It is difficult talking into a microphone, and I've noticed this increasingly about myself that my brain breaks all of the time. And I think when you're talking to somebody in like normal life and you sort of stumble, they just fill in whatever word you're looking for. But when you're on the microphone, you're like...
That's why you got to do it like a bazillion times.
So Scott is incredible. Like his brain, his brain does not break on mic. Josh's brain, your brain does. I always lose the sentence or lose the word that I'm looking for all the time.
Oh, I don't think you always do. I mean, it feels, I feel like I do it a lot. You never do.
I just, I told you I listened to your conversation with the guy from Hamilton Lane that you did it this week.
I thought it was like one of the best things I've ever heard you do. And maybe it's because you know the topic so well. It was like a flawless, no um, no uh, like as soon as he finished a sentence, you had the next question ready.
Well, thank you. I did...
You reacted to his statements immediately.
I don't think I barely feel like I finished something. I'm like, 'Yeah, that was pretty good.' But I did do a good job. But also...
I know the content. And he was very good. So he made it easy.
He was very good, too.
When you do a bad show, how long do you think about it for?
I used to think about it for weeks. Now, I think about it for hours.
Yeah. But the thing I've learned is that the more you do, the less each one matters.
And there's also less time to work because you have another pod the next day.
You got another thing to figure out.
And nobody cares. Are you in a lot of two-way conversations with your audience? Like, do you get a lot of feedback and do you respond to any of it? Do you use that to try to improve the show?
Yeah, I do. But I think you got to take everything with a grain of salt because people online are generally insane.
I don't mean comments. I mean like literally somebody comes to you and says something constructive but critical.
Not in an open forum. If they're right.
Sometimes they're wrong and I'm like...
So I feel strongly about this. I very much try to avoid the comment section.
Because Bill Simmons said this and it crystallized with me because I obviously, like everybody else, it's weird to see strangers talking about you in a personal way. It's just a bizarre feeling.
I understand and I'm thrilled that people even care to listen to us. So I'm not mad that people talk. That's part of the game. I accept it. I totally get it. But I don't want to subject myself to it. Number one, like it's not healthy. But number two, and more importantly for the audience's point of view, I want to say what I want to say...
...without a random anonymous person's comment in the back of my brain. It's not fair to the audience because the audience isn't reading the comment section for the most part.
So why would you steer the conversation away, try to avoid language?
Just to please them, you're going to say the things they want. Then you're not going to say your opinion anymore. And I think the most lame thing that a podcaster can do is talk to the negative comment section. That's not what the audience is there for. They don't give a damn that somebody...
Oh, fight with the comments.
...threw a tomato at you. Like they don't care. Just...
You got to move on. But I think some people say don't read them at all. Zero. I think that it's worth just checking them out and getting a sense of where people land.
But at a certain point you don't have to anymore.
Yeah. I think read them and don't take them too seriously every now and then.
Here's a happy medium. Skip the comments entirely.
But pay attention to the analytics.
Of course, pay attention.
That's people putting their money where their mouth is. And by money, I mean their actual time and attention.
So if you do a show, which we had Ed Zitron on.
He's crazy, but he's a great guest and we really enjoyed the conversation.
And by the way, he does his homework and so this is the thing.
Listen, I don't have to agree with his conclusions, but he's doing his homework.
75 numbers ready. He was very prepared.
Correct. He's a great guest for any show that wants to have an opinionated, spirited discussion about AI, which is the most important topic of our time.
That's what you want. You don't have to agree with him, but you know, his view is valid, I think, because he's doing his homework. He's looking at the numbers and he's coming up with a view. You don't have to agree with it, but I think that...
So we did a great show and then I looked at the analytics. First of all, the show did huge numbers.
So yes, we had people in the comments like, 'Oh, that guy's the worst,' or 'charlatan, he doesn't work in finance, he's saying don't even invest on what my opinions are, I'm just spouting off' — like every criticism you could think of. But you know what? Forget about the views. People stayed for the whole thing. The thing that I look at — if I have a guest where people drop off two-thirds of the way through in big numbers, I'm paying attention to that. That's signal.
That's not one guy spouting off in the comments. I never want my fans and my audience to be like, 'This is not a good use of my time.'
And you don't have to agree with them. And that's the point of having guests on is to have a different point of view. You guys have your point of view, have the guest on, then bring on another guest who has a different point of view. This is the whole point of all of this.
To be clear, if anybody heard me say the audience, that's obviously not what I'm saying. I just say I think...
He knows the cameras are rolling now. He's got to hedge his...
No, no. I think you're doing your audience a disservice if you focus on the negative comments. Bring it into the show. That's my point.
I agree. And you can feel the anxiety on the host like, 'Well, this is what I said.' You're trying to justify that. Just, it's okay. You have an opinion.
That's out of respect for the audience. Michael is not going to slavishly focus on what the 1% of the audience thinks just because they're the loudest.
Exactly. Then you wouldn't be here. You wouldn't be telling your own opinions that come from your own experience, from your own view, your own data. You wouldn't be here telling your opinions if all you cared about is what other people think.
So did you go to school for finance? Like, what is your...
I went to Princeton and I studied classics and I was, you know, deep into...
I was going to say, are we looking in a mirror?
That's crazy. Wait, what dorm? What dining club did you...
All right. So, Princeton Classics. Go on.
All right. I would have said Princeton because you told me the story, but I don't remember.
And then, but you had a great story of how you met Scott.
Let's hear it because I know it but Michael doesn't know.
Well, I was a huge fan of Scott for a long time. I was listening to his podcasts. And then I found out that an old roommate of mine — I was listening to one of his episodes and I found out that an old roommate of mine, his mom was a guest on the show. I'm like, 'Oh my god, this is my hero.' So I try to get connected. I have a kind of awkward phone call with Scott, but I try to make my pitch. I end up being his intern. He hires me out of college. I work with him doing everything, helping him with his books and his research and his decks. And then one day he says, 'Okay, I want to start a podcast. I need someone to read the news.' And I said, 'Okay.'
I was on one of those first episodes with you guys.
Right. And I would literally read the news. I would just boom. I had a script. I'd write up what happened the previous day and start to react. And eventually I start realizing, okay, I have a viewpoint. I want to start expressing it. So I do.
The podcast grows and now we have a relationship where it is a co-hosting dynamic. I host now four out of five episodes a week and Scott joins me on one of them. We review the previous week in markets and finance and tech and anything that is relevant to us. And yeah, and here we are.
The show is on fire. You guys, it's a great dynamic when the two of you are on because you're not the same generation. You don't agree on everything.
He actually is a very good listener on the show with you. And I think him and Kara is obviously chocolate and peanut butter. It's a great combo and they have great banter. I think with you, you probably say more than she says. I think he really gives you room to express an opinion. He's a great listener.
Yeah, he is better than I am at that.
As you'll come to find out shortly.
No, I think that's cool, man. Like I watch you guys on YouTube or whatever, or on Spotify, and I'm looking at the screen and there are long stretches where he's just sitting quietly listening and nobody would expect that because it's Scott Galloway. Of course he's doing the talking. So I think that's a really cool dynamic that you guys have, and then you have a lot of the episodes on your shoulders.
I think it just works. My wife listens. She started out as a Pivot fan and then discovered Prof G Markets because I was on the show, and you're in her rotation. She loves you guys.
Well, I totally agree and I think it's credit to Scott for deciding to take a bet on someone who is young, on someone who a lot of people would say, 'Why does he get to have an opinion? Why is he talking?' I understand it. And if you don't want to listen to what I want to talk about, then don't listen to the show and that's okay. But I think what he has been really good at is identifying that actually young people, if you give them a shot, they can be pretty good. And I think that show is sort of the evidence of that. And so I'm super grateful that he was willing to take a bet with me and he didn't have to do it. But he did.
Shout out to Professor Scott Galloway for sure.
You know, Barry and I know Scott for a really long time. He's been at some of our events. We think he's one of the most interesting people we know.
He is a fascinating character.
100%.
And very polarizing.
Yeah. Well, so that's another interesting thing that I'd love to just hear from you on. So he's not afraid to go there. Like he talks about sex and tons of politics especially lately and I'd imagine that'll get louder as we go into the midterms. He has people on both sides of every debate. He's not afraid to listen to what people have to say. How much blowback do you get depending on who his guest is? And is that something that you can just let run off your back or what's that like?
Yeah, I sometimes get blowback if people are upset with Scott and they say, 'Oh, screw you, Ed, for working with Scott.'
How would you sit there silently while he let so-and-so say blank?
Exactly. So I personally don't worry about that because I know that it has nothing to do with me, but it's a very small minority of listeners. I think most listeners listen to the show and they know that we're different people. We have different views. We have different priorities, different values. And you hear that on the show every episode because we're arguing with each other. I'm bringing a different perspective. I'm bringing the young person's perspective. He's bringing his perspective. So we push back and forth. And so I think generally speaking it's not a problem for me because I think people recognize that I am a different entity, different mind, different view. And I think most people can understand that.
Yeah, I think your audience is very — you probably have a very highly educated audience. Just given the nature of the content and Scott's persona, I think you're talking to people that are accustomed to being confronted with other points of view.
Yeah, I think so.
I would hope, like that's what we aim for. I try to make sure that the show isn't too esoteric or annoyingly intellectual. It's important to me that we're very clear about everything. Everything's simple to understand. Anyone can understand what we're talking about and understand the point of view. But I think one thing that is probably true about the show is that people who listen to it are ambitious and they give a damn about what's happening in the world, what happens in markets, what matters, why money matters. I mean, all of these things I think is probably true of most of our audience, which I love because that's how I feel about the world, too.
Very cool. Well, get ready for some esoterica on today's edition of the company. How we doing, guys? All right, we got John Grayson and Daniel Parah in the house in the control room right now. They will be piloting the show. Thank you guys as always.
Hey, John, what episode is it? Oh my.
This episode is sponsored by ROCY and ROCQ from J.P. Morgan Asset Management. Two new option-based ETFs for investors looking to complement an income approach led by Hamilton Reiner and an experienced team. ROCY and ROCQ seek to provide monthly distributions. These distributions may come from stock dividends and selling options and they may also include return of capital. Both also seek equity upside potential with less volatility.
ROCY offers exposure to U.S. large-cap stocks while ROCQ offers exposure to the NASDAQ 100. Investing involves risk, including possible loss of principal.
Investors should review a mutual fund or ETF's investment objectives, risks, charges, and expenses before investing. This and other information is in the prospectus. Read it carefully. To obtain an ETF prospectus, call 1-844-JPM-ETF.
J.P. Morgan Distribution Services, Inc., member FINRA. To learn more, visit jpmorgan.com/income or search ROCY and ROCQ.
Ladies and gentlemen, you are now rocking with the greatest investing podcast of all time. I listen to all of them. You can take my word for it. Not the only good show, of course. The control board, just a gold standard. All right, guys. My name is Downtown Josh Brown. First-time listeners, welcome to the show. With me as always, my co-host, Mr. Michael Batnik is with me today as well. Michael, say hi.
All right. And we have podcasting royalty. Today, my buddy Ed Elson. Ed is a business writer, analyst, and co-host of the Prof G Markets podcast with Professor Scott Galloway. In 2025, he received the Webby Award for best business podcast. Did we ever win that twice? Amazing. And before that, the Honore Award for best co-host. I like that. That's cool.
In 2025, he was included in the Forbes 30 Under 30 list. So, are you stealing money yet?
I'm going to jail next year.
When do you get indicted?
His work has been featured in multiple New York Times bestsellers, including The Algebra of Wealth, A Drift, and Siege. Ed Elson, I can't believe we haven't done this yet. So happy to have you.
So happy to be here. So excited. Big, great, big fan of the show.
Greatest intro music in all of podcast.
I definitely agree with that. I definitely agree with that. All right. Question one. AI. It's good.
Is that how we're starting?
Speaking of — all right. Speaking of giving you room.
We're going to tackle this from a markets perspective, but we're also going to get into this from a societal perspective. And I know you're very passionate on that end of the debate. I am too. I just don't spend a ton of time on it just because the focus of the show, but I do want to do that with you.
We have this setup where tech is the everything cycle. Are we starting there?
Yeah. So just to frame the conversation, A16Z put out their state of the markets and one of the first charts is, as Josh mentioned, this is how we got here. Daniel, turn it on, please. They have a chart showing that tech is now 55% of all U.S. capital spending.
And they zoom out from the 1960s to today. And this really does a great job visualizing the story of why all of the most famous and most powerful CEOs in our country were at the White House yesterday talking about technology and the next wave and AI and superintelligence and this is it.
Yeah, that's exactly right. And it really — I mean, I know there is an end in sight, but it's not in sight today. It really — I mean, it looks like it's hockey-sticking even from these levels.
I mean, could tech become 80% of U.S. capital spending? Yes. As inconceivable as that sounds, again, in 1960 it was 15 and today it's 55. It just seems like we're in an — like these are the times that people will look back 20 years later and say, 'Holy cow, that was it. That was the revolution.'
All right, so let's start here.
From a markets perspective, it's a pretty exciting time for you to be covering the markets.
Every day is an incredible new story. I mean, the White House meeting of all of the AI leaders, I thought was incredible entertainment. I thought the clips were amazing. Trump jabbing at Dario Amodei, trying to get him to say something. Him feeling awkward. Say something positive. Him calling Sundar Pichai a monster but then saying that as if it's a good thing and then Sundar Pichai saying, 'I don't know if that's the right label.' I mean, incredible entertainment.
He pointed at Sundar Pichai at first — monster who no one knows.
He couldn't remember his name at first. Eventually he goes, 'Sundar.'
It sounded like he was searching for it.
So he was like, 'Look at this monster.' Sundar Pichai, of course, doesn't know what to do with that because he's never confronted an individual like Trump.
So incredible entertainment and that's it. That's all that was that meeting. I think that there was nothing that actually came out of it that actually mattered. We thought that maybe they were going to come up with some level of regulation that was actually important coming out of this AI hysteria, which I know you and I have talked about and we could talk about again. In terms of what actually got done in this meeting, nothing really got done other than...
Wait, what do you mean? This superintelligence. We got a new name. What are you talking about?
It's not artificial because artificial is...
Superintelligence. Exactly.
Super. Did you watch any of this?
Okay. You got to find some clips of this. I was watching it yesterday.
Melissa Lee and Michael Santoli were — I think their show is the one right after the closing bell, like right after 4:00. I guess it's the close. I don't know, doesn't matter. They were like stunned because it's $2 trillion of market cap in an impromptu press scrum in front of the White House. Nobody has ever seen anything like this. This would be the equivalent 100 years ago if you saw like — I don't even know — Warren Harding standing in front of the White House with John D. Rockefeller, Cornelius Vanderbilt, J.P. Morgan. Like this is totally unprecedented.
And he's demanding and pointing fingers. Okay, now you speak. And they go in there and they bumble their way through trying to talk about how AI is actually a good thing. Dario Amodei is saying this. Meanwhile, the Anthropic S-1 is literally saying that it might extinct humanity. So I mean, the extent of the hypocrisy here, the way that these guys are folding themselves into bits is astounding, but they have to do it in front of the big man. And that's why I think this whole thing was total performance art. Donald Trump said, 'I want you to get in there and I want you to say that AI is going to be nice and AI is going to be safe.'
You think they needed the coaching? You don't think they know?
You watch those clips? Those guys do not know what they're talking about.
No, no, no. But you think they needed him to say, 'All right, I want you to be positive'? No, they know like my role here.
I think Dario — yeah, I think they eventually showed up and they realized, 'Okay, the memo here is President Trump wants me to come in and disavow all of the things that I've said in the past about how AI might extinct us. And I'll just forget about the fact that my current employee of the company said that there is a 10% chance that my own technology is going to wipe out all of humanity. I'm just going to try to put that aside because I know that the president doesn't like that. I know that generally speaking, the business world isn't a huge fan of that. I'll show up and say what I'm supposed to say.' And you listen to the way that they delivered those messages. Those were not honest expressions of their views of AI. That was a hostage statement. That was President Trump putting them on the spot, doing it well — I will give him that — of saying, 'Okay, now tell us how actually AI is fine.' And who knows what they believe at this point. And this is the problem with the whole conversation, that no one actually really knows what they're talking about. None of it is based in evidence. None of it is based in data. That 10% calculation...
Well, we're going to go there in a sec. Let's put a pin in that. I want to show a few pictures. Okay.
Daniel, you have some shots here.
All right. So, beautiful.
All right. I thought this was interesting. The seating chart. So the guys standing in the back row were obviously seated on the side of the table where their backs would be to us. So, but look at who's sitting where. It's almost — it's Da Vinci.
It's almost like Last Supper-ish.
All right. So Trump in the middle, Elon on one side, Jensen on the other. Let's just start there.
Jensen is the best — I think the best politician in that room, right? Would you agree with that?
I think Jensen Huang has been really good on the messaging. I think that there's a lot of people who would say that he's made a lot of screw-ups as well. The first one that comes to mind for me is in the Ezra Klein interview. I thought that actually a lot of what he laid out about AI regulation was directionally correct. His only mistake was saying, 'I don't know my home address.' That was weird.
And he said, 'I don't know.' 'Are you sure? You don't know?' He said, 'Yeah, I'm sure of it.'
But it went from 'We started a company at Denny's' to 'I don't know my home address.'
If you're trying to communicate that you're in touch with reality, it's not a very good way to do it.
I think there's a lot of hysteria around what AI is going to do to humanity.
And I was saying this to Ben yesterday. I think the term AI slop is so overused. I think people have lost their minds. This idea that AI music, AI language, AI movies, it's all going to replace humanity. Nope. Not even — like stop. Just stop. I reject all of that. And I understand why certain groups of people, journalists for example, feel threatened. They should. I totally understand that part of it. What Jensen said with Ezra that I was nodding my head at — and I think, I don't know if this is controversial or not — do young people need to know how to do long-form division? Do they need to know the square root of 225? I don't know. But then when he said, 'I don't even know my own address,' I think he probably regrets that.
But I was like, wait, what? That's bizarre.
I think that it is a very bold statement and I think a lot of statements are being made that are very hot takes, very bold statements and they're being expressed with a lot of confidence and certainty about where the world is headed. To say as confidently as he did that you don't need to know how to do basic math in this world of AI — to me, I mean, it's a viewpoint, but it doesn't mean that he's correct. And I think that he should be expressing these views with a little bit more caution because I mean, I'm with you. I think that young people need to have a basic understanding of how the world works. They need to understand basic math. They need to understand history. We got to know arithmetic. And this is all important. And we're already seeing that young people's brains are melting — not even because of the AI but because of the years of screen addictions that have led us to this point. And we're seeing it in the testing scores in science, in math, in reading. It's all going down. And then we see these new reports that AI actually makes it worse. That if you're using AI daily, your test scores on science tests is actually lower than if you were not using...
I have a very different point of view. Okay.
I have college-age kids and in no way, shape, or form — and this might be very specific to where I live or the level at which I'm trying to have my kids compete — in no way, shape, or form would I say their age cohorts, one is 17, one is 20, have melting brains. It is so much harder to get into the schools that my generation was able to walk into with 70s and 80s GPAs. And the level of work they're doing in high school looks more like college. If they want to go to business school and they're not taking advanced placement classes, where I live, don't even bother applying. The average SAT scores for the 50 schools that we've all heard of, like whatever that list is, they're probably up 30% from the averages of just 30 years ago when me and my friends were applying. Schools that, like, no offense to anyone listening to this, you could waltz into Ohio State or University of Maryland in 1995 with the bare minimum amount of work done. They were filling quotas of out-of-state kids. Now these schools are almost — the schools in Florida, you cannot — UF might have a lower acceptance rate than Princeton right now. You might not believe me, but I think that's actually accurate. So I don't think that these kids — screen addictions, and these screen addictions are very real — I don't think that's universal to all kids.
Correct. I think we're still creating stars and performers and I've met with tons of college-age kids over the last couple of years, both friends of the family and randomly just speaking at universities. These kids are so much further ahead than we were. The last thing is...
I'm going to let you. The last thing is I think it's fine if they're quote-unquote cutting corners with AI because that's the world they're going to graduate into. That's what the adults are going to be doing who are handing them tasks. There's no brownie points for taking a number two pencil to a piece of graph paper. No, you don't win any awards for doing that in most vocations. When I went to school, we didn't even have computers. There was no internet, no phones. What we had were word processors where you would literally type and there was an eight-digit LED screen to read your work. Why would we hold that as the standard? Like, are we Amish? Like, why would we force kids to — they're not teaching them cursive anymore in elementary school. Good riddance.
Somebody sent me a letter written in cursive today. I threw it out. I can't read it. Like, why would we still be teaching that? It's ridiculous. It has no value. So I don't know that I fully buy into this idea that the screen addiction and the AI is melting brains. I think it's just changing what we as a society say is important of the top-performing kids that are coming through the world.
So what I would say to you is I agree with your point that...
His mic, that's good enough. We can move on.
...that the standards are high and especially when you look at some of these top schools that you're talking about and I have no doubts that your kids are excelling in their education.
You could have some doubts.
They're doing okay though.
What I can tell you is in the same way that we have a wealth inequality problem in America that is pervasive and getting worse, we also have an educational inequality problem.
And I can tell you that the gap in attainment between the top 10% of students, the bottom 10% of students in America is higher than any other developed nation in the world aside from Luxembourg. I don't know what's going on in Luxembourg, but that's what I can tell you. And I can also tell you that the gap is getting worse. I can also tell you that just in terms of the median average score, we have the test scores. The test scores are going down. It's just a fact. And I think it's possible...
There's a wealth inequality story there. I agree with you.
And the top students are doing well and they're getting better. The bottom students are completely falling behind and we're seeing it literally in the test scores. And so I think that's going to be the real problem here is that how do we make sure that we have an equal platform for everyone? And in terms of AI, I mean, I think there's a question about work ethic.
Wait, but wait a minute. What if another version of this is that AI does not only not exacerbate what you just described, which by the way was pre-existing way before we had LLMs.
And you're dead right on it.
What if that is part of the solution?
It might be, but I'd like to see a bit more of a thoughtful conversation around it. I would rather not see Jensen Huang going on and just randomly saying you don't need to do long division. And there are, to be fair, there are thoughtful conversations being had, but it needs to be about, okay, how do we think about work ethic? How do we think about attainment? What does intelligence actually look like? If the answer is, oh, use AI because it's better to cut corners, I don't think that's going to make for a very...
No, use AI to tutor you because your parents can't afford $150 an hour for a biology tutor. But here's the good news that I like.
No longer necessary. That's the story that these guys, I think, have done a very poor job at telling, but it's actually one of the most immediate consequences of intelligence at everyone's fingertips for $10 a month.
It literally could be revolutionary for the kid who does care and whose parents care but lack the resources so that they can't put their kid into a private school or bring in tutors.
But it's a question of learning versus just being given the answer by AI. Ask those kids. Let me ask a question.
They have to learn because they can't pass the tests if they don't learn.
I think I understand that there is a large difference. I'm a 41-year-old male versus children. But I think you guys would agree that AI is making us better at what we do. It is making us more knowledgeable. We know more things. We can do more things better, faster, easier. Why do we think that it's going to be the opposite for children? Why is it going to melt their brains? Because I'm just looking at the numbers that was reported on by the OECD, their PISA test, where they literally just looked at the frequency with which students are using AI and compared it to the students that aren't and just seeing the test scores are better because they're using it as a crutch because you can just get the answer from the AI and then eventually when you get tested and you have to use your own knowledge, you don't have it. You haven't done the memorization. You haven't gone through the pain of actually learning something. I think this is true of any skill that...
But I'm saying tell the AI, 'Test me. I have a test coming up.'
Here are the seven websites that the teacher has directed us to study from. Test me until I can get a 90th percentile. Like...
That's the practical. I'm not saying everyone should be doing that, but it should turn into...
That's what it should look like. But that's not really the conversation that's being had. And to be clear, I'm not an expert on education. It's just something that I think is important. And so I've done a lot of reading and research on it. But I think that that's the direction that we need to be going in is, okay, what does it actually mean to be a productive member of society? What does it mean to learn? What's the point in having a work ethic and working hard? What do we think about BSing tests and cheating and cutting corners? Do we think it's okay or do we think that actually we need to go through the hard and painful process of learning a skill and applying it, bringing your own originality to a problem? Those are the skills we need to think about and we need to think about how to thoughtfully put AI into that system. But if it's just, you know, his kids are under 10, their high school classes are, I'm guessing, all going to revolve around how to use AI.
Mhm. Like all of them. Like in other words, they'll be in chemistry, biology, European history, American history, global history, whatever they're studying, that's not going to change the way that the teachers get the curriculum across and test the kids 100% will involve LLMs.
Well, instead of textbooks, it'll be Claude or whatever. So, all right.
I'm not against that. I'm not against that. I just think we need to have a thoughtful conversation about how we inject it into our education.
We definitely do because that's our economy right there. That's our...
This is a potentially dangerous statement. But I really believe that a lot of times in the discourse the numbers either are lying or they're taken out of context or there's not enough zooming out...
Like, what were the test scores before? Were we going in the right direction and then all of a sudden AI came out? It's like we're so — they were bad and they got worse. So I also think it's — I think so that could be true. Obviously people are going to cut corners and get lazy, like yes, that is not not happening. But I think it's also very early and I'm not here to be the math police of artificial intelligence or superintelligence. I don't know. We don't know how the story is going to play out. It's very, very early to make declarations of it's good, it's bad, it's this or that.
But you hear me? I'm not declaring that it's bad. I'm saying that I'm in agreement with you. I'm saying let's stop declaring that it's good, declaring that it's bad. Let's have a thoughtful conversation grounded in evidence, grounded in scores. Let's have a more...
So I sort of think that massive rabbit hole.
So I've — you know, I'm very...
You and I in some way are opposites of each other. I think...
No, no, no. I think — no, we agree on a lot. So I don't want to phrase this wrong. I'm just a lot more cynical than you are. I don't think it's just age. I think I've always been this way.
My opinion is a lot has to go wrong in order for us to learn anything. And this idea that we're going to be thoughtful and have this national conversation.
We can't even agree on what gender Obama's wife is. What? We're going to have this thoughtful bipartisan national debate about how to implement AI?
No, but like think about the things that we're arguing about right now. We can't even say terrorism is bad. We can't even agree on what day of the week it is. So that's not going to happen. What actually is going to happen is terrible things are going to take place and then we're going to pick up the pieces and create regulation. Hopefully somebody's left alive in the aftermath. I already know that's how this is going to go. So I'm not hand-wringing over, oh, we need thoughtful regulation. Good luck. We 100% will have a series of crises and then things will — because that — and by the way, that's only the way it's always ever happened and that's why I know it won't be different this time.
I can hear where you're coming from and I'm completely with you that the national conversation is a show on basically every dimension.
But here we are having a thoughtful conversation on a big podcast. What else are we supposed to do? We want to show — can I show you around and just say AI is great? No, we're here to have a conversation. I think people will listen to this. I think they'll start thinking about it. I think it matters. I think it's worth having that conversation. So I would disagree that it's all pointless. Why talk about anything?
I don't think it's pointless. I want to show you what was actually announced. Okay. And I sort of know what your take is going to be, but I want you to get a chance to say it. Okay.
Put that last picture up with Elon in the middle.
So this is the part we've never seen before. To your point, you know, like the scrum. They're like, 'Mr. President, Mr. President, Mr. President.' He picks somebody, probably somebody that he's somewhat familiar with. Let's assume, right? Like he's not picking somebody from the back that he's never seen before. But even then, I like this. Rather than answer the question in Trump, which we know would be hilarious, he goes, 'Jensen, you got this one. Elon, come on up. You take this one. Zuckerberg.' I actually like this as a format going forward. Anytime there's some sort of industry-related regulation being considered, I like him putting them on the spot. We can argue about how genuine their answers were, if it was just happy talk. But like I sort of think the American public needs to see these people's faces, see how they react to questions, and listen to the words that they're willing to say into a camera. What do you think about that?
I'm okay with it. I don't feel that strongly about it. I don't think it tells us much because, as we know, they've been put on the spot. They've been given a rough instruction of what exactly they should be saying. So the words that come out of their mouth, I don't find that significant. It's all politics. It's all entertainment. It's all performative.
Let me ask you this. Getting back to what Jensen said, part of what he said that I strongly disagreed with was don't ship a product if it's not safe.
100%. We need — and Ezra pushed back rightly so — we need regulation. These companies cannot always self-regulate. Obviously we've seen that a million different times. Dario is saying, 'Regulate me. Regulate me.' Jensen is saying just ship safe products. And then interestingly the next day or two days later Jensen says, 'Today with over 100 industry partners we are introducing the NVIDIA open agent safety platform.' Whatever. And then to the point of it's all performative nonsense. Literally three days later, OpenAI scraps release of new AI model over safety concerns. Who knows if that's true? Were they doing that just to show that they're playing ball? Like, it's very hard to be...
He actually said. He actually took it further.
After this White House meeting, Sam Altman said, 'We will not go public until we can certify that what we're doing is safe.' Meaning he doesn't want to put out an S-1 that has those safety concerns that Dario just put out at Anthropic. Here's what they actually...
It's four things. It's four things. This is Trump on Truth Social. This is the official statement of what came out of yesterday.
One, implement robust internal controls. I would assume they're already trying to do that. I would hope. To monitor the capabilities and alignment of models during training and deployment. I think like Brad Gerstner probably helped write this or some of the All-In guys that are right because he's not — I don't...
He doesn't know what he's talking about. We know that.
But I don't think he would even represent that he wrote this. I don't think his people — like Trump White House people are writing this. I think the industry is writing their own regulation, which usually goes well. Two, empower an internal team to ensure all of the controls, monitoring, and detection are operating as intended. Again, something I would hope is already happening. Three — there's two more. Three, partner with an independent external auditor or evaluator to carry out independent assessments of whether the controls, monitoring, and detection are operating as intended. The partner-with part is — oh, sure, we brought in a partner. They're going to monitor. Okay, fine. Four, designate — all right, now we're getting somewhere. Designate an independent committee of the board of directors to oversee and receive reports from the teams operating the controls and the internal and external auditors and evaluators. Yeah, like somebody's ass has to be on the line.
You forgot the most important part of the contract though as the final piece of it. Make America great again. Signed, President of the United States. Listen, we're moving at rapid speed and it's AI and it's...
Wolf speed AI regulation.
You know, we're spitballing.
Well, look, the part I actually agree with Jensen on a lot of this is that there is a fundamental lack of accountability in this entire conversation. And I think that they're doing their sort of Trumpian way of getting to a place of accountability. It's kind of scrappy and it's not very thorough and it's a little bit...
What would you rather? All right. So tell me in your...
My dream scenario is a data center kill switch. I don't trust anybody. I don't trust any of these guys. I want the FBI to be able to shut down any data center that's hosting models that are endangering — like on-site. And I think we do have existing law enforcement and we need to maybe have a beefed-up AI, cybersecurity, whatever division. I like just shut the whole thing down if these guys can't get their act together. Let's immediately turn off the revenue knob the second we have a threat that these people should have been responsible enough to have prevented. That's my opinion. What would you have preferred to hear?
I think that's probably pretty severe. I would say — I mean, the thing that I wanted after we had that Jacob Coxon tweet which literally melted the entire world essentially — this guy says, 'I've quit Anthropic and everyone at the company thinks that humanity is going to go extinct.' He didn't reveal anything we didn't already know. We already knew about the Hugging Face incident which happened months before. The only thing he said was an opinion. Then a current Anthropic researcher endorses that and says, 'But by the way, it's a 10% chance of that happening.' Does not cite any numbers, does not cite any calculations. He just attaches a number to an opinion and then suddenly everyone says, 'Oh my god, it must be true because it's got a number attached to it.' Then an OpenAI employee says, 'Oh no, I think it's 70%.' And I think it's not going to be within the next 10 years. I think it's going to be in the next three years. And then suddenly, oh...
My God. And then the takes start to roll. Then everyone starts to pile on with the opinions. None of it's based in reality.
Well, here's your actual words.
Edson, can I do it in your voice?
Please. I'm excited to hear.
I've spent the past two weeks discussing this topic with everyone from engineers to computer scientists, policy makers and philosophers. I've talked about it on my own podcast and on national television after hours of rumination.
You like how terrible it makes me sound.
I finally... because you have like a posh British... I'm not good at distinguishing between... I just do Guy Ritchie or Michael Caine.
I finally reached my conclusion. This is the dumbest conversation of the year. And as soon as you said it, I said, 'He's right. These guys are saying there's a 17% chance we will wipe out humanity in 38 months.' Based on what?
What are you talking about? Okay.
But they're guessing. They're guessing.
What else can they do if they're worried?
Well, here's what you do. If you take their statement seriously, and we should because we're talking about humanity going extinct, if you take your statement seriously, then you have an obligation. It is on you as the government to investigate the out of their claims. And to me, what does that look like? It means handing a subpoena to OpenAI. It means handing a subpoena to Anthropic. It means bringing those people into Congress for questioning. It means getting to the bottom of their claims and asking them, 'Hey, you just released a corporate communication from a current employee. Forget about Jacob Coxin. He's no longer with the company. This guy is saying this.'
The CEO didn't double down on the 10%, but he didn't say anything about... he didn't say it wasn't true. So what are we supposed to believe? So then you need to bring them in and say, 'Okay, what was that based off of? How did you come up with a 10% probability that your technology, your product is going to kill all of us?'
It's that meme. I made it up.
