important

01:21:24.260 that what leaders do reflects to some degree

01:21:28.180 what the population wants?

01:21:31.940 I think, first of all, that's, to me,

01:21:34.260 it's just inherently right.

01:21:36.680 It would be wrong for the government

01:21:37.820 to do what people don't want.

01:21:39.600 But I think it's also just the only way

01:21:42.200 for it to be sustainable and stable

01:21:44.040 is for government to,

01:21:47.500 at least in some sort of medium term,

01:21:51.460 do what's best for people.

01:21:52.580 sometimes government should do unpopular decisions for the short term and you know and i think um i

01:21:59.540 think it's healthy to have a little bit of separation from the people you know i think one

01:22:05.540 of the uh representative democracy i think is is a genius concept because it creates a little bit

01:22:11.620 of separation where people entrust a representative and then that person makes decisions um some of

01:22:19.620 which may be unpopular but you know ultimately has to come back to uh the will of the people

01:22:25.300 to be re-elected or so on um so but if if a government strays from the will of the people

01:22:32.420 for too long it just becomes unsustainable and um and i will say tucker i think that

01:22:42.420 something that worries me in general about the state of the world is across the board i feel like

01:22:47.620 governments have strayed from what people want you know um i i love america i'm american and i'm

01:22:56.660 uh most hopeful about america's system of government but it feels even here that the

01:23:02.900 government doesn't reflect what the people want but this is also true in in lots of other countries

01:23:08.820 and um and that's worrisome and you know and i uh i feel like it should be a solvable problem

01:23:15.860 you know we have we have humans doing such incredible feats in other domains right yes

01:23:22.340 we're sending spaceships into outer space we're taming nuclear fusion we just had a flying car

01:23:29.860 you know we talked about artificial intelligence and yet the system of government is feels a little

01:23:36.420 bit like it's regressing and going backwards towards the the middle ages instead of modernizing

01:23:44.580 and and uh and doing better or even worse it feels like a medieval form of government with

01:23:50.020 like space age technology which is like the most terrifying combination i can imagine yeah yeah and

01:23:59.140 um a lot of the smartest people and i i you know i work with some of the smartest people on the

01:24:06.260 planet this i i run neo which is a mentorship community and um like a school for startup

01:24:12.580 founders and the smartest young people in tech are drawn to the cutting edge technology much of

01:24:19.540 which is some of the things i just mentioned like ai or space exploration or so on um but i believe

01:24:26.660 that one of the biggest areas for impact is just the domain of social engineering things like

01:24:32.740 social media and information distribution or with what we're doing at sway

01:24:42.260 citizen organizing and influence in government to bring government back to reflect the will

01:24:49.940 of the people and i'm hopeful that more and more smart people will realize that even if it may not

01:24:56.340 be as techie that this realm of communication and social engineering is where some of the most

01:25:02.780 important impact is to be had. So if you could just bottom line on your way out, are you, I mean,

01:25:09.980 you're a tech guy, you're an investor, you have long-term investments. So just by definition,

01:25:14.940 you're an optimist, obviously. I am, indeed, indeed. You haven't gone to gold.

01:25:18.500 um but are you still optimistic knowing everything that you do no i i am optimistic tucker um i

01:25:30.800 yeah um and yeah it's true i was thinking as a as a tech guy not just as an investor but i've

01:25:37.800 found in multiple companies i think you have to be not just optimistic but even a bit naive to

01:25:43.880 to start companies so i'd probably rather be accused of being naive even than than to be a

01:25:49.740 pessimist um but i'm optimistic because i i believe in in the inherent goodness of people

01:25:56.560 and um you know even in this conversations where there's been a few points where you kind of said

01:26:03.300 oh this person these people must be all cynical i always think well maybe there's actually a

01:26:09.800 explanation for the behavior that they're they're doing what they think is best and could be

01:26:14.800 reoriented towards the right thing i i tend to believe in that um and you know coming back to

01:26:21.280 the um to the starting point of the conversation of the ai topic um you know i think that it's very

01:26:31.100 um very solvable to prevent ai from destroying humanity um it's been it's been a topic that's

01:26:39.080 been like a debate for a while um and sometimes the solution is just reorienting it to let's look

01:26:45.720 at it this way how about redefining it to just making ai abide by laws and once you see that

01:26:52.260 switch it makes the solution much simpler and uh something that was previously seemed divisive and

01:27:01.000 hard to figure out, become much more kind of plain and more about execution. So I think even

01:27:11.840 the there's lots of problems I don't know the solutions for, but I, I believe that there are

01:27:16.540 solutions out there to be found. And often finding them is just about looking at it a different way

01:27:23.680 and, and kind of seeing an elegant solution that was always there in plain sight.

01:27:32.000 Ali Portovi, thank you very much.

01:27:34.160 Thank you, Tucker Carlson.

01:27:35.000 That was great.

`witch hunt`	Both words, anywhere in the episode.
`"witch hunt"`	The exact phrase.
`witch OR hunt`	Either word.
`"alpha male" -"high value"`	The first phrase, but not the second.
`(feminist OR feminism) AND (marriage OR divorce)`	Parentheses group alternatives.
`"decline civilization"~5`	Both words within 5 words of each other.

`misogyn*`	Any word starting with "misogyn": misogyny, misogynist, misogynistic…
`feminizm~1`	Words one letter off: catches typos and transcription errors.

`person:"Andrew Tate"`	Episodes where they host, guest, or are mentioned.
`host:"Glenn Beck"`	Episodes they host.
`guest:"Jordan Peterson"`	Episodes they appear on as a guest.
`mentioned:"Andrew Tate" -host:"Andrew Tate"`	Talked about on other people's shows.
`show:"Bannon's War Room"`	Episodes from one show.

`topics:abortion`	Episodes tagged with a topic (see the Topics page).
`summary:"Andrew Tate"`	Words in the episode summary. Summaries are machine-written and can be wrong.

`"great replacement" AND episode_date:[NOW-1YEAR TO NOW]`	From the past 12 months.
`lockdown AND episode_date:[2020-01-01T00:00:00Z TO 2020-12-31T23:59:59Z]`	From one year. Dates need the full time, as shown.
`"high value" AND misogyny_count:[100 TO *]`	Episodes with 100 or more sentences flagged as misogynistic. Also: hate_speech_count, toxicity_count, word_count.

The Tucker Carlson Show - October 31, 2026

Forbes #1 Tech Investor Warns of Revolution, AI Bank Heists & How to Avoid a Dystopian Future

Episode Stats

Transcript