YT YouTube2Text

YouTube2Text → Transcripts → AI Cyber Panic, Ken Griffin’s Bubble Warning, The Fall of France?

AI Cyber Panic, Ken Griffin’s Bubble Warning, The Fall of France? — Transcript

by All-In Podcast · 20,879 words · 3,280 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:00All right, everybody, welcome to the
0:01literal number one podcast in the world,
0:043 weeks running. We got a lot to get to.
0:06Sachs is out this week. We've got a
0:08special guest coming up. Let's get to
0:10our first story. Ken Griffin says AI is
0:13almost certainly a bubble, and Chamath
0:16has responded. Chamath responding and
0:18moving markets almost every week here on
0:20the pod. Look out, France.
0:22>> [laughter]
0:23>> The Citadel founder made the comments
0:25this week at the company's annual future
0:28of global markets conference. But he
0:30says generally, hey, bubbles are normal.
0:32Any radical transformation of our
0:34economy is generally intertwined with a
0:35bubble. He said he's bullish on AI's
0:38long-term potential, but his concern is
0:41that companies are struggling to turn AI
0:43spending into productivity, something
0:45we've talked about here, I guess last
0:47year.
0:48And so, companies are, quote, looking at
0:52their bills from Anthropic and OpenAI
0:54and asking where the productivity is. We
0:56we had this discussion over and over
0:57again. On Thursday, tech continued to
1:00get
1:01crushed on the
1:04data center and CapEx front. So, people
1:06who are either making data centers, like
1:08CoreWeave, down 10%, or having a huge AI
1:12spend. Meta, which is an extraordinary
1:14one there, it's down 6% after earnings.
1:16AMD down 5%. SpaceX was down 5%. But
1:19these things have been swinging up and
1:20down a bit and generally trending up.
1:23Chamath, what's your take here on the
1:26market sentiment, October 22nd,
1:292026? What are your thoughts here on how
1:32markets are viewing the AI,
1:34quote-unquote, bubble? And do you have
1:36concerns, cuz
1:37Jensen from Nvidia thinks
1:40money is available and money is going to
1:42want this opportunity to finance it, and
1:45I think we all think companies are
1:46getting a ton of value out of AI
1:49technologies.
1:50>> My Social Capital research team had a
1:53really great call with Ramp's
1:56chief economist, Ara
1:58The first question that they had was
2:00because like I was doing my own math
2:04and I tweeted this thing, Nick, maybe
2:06you can put it back up there, where I
2:08said, "Okay,
2:09my back of the envelope math, knowing
2:11everything that I know and everything
2:12that I see, is about 1%
2:16of
2:18all of the companies represent 90%
2:23of all the revenue. And
2:26>> Revenue for AI spend?
2:28>> Yeah, so what I said was 1% of companies
2:30represent 90% of AI spending, so that's
2:32tokens. And of that 90%,
2:3480% of them
2:36are from unprofitable VC-backed
2:38companies. So, the cycle here, just to
2:40be very precise, is
2:42VCs raise from pension funds and LPs and
2:45foundations and family offices.
2:47They put money into a startup.
2:50That startup
2:52takes that money
2:54and invests in growth.
2:57Part of growing is by consuming other
2:59people's AIs.
3:01And so, that company then starts to show
3:03revenue traction.
3:05Those folks then say, "Hey, look how
3:06well we're doing." Raises a bunch of
3:08money from the same VCs, and it's just
3:10this circularity.
3:11Why that's important
3:13is what Chamath is pointing out is that
3:16for AI to really scale as a predictable
3:19driver of long-term earnings,
3:22it has to diffuse out of the startup
3:24ecosystem.
3:25And so, you can't be in a situation
3:27where on the one hand you're talking
3:28about breathtaking valuations,
3:31but where 75 cents of every dollar is
3:35actually just a startup sending it back
3:37and forth so that they can mark
3:38themselves up. That's I think a a
3:40structural problem. Now,
3:42we've seen that structural problem in
3:43the past. That happened in the early
3:45days of cloud when AWS initially started
3:48to
3:50expose EC2 and S3, the consumers of an
3:53EC2 instance or an S3 bucket were
3:55startups. But the difference there is
3:57that AWS was abstracted from Amazon
4:00retail, which was already then a roughly
4:03break-even profitable business, at least
4:05on a free cash flow basis. So what's
4:06different here is you have a bunch of
4:08money-losing entities just spinning a
4:10bunch of money around, and it has to
4:11diffuse more broadly. So I think that
4:14that's what Ken is observing. Now, when
4:16we brought that data to Ramp, the quote
4:18is incredible because this is what my
4:19researcher told me, "I prepared a list
4:21of factors to see if your thesis" This
4:23is him texting me.
4:25around VC money being 80 to 90%
4:28of the top 1% of AI spend. But he said,
4:30"Chamath's actual estimate is pretty
4:32much correct." And then the second thing
4:34though that he said, he sent us and Ramp
4:36hasn't published the chart, so thanks to
4:37Ramp and thanks to Arav. He said we
4:39could we could be the first ones to
4:40publish this.
4:41>> And just to be clear for the audience,
4:42Ramp provides financial services, neo
4:44bank, credit cards, primarily to
4:47startups. So they know startup data.
4:49There's a free commercial.
4:50>> Yeah. I think they know broad-based data
4:51from other data sources as well, and I
4:53think Arav has access to all of it. But
4:54they're But what he said is
4:56Now, look at this chart. So what this
4:58shows is how are the top 1% of spenders
5:02in AI spending their money?
5:05And uh I think that Ramp is going to
5:07publish this soon, as I said, so thanks
5:08to them for letting us preview this.
5:10What's amazing to me is the blue box in
5:13this histogram. What this shows is what
5:15percentage of money they are spending on
5:19GPU compute. Now, why does that matter?
5:21Well, the rose-colored box is LLM
5:24traffic. That's the typical spend you're
5:25going to spend on the big frontier labs
5:28closed models effectively, either
5:30through things like bedrock on AWS or
5:32what have you.
5:33What's really interesting is that the
5:35blue box is ticking up. The blue box is
5:36people buying their own GPUs because
5:38they're moving to open source.
5:40And so what I think is really
5:42interesting is you have this two-pronged
5:43thing that's happening. One is
5:47the diffusion we need
5:49within AI is happening a little bit
5:51slower than we would all would like.
5:53I see that at 80 90s well. I think the
5:55large companies that we work with
5:57you can almost pinpoint it on a chart
5:59ever since Alex Karp's commentary on
6:02CNBC about data sovereignty
6:04buying patterns have changed. And I
6:07think people are being a lot more
6:08methodical.
6:09They are building out their own data
6:12centers or at least moving into VPCs
6:14virtual private clouds.
6:16My team yesterday was at the CIO event
6:18of let's say one of the largest banks in
6:20America.
6:22You know what they said? I'm not going
6:23to say the name of the bank. It was the
6:25most incredible thing.
6:26They said, "We have decreased our spend
6:28on AWS and GCP. We now have six
6:32completely independent ground-up data
6:35centers that we built from scratch for
6:36ourselves." And I was like, "What is
6:38going on here?"
6:39>> Wow, we're back to the future. People
6:41standing up their own data centers,
6:42yeah.
6:43>> It's back to the future. So, I think if
6:45you look at this I didn't mean to have
6:47any commentary other than just observing
6:49what I think, but yeah.
6:51AI has a diffusion problem. It is too
6:53concentrated and ripping around in a
6:55small little ecosystem. We need to
6:56diffuse out, but as it diffuses out
6:58people are picking open source is what I
7:00would tell you.
7:00>> And we we we were early on that
7:02discussion 2 years ago here on the
7:04All-In pod. I will say Ken Griffin, and
7:06I didn't hear the full
7:08um quotes here
7:09um cuz I wasn't in the room where it
7:11happened. However,
7:13um
7:14based on the quote, I think he may um
7:17have something slightly wrong. When
7:18startups flood to a technology, that is
7:21a strong signal. Why? Startups and I
7:24invest in a hundred a year and we
7:26evaluate over 10,000 a year at my fund,
7:29they
7:30look for the cheapest, best, fastest
7:32edge they can get. What is the biggest
7:35edge they could get 20 15 20 years ago?
7:38It was cloud computing. They saved a lot
7:39of money. It it WeWork. They saved a lot
7:41of money. It was more efficient. And
7:43then it was Facebook ads. Facebook ads
7:44were mispriced, they were super cheap.
7:46They
7:47dumped all their VC money into Facebook
7:49ads. Then and now it's become AI. For
7:53the last 2 years, yes, they've been
7:53dumping into it. So, if you see startups
7:56rush to a technology or rush to a trend
7:58and start spending all their VC money on
8:00it, it means this little colony of two
8:04to 200-person companies have determined,
8:06and they're the most nimble, that this
8:08is the best solution. So, that's a tell
8:10that things are going in the right
8:11direction. The big, slow, cumbersome
8:14enterprises, they're the slowest to the
8:16new technology, and that's why they buy
8:18startups all the time. Speaking of
8:20buying startups all the time,
8:22my guy Nikesh is here from Palo Alto
8:25Networks to tell us
8:27if
8:28this is the end of days and all of our
8:30bank accounts and uh if Bitcoin is going
8:33to get hacked this weekend and buy more
8:36guns and chickens and eggs and drill
8:38well. Welcome back to the program,
8:40Nikesh Arora. How are you, sir?
8:43>> I'm doing great. Thank you for having
8:44me, guys.
8:45>> And you had a nice event last week,
8:47didn't you? A little nice event in Napa.
8:49>> I did. We had a bunch of CIOs, not just
8:52one bank CIO like Chamath said. But
8:53before we jump into the end of days, I I
8:56do want to say something to what you
8:57guys were talking about earlier. Yeah,
8:58but
8:59the open source versus the the sort of
9:02the hyperscalers providing info. So,
9:04then we had this peculiar problem. So,
9:06we'd like to embrace AI as quickly as
9:08many of the startups. And we've gone
9:10gone through all the understanding, all
9:12the analysis, open source versus cloud
9:13models, and we
9:15preference open source open weight
9:16because it allows us to modify the
9:18weights and train them. And something
9:20which most people don't talk about is it
9:22allows us to
9:24interrogate the chain of thought.
9:27Cuz over time what's going to happen is
9:28all of us enterprises are going to need
9:30explainability, auditability, and verify
9:32stuff.
9:32>> 100%. 100%.
9:34>> If I can't validate, it's not going to
9:35be okay to go to your boss and say, "I
9:36don't know what happened, but the IT
9:38guys put an AI model in there and it did
9:40it." That's not explainable to the
9:42regulator in financial services and
9:44health care. So, they want
9:45explainability, verifiability, and and,
9:48you know,
9:49uh auditability. And we can only get
9:51that with open source open weight
9:52models. We can interrogate chain of
9:53thoughts. That's kind of one reason
9:55we're gravitating towards them as
9:57opposed to getting caught into the
9:58closed uh race. The challenge though, I
10:01think you guys talked about and you I
10:03saw Chamath's chart, the challenge we
10:04have is is very hard to find a global
10:09resilient highly available inference
10:12provider outside of the hyperscaler
10:13community right now. So, if I'm a
10:14startup, great. I can go rent 10 MW in
10:17one data center in one part of the US
10:19and if it goes down for a few hours, I'm
10:22out of luck, but I'm a startup, so I'm
10:23okay. But I'm providing resilient
10:26services at 99.9%
10:29reliability, it's very hard for me to
10:31find a global inference provider who
10:34does not want to sell me their model at
10:35this point in time
10:37at the price that just makes it
10:39economically useful to use open source.
10:41>> Let me ask you a question though. Do you
10:42think that there's a world where Palo
10:44Alto Networks says one of two things.
10:46One is, you go to the hyperscalers and
10:48say, "Okay, listen guys,
10:50you run it for me, but I want the bare
10:51metal, I want it totally air-gapped, and
10:55it's an AWS facility, but I'll just
10:56lease it from you. Otherwise, just get
10:58out of the way so that I can have
10:59everything inside of it." Or do you do
11:02what this one big GSIB bank does and
11:05say, "I'm going to build my own inside
11:06the United States."
11:08Is there a world where do you think
11:09other companies go to that limit or do
11:12you think that's an extreme case that's
11:14probably not representative of the norm?
11:17>> So, this is the interesting debate from
11:18you know, six, eight years ago when I
11:19came to Palo Alto, I shut down all data
11:21centers. I said, "We cannot run a global
11:23data center sort of footprint at the
11:25scale, speed, and reliability as I could
11:27probably rent it from hyperscalers." And
11:29life was great and And what's happened
11:30is in the current sort of incarnation of
11:33where we are, compute is highly dear.
11:35It's very hard to find compute in the
11:37world, period.
11:39So,
11:40I don't I'm I'm a I'm a buyer. I'm a
11:42taker of terms, not a dictator of terms.
11:45It's very hard to go and dictate terms
11:46anywhere because yeah, sure, you can
11:48make any plans you want, but those plans
11:50are you know, you talk to Jensen, he
11:51can't give you chips till 2028. You talk
11:53to you know, when you get land power and
11:55shares and shells. So, it's very hard to
11:57go say yes, I can go build my own data
11:59centers, but it's kind of an empty
12:00threat. It's very hard to go build it. I
12:03would be resilient. And by the way, I
12:04don't run 50 data centers around the
12:06world. We're only running 10 to 20
12:07megawatts of inference. So,
12:09it's a it's a good conceptual sort of
12:12idea that I should go become
12:13independent, build my own inference, but
12:15most likely the answer is somewhere in
12:16the middle. I'm going to have to get
12:17bare metal, yes. Most likely I'm going
12:19to get from either a hyper scalar or a
12:21cousin of a hyper scalar,
12:23or somebody's going to be stitching, you
12:25know, sleeves of compute from around the
12:26world
12:27to meet my needs in the hope that I'm
12:29going to migrate to something more
12:31resilient and more robust over time.
12:33>> All right. So, Nikash, you have a hybrid
12:34solution down there. One second. I just
12:36want to complete that. You have a hybrid
12:37solution down
12:38>> stuff before we run out of time, you
12:39know.
12:40>> [laughter]
12:41>> Then you go and you basically get a
12:43bunch of open source models so that you
12:44can interrogate chain of thought tokens.
12:46Is that a representative solution for
12:48most of the Fortune 500 or Global 2000?
12:51>> I think we get there. I I expect at some
12:53point in time we will start seeing hyper
12:54scalars giving bare metal and letting us
12:56run open source models. They already do
12:57to some extent. So, that's probably
12:59what's going to happen because
13:00you know, it's a matter of the profit
13:02pool, right? It all depends on where you
13:03shift the profit flow. Today they all
13:04live in Nvidia and and the memory
13:06industry and they're being shifted
13:08towards the hyper scalars, people who
13:09have large compute available because
13:11it's it it at a point it'll become less
13:13about the chips, it'll become more about
13:15the availability of compute because I
13:16think demand is insatiable. There's
13:19enough demand, there's more demand than
13:21you and I can imagine. So, yes,
13:22>> Right.
13:22>> I understand the enterprise diffusion
13:23comment, but it's going to get diffused.
13:25The question is, even if you try if
13:27every company in the world today said,
13:29"I'm in. I'm in. I'm ready. I'm
13:30embracing AI. Let's go as fast as we
13:32can." There's not enough computer in the
13:34world to satisfy us.
13:35>> Right.
13:35>> We asked you to come on, Nikesh, not
13:37just because you're awesome and uh
13:40>> [laughter]
13:41>> you were you're now a bit uh
13:43based
13:44>> He's a bestie.
13:44>> He's a bestie. He's a bestie. He was
13:46very quiet under the radar.
13:47>> I think I would Nikesh every 48 hours.
13:50>> I know, David. I keep [laughter] We keep
13:51running into each other.
13:52>> I mean, every time there's a Knicks
13:54Warriors
13:55>> game, Nikesh invites me to sit with him.
13:56So, it's awesome.
13:57>> Um is there one [laughter]
13:59>> That's not true. I'm just manifesting.
14:01Um Scott Great States, I can manifest.
14:03Boy, can I manifest. Kid from Brooklyn's
14:05manifesting.
14:05>> As long as you want to go with my
14:0711-year-old, you're welcome anytime.
14:08>> Okay. I'm totally fine with that.
14:10>> He's always looking He's always looking
14:11for somebody to go with. So, you could
14:13be next, Jason.
14:14>> No, I'm I'm happy to babysit if uh with
14:16the courtside seats for a second row.
14:18>> He'll He'll babysit you.
14:19>> Exactly. Uh but there's a cyber panic
14:21going on. All seriousness, uh the last
14:23couple of days, AI found two major
14:25exploits that impact most of the
14:28internet. Here's what happened. PHP,
14:31which is an open source coding language
14:32powers about 70% of all websites, think
14:34WordPress, Wikipedia,
14:37is at risk. In August, security
14:38researchers using AI found two, one,
14:41two, critical exploits nicknamed PHP two
14:45shell. PHP security team gave this a
14:47severity rating of 9.2 out of 10.
14:50Basically, DEF CON 1 for these security
14:53teams. Since August, developers and
14:55companies have been working together on
14:56fixes. This is standard operating
14:58procedure. And the plan was to roll
15:00these out gracefully in one major patch
15:02on October 22nd, which is the day we're
15:04recording this very podcast. In the days
15:06leading up to the release, information
15:08about the exploits leaked. Nobody knows
15:10how it leaked. That's going to be part
15:12of the postmortem of this cyber attack
15:16week. So, PHP moved the release up just
15:19by a day to Wednesday, October 21st. and
15:23on Wednesday all hell broke loose. AWS
15:26reported a jump in hacking activity
15:28while Cloudflare CEO, friend of the pod,
15:30said they saw a millions of hacking
15:32attempts as well. A bunch of
15:34precautionary shutdowns happened.
15:36Governments across sites in Europe,
15:38France, Germany, Netherlands, Denmark,
15:40Finland, etc. They started taking
15:42websites down. Wikipedia took down some
15:44of its systems just to get them patched.
15:47In the United States, the NSA and the
15:49FBI put out a bunch of warnings and now
15:51everybody started panicking. People are
15:53starting to think that Bitcoin's going
15:54to get hacked, their website's going to
15:56get hacked, their bank accounts are
15:57going to get hacked. It's a bit of chaos
15:59and then there's a second sub-issue
16:01around agents accessing banks that
16:03dropped at the same time. Now you're in
16:05the security business, you've got lots
16:07of different products and services here.
16:09When did you first learn about this and
16:11how severe is it? Is this an acute issue
16:14or is this a harbinger of things to come
16:17and is this going to happen every 10
16:18days for the next year as these models
16:21have gotten incredibly powerful at what
16:23we call cyber, which really means cyber
16:26security, not just cyber punk, but cyber
16:28security.
16:30>> Both Jason, I think 10 days is too long.
16:32>> Okay.
16:33>> these things going to happen much faster
16:34than every 10 days.
16:37I think part of the challenge we're
16:38seeing is as we've been talking about
16:39technology debt that we've all have to
16:42go pay sooner than later. And what's
16:44happened is people have not paid
16:45attention to the idea that they have to
16:46go fix their stuff. And that to customer
16:49once they were about And I talked to the
16:51CIO this morning, they had a PHP
16:52problem. They have a bunch of websites.
16:54Some are running the most advanced
16:56pieces of software so they can put the
16:57patch. Some are running older versions
16:59which they can't patch. They've got to
17:00bring them down and go build them up.
17:02But you know what? They never did
17:04anything about the fact they were
17:05running older versions because it wasn't
17:06important. So now they have to go to go
17:09take those websites down.
17:10They're going to go
17:12upgrade them, put the patch in, bring
17:14them back up. We found out about this a
17:16few days ago.
17:18We scanned everything in our
17:19infrastructure to make sure that there
17:20were no issues. We made sure that our
17:22products had patches available for our
17:24customers who deploy them. So, there's
17:25two ways to solve this problem, right?
17:27One, you can go in and patch the
17:29vulnerability. You can go in, find the
17:30code, patch what has been given to you
17:32by PHP, and make it work. The other is
17:35you can actually put it in a perimeter
17:37defense product like ours or Cloudflare
17:39has one or other firewalls have one. You
17:41put the patch there. It's a virtual
17:43patch, so any attacker trying to access
17:45that vulnerability gets stopped at the
17:46perimeter because we know what they're
17:47looking for. There's two ways to solve
17:49this.
17:49>> to a CDN, content delivery network
17:51first, and that acts as a first line of
17:53defense cuz they're not getting to the
17:54actual servers. But, if they do get to
17:56the actual servers, my understanding is
17:58they can run code on these servers, and
18:01there were some security certificates
18:03that were incredibly brittle and easy to
18:07hack.
18:07>> The closed models have
18:10basically said, "Hey, these things are
18:12cyber risks, and so we need to be very
18:14thoughtful." I know Anthropic did that
18:16with Mythos, where they brought you in
18:19and a bunch of other people and like 50
18:20companies in and said, "Okay, now go
18:22harden your systems."
18:24Were the closed models helpful here?
18:26Were they the ones that helped you find
18:27these CVEs? Bring us behind the curtain.
18:30>> The closed models are can be helpful
18:32because both OpenAI and Anthropic, and
18:34most likely Google as well, have very
18:36powerful models now, which can go, if
18:39you give them enough compute, enough
18:40persistence, they're going to go find
18:42vulnerabilities. They're going to even
18:43chain vulnerabilities to find an attack
18:45bot, right? If they can solve neighbor's
18:46stokes, they can chain four
18:48vulnerabilities. And if you give them
18:50enough immense compute, they will try
18:51every way possible to be able to get
18:53Actually, the PHP problem was solved by
18:55persistent compute. This problem was
18:58identified a few years ago, but nobody
19:00thought it could be solved because it
19:01required too much of coordination. So,
19:02what happened was if you throw computer
19:04at the problem, it persisted and it
19:06found a way to crack that. Once it found
19:09a way way crack that and they published
19:11it or got to reverse engineer it. Every
19:13hacker knows now how to go after it
19:14because the challenge you have with
19:15patches is hackers are smart. They look
19:17at a patch, they see what the patch is
19:19trying to do, they reverse engineer from
19:20there, and say, "Okay, if I follow the
19:23reverse engineer path, if somebody's not
19:24patched, I can break in."
19:26>> It seems like a bunch of the hullabaloo
19:27today that's being drummed up, and this
19:29seems like it's going to change every
19:31day,
19:32is all about these bank accounts getting
19:34accessed. Is that Do you think primarily
19:37through kind of the front end, through a
19:38website using an agent that's logging in
19:40on a browser-based system on behalf of
19:42its user?
19:43Because it seems like a lot of this may
19:45be coming from news or some of these
19:47other sort of like personal agents that
19:49are logging into your bank account to
19:51get information about transactions and
19:53what you're doing.
19:54And the banks are seeing all this
19:55activity and they're like, "Oh my gosh,
19:57is this authorized? Is it not
19:58authorized?" The traditional way that
20:00banks have provided access to account
20:03data has been through APIs using
20:05verified applications,
20:07where I as a user, as a consumer, would
20:09log in to my,
20:11you know, mint.com account, and it had a
20:13a legal access point to
20:16uh to Amex or to Bank of America. It can
20:18do that.
20:19>> there.
20:19>> Yeah. Plaid is an intermediary.
20:21>> Yes.
20:22>> And so, there's a bunch of these sort of
20:23intermediaries that then provide that
20:25capability to apps. Is that ultimately
20:28going to lead to a world where the banks
20:30are all going to just turn off website
20:31access and say no one can log into a
20:33website because technically agents could
20:35do it and they're not able to kind of
20:37manage that. Um how do you think this is
20:39going to evolve as it relates to
20:40consumer financial services? Cuz I
20:42actually think this is going to be a big
20:43part of the press cycle over the next
20:45week or two.
20:46If you have an answer.
20:48>> Yeah, I think two different things,
20:49David. I think the reason banks have
20:50found
20:51uh
20:52that in sort of unauthorized access to
20:54their accounts is possibly because of
20:56PHP problem. Because their websites are
20:59hackable and they've been able to get
21:00through them and get to credentials. I
21:02don't think there are two different
21:04problems. One is unauthorized access
21:06happening because people have not been
21:07able these sites have become hackable. I
21:10think the second problem you're
21:11mentioning is the personal agent
21:12problem. So far, most personal agents
21:15are using your credentials to log in.
21:17So it becomes hard for a bank to decide
21:20whether it's actually you logging in or
21:22is it your Muse agent who has your
21:25credentials logging on your behalf. They
21:27cannot distinguish between your traffic
21:29or agent traffic. They're both logging
21:30in from the website. They're basically
21:32logging in as you and your browser and
21:34trying to conduct a transaction. Now, as
21:36you know, there's this technological
21:37capture which tries to see that it's not
21:39a tech it's not like an agent that's
21:41sort of or a it's a sort of a bot or an
21:43agent accessing the account instead of
21:46you. Naturally, there are ways to you
21:49know, validate capture with the eye so
21:50you don't even actually have to have the
21:52human on the other end. So that's also
21:53>> good at it.
21:54>> It's pretty good at it. It's actually
21:56sometimes I have to squint to see what
21:57the hell is it I want to type.
21:58>> Yeah, get click all the bicycles in this
22:00picture. [laughter] Like it takes you
22:01three times at our age with our
22:02eyesight. But the interesting thing here
22:04is that there is a hybrid as well
22:06Freeburg, which is your
22:08um
22:09your Muse agent, your Grock bot is
22:12running a virtual browser. It will pop
22:14up the virtual browser and say, "Hey, do
22:16this captcha." and then take over. So
22:18Nikesh, is this fear, uncertainty, and
22:20doubt, the old FUD by the banks to try
22:23to neuter or sandbag
22:26these agents from getting the the front
22:30door to banking cuz then they could just
22:31route to a bank that's cheaper, better,
22:33has better rates on behalf of the agent.
22:36>> I mean cuz there's I think that would be
22:38giving them a lot of credit.
22:40>> Okay.
22:40>> I think I think I think right now
22:42they're very diplomatic. The fundamental
22:44challenge they have is the liability
22:45problem. Look,
22:47it's kind of like if you call customer
22:49support at your favorite bank and say,
22:51"I don't know how $50,000 vanished from
22:53my account. I never did this."
22:54>> Yes.
22:56>> Right?
22:56>> What's the Spirit of the Rhino? I don't
22:58know what this is.
23:00>> [laughter]
23:02>> And and the challenge they have is they
23:03keep
23:04keep
23:05>> Keep going, Nikesh. Don't be distracted.
23:06Don't be distracted.
23:08>> The challenge is we're number one for a
23:09reason.
23:11>> The challenge is you you've given your
23:13credentials to an agent. You didn't even
23:15realize that it got your credential.
23:17It's living on your browser, living on
23:18your laptop. It's going to get
23:20innumerable amount of credentials. And I
23:22think the biggest problem we're going to
23:23have is who owns the liability and who
23:26has accountability.
23:30>> Banks say, "Do not I do not allow you to
23:32interact with me with your agent,
23:33period." Right? They're yelling that
23:35disclaimer from the top of the roof
23:37saying, "It's not going to be my fault."
23:38Look, banks had to go Don't forget, it's
23:41not even the probably the cost. It's
23:43like the liability secondary. Remember,
23:45banks We all know this at credit cards
23:48below a certain amount of money, they
23:49don't want to dispute it. They pay you
23:50because the cost of the customer support
23:52call is higher than than they're going
23:54to have to deal with. Now, if you start
23:55having tons and tons of these agents
23:57doing and you start calling your bank
23:59for explainability, you're going to
24:00create huge amount of cost. I think I
24:02think that's where they're going, to be
24:03honest.
24:03>> Okay, you have a unique You have a
24:05unique seat, Nikesh, because you you
24:07talk to the people who are
24:09being hacked, have been hacked, and you
24:11help them prevent being hacked with with
24:13this all these suite of services over
24:14Palo Alto Networks.
24:16So, I want to take you through a couple
24:18of the theories people have about AI's
24:20impact on security. You tell me if you
24:22agree with them or not. The first one is
24:25this is going to be largely an issue of
24:27who has the most compute. And because
24:30the white hats have access to the most
24:32amount of compute, they will be able to
24:34out compute
24:35compute the uh black hats, in other
24:38words, the hackers. Is that a true
24:41statement or is that not yet a true
24:43statement and maybe it will be in the
24:44future. In other words,
24:46Palo Alto Networks or or or OpenAI,
24:48whoever it happens to be, could fire up
24:50so much compute to defend systems
24:53that a bunch of hackers in North Korea,
24:55Russia, China, etc. just don't have the
24:56the firepower and they'll be outgunned.
24:58True, not true, when will it be true?
25:01>> True, not true today because the
25:04underlying systems are not are not
25:06architected or orchestrated in such a
25:08way that we can launch these, let's call
25:10it defensive swarms because everything
25:12is set up in disparate
25:14situations, right? Like if everything
25:16was true, then by definition you could
25:18send out a bunch of
25:20agents to go patch PHP if you wanted. A
25:22company's got 5,000 websites and guess
25:24what? Their priority is going to catalog
25:26each one of them because they're they're
25:28actually aggregating everything, putting
25:29it together, understood their so So, I
25:31think we need to get our infrastructure
25:32to base level of preparedness, which is
25:34the tech the technological debt lives.
25:37>> Let me take you to the next thesis and
25:39theory people have here, which is
25:41persistent computing
25:44reviewing code and people taking
25:46proactive measures more seriously is
25:49going to solve this problem and right
25:51now we just have this interim period
25:54where people have accumulated technical
25:56debt you've talked about, but AI is
25:58going to solve that problem cuz you'll
25:59have persistent tools from companies
26:02like yours and others or in-house ones
26:04that are constantly banging on systems
26:07to find these before hackers get to
26:08them. Is that going to be the end game
26:10here? True, not true, short-term true,
26:12long-term true, take us through it.
26:14>> True. Today, if you can see almost every
26:16large company in the enterprise space is
26:19out there busy using Metasploit
26:20equivalent to try and penetrate their
26:22own defenses to see where the holes are
26:24and they're trying to patch them. The
26:26challenge as you'll appreciate is that
26:27the next model is going to be more
26:28powerful. It's going to find more ways
26:30to get into your infrastructure. So,
26:31you're going to be playing this cat and
26:32mouse game and you're constantly
26:34patching. I think there's a slightly
26:35different sort of version of this,
26:37Jason, where for a second let's assume
26:40that you will get breached. Okay? Cuz
26:42there are so many vulnerabilities out
26:43there you can spend a lot of money
26:44patching everything. The question
26:45becomes less about how I got breached
26:48and which vulnerability is going to let
26:49people through. The question is how
26:51quickly can I find them and how quickly
26:53can I eject them?
26:55>> Mhm. Okay.
26:56>> our advice to people is
26:58build the infrastructure in such a way,
27:00which is possible today, where you're
27:02aggregating all the signals, all the
27:04data in such a way that you can actually
27:06look for anomalous behavior with that
27:08white hat compute you talked about,
27:09right? Because I have more compute. Go
27:10ahead.
27:11>> Yeah, and then the third one is the
27:14frontier models are causing a lot of
27:16this chaos because they
27:20are gatekeeping the most powerful
27:22solutions
27:24to do cyber defense and that the open
27:28source community is giving more
27:30functionality to the white hats, to the
27:32security teams, to the CISOs to actually
27:34solve this and that the frontier models
27:37need to change how they deploy this
27:40technology in some way. Are they causing
27:42part of this problem by gatekeeping this
27:43technology and not giving the raw power
27:47to the white hats?
27:49>> What you said earlier about is a
27:50diplomatic answer about giving too much
27:52credit.
27:53>> Yes.
27:53>> I think I think your theory is giving
27:55too much credit. Again, I think that
27:57they You remember majority of them are
27:59researchers to start with. They're
28:01sitting there and saying, "Wow, this is
28:02extremely powerful. What do I do with
28:04it?"
28:04And they're talking to us and saying,
28:06"What do I do with it?" Now, remember
28:07these models that are so far ahead,
28:09they're still trying to figure out how
28:10to release them, how to package them,
28:12how much computer these guys going to
28:13take. And they're all computer
28:15constrained, right? So, they're all rate
28:17limiting all of us, even though I have
28:18access to the most recent model, which
28:20is available publicly, I'm still rate
28:22limited.
28:23Right? So, there's tons of issues that
28:25are going on. So, I think they're not
28:26causing the problem. They're just trying
28:27to grapple with the fact that these
28:29things are extremely powerful. What they
28:31would like they would like to figure out
28:33how to monetize it.
28:34>> Mhm. Because they want [clears throat]
28:35to compete with you?
28:37>> No, they they just need money to grow.
28:38They're like they're all listening to
28:39your conversations about compute needs
28:41and where the money's going to come
28:42from. They're like they're not trying to
28:43compete. They're just trying to go after
28:44profit pools because they want a lot of
28:45money cuz they got to go give it back
28:46to, you know, Jensen.
28:48>> So, that's their main issue. Yeah.
28:50>> Let me ask you a a hypothetical
28:52question. I'll ask all of you.
28:54I use one of these
28:57agentic services. I'm a consumer. I give
28:59it my credentials to my bank account.
29:01Let's use your example. $50,000 goes
29:03missing.
29:05I don't know what happened. Maybe I
29:07escalate it to the agent I use, you
29:10know,
29:11Instinct, Muse, dot
29:14or I copy of A.
29:16Everybody's finger pointing at the other
29:18one. Maybe there is a prompt injection.
29:21Who knows? Maybe the agent just went
29:23rogue and transferred the 50K. Let's
29:25just assume that that's possible. 5K,
29:26you know, cuz it's under some threshold.
29:29Who is responsible? Okay, Friedberg, who
29:31is responsible?
29:33And who should be responsible? Who is
29:35responsible? Who should be responsible?
29:37>> Well, the the terms of service of the
29:39bank
29:40say you can't use an agent to access the
29:43bank account through the website, right?
29:45>> Does it say that now?
29:46>> Yes.
29:47>> Yeah.
29:48>> So, given that that's the terms of
29:50service,
29:51if the agent hasn't built in a capacity
29:54to recognize that and stop that access,
29:56then it is a software liability issue
29:58with the provider of the agent, I
29:59believe.
30:01Unless it alerts you and says this
30:02violates the terms of service, do you
30:04want to override my block, and do you
30:07want to do it? Now the liability shifted
30:09to you as a user.
30:10>> And we have evidence of this, Chamath.
30:12Um not in banking, but in e-commerce.
30:15Amazon has blocked these, and
30:17subsequently Grok now will tell you in
30:20my Amazon shopping bot says, "I can no
30:22longer check out for you, but I can put
30:24it into your shopping cart. Tell me
30:27which address, which card you want, and
30:28I'll tee it all up for you, but I won't
30:30hit the buy button." So,
30:33it's being
30:35adjudicated in real time right now, and
30:37I think that's probably where we wind
30:39up.
30:40>> Nikesh, what do you think?
30:42>> You are.
30:43>> Yeah.
30:44>> I think I think come on, we all have to
30:45take responsibility to empowering agents
30:48and giving them our credentials. You
30:49can't say I gave my credit
30:51>> It's the only solution.
30:52>> I gave my credit card to Chamath. He
30:53went and burnt $20,000 on it. Who's
30:55responsible? The bank? No way. Is it
30:57Chamath? Yeah, he is, but I gave it to
31:00him.
31:00>> Yeah. But what if the agent Chamath uh
31:03says, "I have been programmed to be
31:05maximally helpful." You know, these
31:07sycophantic agents, and it goes and
31:09says, "I found a way. You know, I found
31:11a way. Actually, if I fire up this type
31:14of browser and I simulate a human
31:16behavior, I think I can get this done."
31:18That's what a lot of these agents will
31:20do is they'll try to please you, and
31:22they find that then that to me would
31:23lean towards the platform providing the
31:25agent No, Jason, Jason, no, that's you.
31:28It's your responsibility.
31:30>> Yes.
31:30There's a word in agent, it's called
31:33agency. Somebody has to give the agent
31:36agency.
31:37>> Okay.
31:37>> Who's giving the agent agency? You or
31:40Facebook? I think you're giving agency
31:42to act on your behalf. And if you're the
31:44one who's giving agency, you have the
31:45liability.
31:46>> Yeah, I think as a practical matter,
31:48it's the only scalable solution because
31:50what will happen when these lawsuits
31:52happen is everybody's incentive,
31:55everybody meaning the corporate
31:56incentive. B of A will point to
31:58Facebook, Facebook will point to the
31:59user. It's the only way that this thing
32:01continues to scale like this.
32:02>> but Chamath, there's an exception case.
32:05And I think Jason said it in his words,
32:07you said it, but we didn't dwell on it.
32:08It's called the rogue agent. Now, that's
32:11a product liability issue.
32:13If a product does not If a product does
32:14not behave like it's intended to behave,
32:17right, then there is a platform
32:19liability. In my mind, it says, "Wait,
32:21you said this is what it does." And
32:22today, you buy any product in the world,
32:24if the product doesn't do what it's
32:25supposed as advertised,
32:27uh it goes rogue on you, then it's the
32:28person who built the product, it's their
32:30fault. And this is where the problem is.
32:31>> dangerous territory, which is, "Hey, it
32:33can do 140 It's capable of doing 140
32:36tasks.
32:38You know, just maybe not really. Let me
32:40ask you a in in the weeds question. We
32:42incubated a company Obliteration AI and
32:46what this company is doing
32:47>> Nice name.
32:48>> Yeah, and obliterated models maybe
32:50people don't know about but they're
32:51going to understand what these are.
32:54This company a little bit
32:55controversially provides
32:58models
33:00that have no constraints, no guardrails
33:02and it's your responsibility when you
33:05have a red team, a team that's going out
33:07and banging on different vulnerabilities
33:09to build up those. What are your
33:11thoughts on this new
33:13somewhat controversial Wait, before you
33:17go.
33:17>> Let me just say I before you go. I find
33:19Obliteration unbelievably scary. That is
33:21a vector of risk that is so
33:25unbelievably underappreciated right now
33:27because to your point Jason, it is a
33:28very long tail thing. Not enough people
33:30understand what it is but I find this
33:33probably the thing more than alignment
33:36that I think
33:37people should really understand and
33:38think about. Sorry Nikesh, over to you.
33:40>> Yeah, go ahead Nikesh, tell us what you
33:41think of these services.
33:42>> So, I think this is where this is the
33:45dark side of AI. And what I mean by that
33:48is, you know, we've had all these
33:48debates at the White House, many of you
33:50were present. We talked about
33:51responsibility and alignment and making
33:53sure we tested them before. The problem
33:55is
33:57you're you're dependent on the weakest
33:59link. If this there's a company which is
34:01going to give you the raw power of AI
34:04without any guardrails, without any
34:05controls that are either built in or
34:07can't be built in, then you're basically
34:09handling a nuclear sort of bomb to the
34:11hands of anybody who wants it because
34:12they're going to afford it. That's the
34:13problem because at one level we're
34:15saying we need global treaties to manage
34:17the debasing of AI or development of AI
34:19and just for good. On the other hand,
34:20you're saying we can have a company
34:22which can allow you to have this most
34:23powerful thing with no control, no
34:25guardrails and no governance.
34:28It's like saying, "Hey, there's a here's
34:30a DIY aspirin kit.
34:34Have a good time. You can build your own
34:35medication. Have a good time. Sometimes
34:37I get it wrong, but don't worry about
34:38it.
34:39>> Or just giving somebody some really
34:41powerful power tools to build things
34:44that make things more secure, like build
34:46a wall.
34:47>> I don't understand what the business
34:48model of your company is. Like your your
34:50company's like basically like inducing
34:52product liability as a service.
34:54>> it to white hats under very constrained
34:58No, it's providing this to white hats
35:00and corporations. They they share the
35:02same
35:03logo list as the cash and Palo Alto
35:05networks. There are people who want
35:07access to these in order to do proper
35:09testing and make things more rigid. So
35:11you do have a higher authority of like
35:13not providing this for free like Cloud
35:15was doing.
35:16>> But Jason, are you telling me that they
35:18they they underwrite that these law that
35:20when you put these things in
35:21infrastructure and do testing, they're
35:22not going to go rogue on you?
35:24>> Well, it's your responsibility to to
35:26manage the model as opposed to giving it
35:28to in this Yeah, if you were the NSA or
35:31if you were
35:32>> How is that How is that How is that
35:34different How is that different from me
35:36going and taking open rate model and
35:37training it to my needs?
35:39>> You could do that. This is just a a
35:41place that has the infrastructure to do
35:42it. So providing that infrastructure you
35:44talked about earlier like who's going to
35:46serve it and that kind of thing. And and
35:47and making, you know, just essentially
35:49making it more efficient for you to do
35:51this with more constraints.
35:53But you know, and letting you take the
35:54liability, but it's a very interesting
35:56place. Why does technical debt
35:58build up in historically have we just
36:01under and a lot of these causes of these
36:04cyber scares is this just all because
36:07corporations
36:08deprioritize security and technical debt
36:11and that we just need to change that in
36:13this new world and that will essentially
36:15solve the problem if we prioritize
36:16safety first?
36:18>> Nobody wants to work on this stuff. I
36:19mean, the problem is when you ship
36:20something into prod, everybody wants to
36:22work on like the next greatest thing.
36:24Nobody really thinks about like
36:25maintenance as like a sexy idea. That's
36:28one thing. Second thing is if you
36:30actually deploy it inside of your own
36:31virtual private cloud, the problem is
36:33the software vendors don't have this one
36:35version that's constantly updated. You
36:37have a drift problem where you have a
36:38version, Nikesh has a version, Free
36:40Brick has a version, they're all out of
36:42sync. And so now, you know, you as a
36:43software vendor are maintaining 19
36:45versions because everybody has their own
36:47pace and cadence for upgrades and
36:49migrations.
36:50Then you have the fact that technologies
36:52are constantly changing and the
36:53underlying way in which things
36:55can happen change so rapidly, but the
36:58way in which you've done it represents a
36:59sunk cost and you can't just rip and
37:01replace this stuff.
37:02And you multiply that by millions of
37:04companies and
37:06hundreds of trillions of lines of code
37:07and this is why it is like it is.
37:09>> And to that also
37:10add to that the incentive alignment
37:12problem.
37:13The CIOs are not incentivized to
37:14maintain the most modern infrastructure.
37:16They're incentivized to keep costs low
37:18and they're incentivized to keep
37:19everything available all the time. So,
37:21every time you want to upgrade something
37:23it introduces risk of blowing up or
37:24going down. It's like, I don't want to
37:26go I don't want to go change everything
37:27all the time because all my users
37:29complain because you know what, every
37:30time you upgrade something, change
37:31something, it causes friction in the
37:33system and then it costs money. When was
37:34the last time your your CIO showed up
37:36and said, "Yeah, great. Have some more
37:38money and have a bigger budget because
37:39of the new toy."
37:40>> this is a real tragedy and and you'll
37:43see this with like say the CIA or the
37:45FBI.
37:46If there's no terrorist attacks, people
37:48are like, "We're spending too much
37:49money." And it's like, "Or
37:51we're spending the right amount of money
37:53because we haven't had a 9/11 again."
37:55And when they come to us and say, "Hey,
37:56we need more because based on what we're
37:58seeing,
37:59you know, there could be another God
38:01forbid 9/11." And
38:02then Congress or whoever is, you know,
38:04got the power of the purse at the given
38:05point in time says, "Yeah, you know,
38:07we're not going to give you more money
38:07because we haven't had one in 20 years."
38:10That that that is really the core of
38:11this problem is they only get credit
38:13>> to change.
38:14Cuz we're going to have a bunch of 9/11s
38:16in cyber. Yeah.
38:17>> Yes, Scott, the phones keep ringing
38:19every day and the conversation always is
38:20now, "How do I get out of this?" Because
38:22I'm tired of once getting punched in the
38:24face every morning when I wake up.
38:26>> [laughter]
38:27>> When people really appreciate the the
38:30the CISO and and and those groups, uh
38:32when these things and the systems come
38:34back online and the system's been
38:35hacked, it's going to be brutal the next
38:37uh year or so, yeah? It's going to be a
38:39big transition time.
38:40>> Nikesh.
38:41>> I think I think, you know, the CISOs
38:43could become the next CEOs. Think about
38:45it. If you believe in three years from
38:46now, every business function of yours
38:49wants to transform themselves with the
38:50AI. Let's go to the positive side,
38:52right? And my my CFO never wanted AI.
38:54He's saying, "Dude,
38:55I could do all this and all financial
38:57analysis. You just appoint some people
38:59to my team. My marketing people want
39:01them. My HR people want them." So,
39:02almost every function wants more AI
39:04people, which has never happened
39:05historically. Everybody's wanted
39:07Everybody wants AI people. And that's
39:08what's going to happen. Every function
39:10is, instead of waiting for AI or
39:12technology people, they all want the
39:14CISO to come and help them. Then who's
39:16the person running the company?
39:17>> What percentage do enterprises spend on
39:19security of their total spend, and what
39:23should it be ideally? I'm just curious.
39:24I mean, we have no insight here.
39:26>> The insight is that typically, if your
39:28IT budget, the cost is somewhere in the
39:314 to 10% range for cyber, okay? And 10%
39:35is typically where you'll find financial
39:36services, the federal government,
39:38because they want to make sure things
39:39are secure and robust. Uh 4% can either
39:42be cloud people, because that's where
39:44the hyperscalers would take care of
39:46infrastructure security for you. You
39:47just have to worry about application and
39:48software security. So, it ends up at 4
39:50to 5%. That's the range. You can get it
39:524 to 5% to 8 to 10%.
39:54>> Yeah.
39:55And that's going to have [clears throat]
39:56to go up. And there's
39:58uh there's an
39:59an issue, I think, as you're pointing
40:01out. If there is AI talent in the
40:02building, we got to build new products
40:04and services for our customers. Uh YOLO.
40:08>> Jason, Jason, there's a whole It's kind
40:10of like, you know, I I was joking the
40:11other day, but it's it's true. Like,
40:12look, there's a whole new category,
40:14which we even don't know how to put our
40:16arms around, AI. Like nobody knows. Like
40:18imagine if I said, "Great. You know
40:20what? Go deploy AI in my company and
40:22let's make sure AI is secure." Well, the
40:24boss says, "Whatever you do, make sure
40:26you can interrogate agents through chain
40:27of thought."
40:28That's easy to say. That's hard to do.
40:31How do [clears throat] I find all the
40:32agents that are running it? So, it's
40:33like we haven't built the AI security
40:35harnesses that are needed. I joke like,
40:36"You know, when they invented planes,
40:38when the Wright brothers were out there,
40:39nobody said, 'Let's go create the TSA.
40:41This is about time. We got to build
40:42security now because it's eventually
40:43going to happen.'" So, people are busy
40:45enthusiastically deploying AI, embracing
40:48AI. You saw I don't know if you saw this
40:49morning that, you know, if somebody some
40:51researcher found that if you send a news
40:53conversation to somebody else, you can
40:55extract private information out of it.
40:57So, now they got to go back to the
40:58drawing board and patch it. So,
41:00everybody's experimenting, news is being
41:02found out. We need to build all the
41:03security harnesses for AI and that's not
41:05even started yet.
41:06>> Bernie Sanders goes on Twitter today and
41:08says, "Look at all the craziness. This
41:10is insane.
41:11Stop AI now."
41:13So, you're in a room with Bernie
41:14Sanders. Someone says, "Nikesh, we need
41:16you to save America. Go meet Bernie
41:18Sanders." [laughter]
41:19You get on the plane. You go over there.
41:21You get in the room.
41:23And the guy And the guy says, "We got to
41:24stop AI now. What is your
41:27response? How are you going to solve
41:28this problem?
41:29Nikesh, [laughter]
41:30Aurora, how are you going to solve it?"
41:33>> I said, "Sounds like a great idea. Let's
41:35go. Why don't you start?"
41:37I'm right behind you. It's like I got
41:39it. And then you know, with all due
41:40respect, I
41:42it's interesting.
41:43This is the first time we have dwelled
41:45so much on the dark side of the
41:46technology history and I don't know why
41:48it is. Maybe you guys can explain to me.
41:50We spend so much time obsessing with the
41:52dark side of technology and my thesis is
41:54it started with the godfathers or
41:57whatever you want to call them, the the
41:58builders of this technology. They've
42:00over-indexed and dwelled harder on the
42:03dark side of this, causing this
42:04narrative. I think we need a narrative
42:05shift. And think about it. Like we we
42:07only hear about all the bad stuff that
42:10that this technology does. Oh my god, I
42:12got missiles. Wait, I got something
42:13worse. My guy's jumped over. Wait, mine
42:15jumped over 6 months ago. Oh, mine are
42:17worse. It's like, dude, this is not a
42:19good narrative. This is a race to the
42:20bottom from a narrative perspective.
42:23>> [laughter]
42:23>> Nikesh, why then do you think Bernie
42:25Sanders and others are seizing on the
42:27moment? I know we've talked about this
42:28before, but like what's going on that's
42:31motivating Bernie Sanders to be so vocal
42:33about stop AI, stop AI? Is it like a
42:36power struggle that they view there to
42:38be an emerging system of power that they
42:43don't control, that they don't have a
42:44hold on, that diminishes their power? Or
42:47is there something kind of
42:49more simple than that?
42:51>> David, this is what I phone a friend.
42:52You guys are much more competent as
42:54answering these questions than I am.
42:55Jason and Chamath are probably better
42:57better equipped to answer these
42:58questions
42:59than me.
43:00>> Yeah.
43:01Yeah.
43:02>> Yeah. It's a It's a populist issue. If
43:06you if if the people at the labs
43:09are going to tell you half of
43:11white-collar jobs are going away and all
43:13the you know, drivers are going away and
43:15it's a 10% of everybody dying. I mean,
43:17what a great issue to align yourself
43:19with if you're a populist, whether
43:21you're America first populist on the
43:23right or a liberal on the left. It's
43:25just a great way to get people's
43:26attention and get votes and you know,
43:28maybe it helps you win another election.
43:31Nikesh, we appreciate you you're busy
43:33and this is like the busiest time you
43:34could possibly have, I think, right?
43:36What is your day like? Have you slept? I
43:38mean, are you just getting woken up by
43:40clients at like 1:00 a.m. when they're
43:42like, "Hey, I need a I need front
43:44forward forward deployed engineers from
43:46you?" I mean, what's going on in your
43:47life?
43:48>> Yeah. I I woke up at 3:00. I flew to the
43:50East Coast. After this, I'm going to get
43:51back on the plane and come back home
43:52because I got to make it time for dinner
43:54with customers.
43:55>> Crazy. I mean, it is just a crazy time
43:57for you, man. You know sleep. I know.
43:59Yeah, I love it. Yeah, I mean, this is
44:00the time to grind. All right.
44:01>> I love that.
44:02>> Uh thanks so much for Nikesh for joining
44:03us. We got to get you on for a full
44:04episode and talk about a range of
44:06issues.
44:06>> Thanks, Nikesh.
44:07>> You're so based. Follow and follow
44:08Nikesh on X. He does it daily.
44:10[laughter] He does one tweet per day and
44:14it's based in Nikesh.
44:15>> All right. Thank you, Jason.
44:17>> All right. Next up on the docket.
44:20Is this the fall of France, gentlemen?
44:21We covered this 2 weeks ago. Situation
44:23in France is
44:25my god. Well, the place is on fire. It's
44:26gotten worse. Over a half a million
44:28people have taken part in the protests.
44:31They've now spread to Italy, Spain,
44:33Albania, and a couple of other
44:35regions. Over the weekend, France saw
44:37its worst riots in decades. And they
44:39love to riot. So, that's a high water
44:40mark. Numbers pretty striking. New York
44:43Times data [laughter] by the French guy.
44:45They're good at it.
44:46This is their wheelhouse.
44:48>> Pro pro-rioters. Pro-rioters.
44:50>> They're professional rioters, exactly.
44:52It's kind of their number one pastime.
44:54For us, it's like football. For them,
44:55it's riot. Over 3,000 businesses,
44:58Chamath, have been damaged. You've seen
45:00the videos, obviously, trending all over
45:02the place. Hundreds of millions in
45:03losses.
45:044,000 police officers injured. 2,500
45:08people have now been arrested. 70% of
45:09them are adult men. Now, you remember
45:11this was a high school protest because
45:14high schools
45:15were dilapidated and underfunded. And
45:18that's what the origin of the protest
45:20was. Now, it's just become complete
45:21civil unrest
45:23and looting.
45:24Multiple historic government buildings,
45:27tragically, have been burned down.
45:28Notably, a 143-year-old
45:31city hall
45:33in a town 250 miles south of Paris,
45:35Limoges. Limoges. Beautiful name.
45:38Beautiful building. Absolutely
45:40heartbreaking to watch it collapse. For
45:42what reason?
45:44So, they could upgrade the schools in
45:46France. Seems like there's a lot more
45:48going on here. France has finally
45:50implemented some curfew some curfews.
45:54Which I don't know how you don't do that
45:55day one. New York Times quote, "France
45:57France's government has accused
45:59ultra-left militants and opportunist
46:02looters of causing the violence.
46:04President Trump and Vice President Vance
46:06have called this a third world
46:08immigration and welfare state issue.
46:11Here's what France's Justice Minister
46:13said. Rioters include, quote, "People
46:15from immigrant backgrounds, people who
46:17are foreigners, but there are also
46:19people who are fully integrated and who
46:21are French." Our partner, Polymarket,
46:23says 18% chance France declares a state
46:26of emergency by the end of the year.
46:29Same time,
46:30investors are dumping French debt.
46:33Friedberg, France's 10-year has
46:35surpassed 5% this week, highest since
46:382002. On the other hand, and uh this is
46:41an ongoing issue here in the United
46:43States,
46:4410-year dropped just under 5% for the
46:47first time
46:49in a month.
46:50Wall Street Journal says this might be
46:51because people are fleeing EU debt and
46:53buying the US debt. Polymarket, our
46:56long-term partner here at All-In, says
46:5891% chance Moody's or S&P downgrades
47:01France by October 31st. So, let's stop
47:04there.
47:05Chamath, when you see this and you spend
47:07your time in Italy in the summers,
47:09obviously, we talk about that here on
47:11the pod.
47:12Uh you see it happening in Italy.
47:14Uh and what's your what's your takeaway
47:15here from what we're seeing?
47:17>> It's heartbreaking
47:19to see what happens. Let's just be very
47:20clear. I think that these people are
47:22absolute criminals. They're all
47:23terrorists and they should all be
47:26arrested. And I think that this shows
47:29sort of like what is at the end of this
47:32quote-unquote empathy that all of these
47:34governments have been espousing.
47:37It's just a disguised way of promoting
47:41urban chaos.
47:43And uh
47:45I don't know how a police officer thinks
47:47it makes any sense
47:49for him to be deployed into this melee
47:52to do what effectively like protect a
47:55Christian Dior store in the 7th
47:57arrondissement of Paris. Like it is so
48:00stupid. So, these folks have created the
48:03tinderbox and the conditions for this to
48:05happen.
48:06And they've done it because they won't
48:08even acknowledge what this is. Even when
48:10Macron went on state television,
48:13he tried to
48:15talk about this in terms of
48:17the suffering that these people must
48:19have. When you look at these pictures,
48:21these are not people that are suffering.
48:22These are people that are uh enacting
48:25chaos and violence and terrorism,
48:27domestic terrorism. And I think you need
48:29to respond to these people with an
48:31incredibly heavy hand.
48:33And then you need to fix the much
48:35broader problem.
48:37And as you said, Jason,
48:39I think what's happening now is that
48:40France is going to be in an incredibly
48:42difficult economic position. We've
48:43already talked about this before that
48:45the that France is is critical country
48:47that's going to set the tone for the
48:48rest of continental Europe
48:50in terms of how it finances its
48:52long-term debts. It's already proposed
48:5350 billion euros of austerity.
48:56That's going to have to go up. I don't
48:58understand how in an austerity scheme
49:00you actually invest in things like
49:02education in a way that would appease
49:04these people if that were the true thing
49:06that they were
49:07rioting and protesting about. The
49:09government is in a no-win situation. I
49:12think they are going to lose the
49:15upcoming presidential elections. And I
49:17think you're going to see a very
49:20complicated reset of the French
49:23bureaucracy and the French economy. And
49:25I think this is the this is the
49:27harbinger of of a lot of change to come
49:29because the average French citizen
49:31should not tolerate what's happening in
49:33their country right now.
49:33>> Friedberg,
49:35what's your take here when
49:37there might have been completely valid
49:40peaceful protests.
49:41High school students protesting is a
49:43beautiful thing.
49:44They they they have some
49:46gripe. They created a lot of the
49:48democratic system that we actually built
49:50America off of. They they were great
49:52pioneers in democracy and protest.
49:54Completely valid form of protest,
49:55peaceful, taking to the streets, and
49:58then it tips over into something that's
49:59obviously at a minimum just incredible
50:03criminal uh and
50:05potentially, you know, in Shamas'
50:07interpretation, you know, terroristic.
50:10And the definition of terrorism, to
50:12terrorize people for your political
50:14goals, yeah, seems like it fits the
50:16definition. So,
50:17for you,
50:18what do you think this represents? Is
50:20this a harbinger of more to come and
50:23something we might
50:25see flare up here in the United States
50:27again? Remember, we had uh the BLM
50:29riots, we had the January 6th riots.
50:31We've had our own uh series of riots
50:33here, and they had a very similar
50:35pattern, you know, people who felt
50:37politically like maybe or socially they
50:40weren't being represented properly, and
50:43then in some cases it tips over to just
50:45criminal looting and incredible
50:47violence. What's your take here,
50:48Freeberg?
50:50>> I think it's a little conflating to add
50:52some of that other stuff to the mix
50:54here.
50:55>> How so?
50:56>> I think those have very kind of acute
50:59political or social political events
51:01tied to them. This one seems to be kind
51:04of much more broad-based. So, if you
51:06watch
51:07some of the interviews that were done on
51:09the street,
51:10there's a very wide range of kind of
51:13complaints if you listen to the folks
51:15that are rioting. Everything from
51:18schools to free Palestine, to the
51:20government's not doing enough for us
51:22with healthcare, with support, with
51:24jobs, we're underemployed.
51:26There's just so much discontent with the
51:29objective being to get the government to
51:31do more for the people. There's a lot of
51:34immigration that happened, and so all
51:36these immigrants that come in, they
51:37expect the government to provide to
51:38them. The The promise was that the
51:40government would provide to them. And
51:42over time,
51:43the government providing things to the
51:45people, as we've talked about countless
51:47times, is never ever ever ever ever a
51:51sustainable answer.
51:53Decentralized government doesn't have
51:55accountability.
51:57There is no progress. There is no
51:59possibility for economic growth. There
52:01is no possibility for creating the right
52:04incentive structure. So, the more the
52:06government tries to provide to the
52:07people, the more the people are let
52:10down. We've seen this in the United
52:11States, I would argue, with the three
52:13things that the government is trying to
52:14provide here, which is healthcare,
52:16housing, and education. And the more the
52:19government spends to try and provide
52:21more of those things, the worse those
52:22things get and the more unattainable
52:25they become. And the same has happened
52:27in France. So, take a look at this
52:29chart. This chart is a measure of public
52:32spending as a percent of GDP. RAND Group
52:34Research published it. And you can see
52:37the comparison of France's public
52:39spending as a percent of GDP
52:42to the peak public spending as a percent
52:44of GDP of major communist nations.
52:47Soviet Union during 1989.
52:49>> chart for you to look
52:50>> Communist Romania, Venezuela under
52:52Chavez,
52:53and then communist China.
52:55>> That's unbelievable. So, do you think
52:57France is a communist state?
52:59>> Well, this is This is like it.
53:01>> Well, it's certainly been a socialist
53:02state, right? Is it working?
53:04>> No, well, that's [laughter] that's my
53:05point. And I think this is such a a
53:07present lesson for the West to look at.
53:10We all think that there's some political
53:11problem with Macron or whoever it was.
53:14The more the government spends, the
53:16worse the conditions get for its people,
53:19not the opposite. And then the people
53:21riot and revolt and say that the
53:22government's not spending enough and the
53:24government needs to be doing more. Now,
53:26take a guess. Pull that chart back up.
53:28Take a guess where the US is today.
53:29>> What percentage of our GDP?
53:31>> 20 20%
53:33>> Yeah, I was going to I was going to say
53:3425.
53:35>> Yeah.
53:36>> Federal government spending is 7.4
53:38trillion. State and local government
53:40spending is 4.6. You add those two
53:42together, it's $12 trillion
53:44out of a $33 trillion GDP. US public
53:48spending is roughly 36% of GDP.
53:50>> Okay, we're right there.
53:52>> France today is um the same as the US.
53:55It's about 120% debt to GDP. France
53:59today is operating its national
54:00government spending deficit is 5.4% of
54:03GDP. The US is actually 6.2%
54:06of GDP. France is
54:09I think it's probably the first to go
54:11because they're only growing their
54:12economy at half a percent while the US
54:14is growing, you know, at 2.3%. So, we're
54:17still kind of hanging on because we have
54:19some degree of economic growth that's
54:21allowing us to kind of hold things
54:22together. But, I would argue that we are
54:25facing the same sort of fiscal cliff
54:27which leads to this extraordinary social
54:29unrest as France.
54:30All the economic measures are aligned or
54:33at the same rate, we're at the same set
54:35of problems, and we're starting to see
54:38the cracks in the system. I think that
54:40France is a bellwether for the West. You
54:42can load in all the immigration stuff
54:44with it because I do think that that's a
54:45big part of it where you open the border
54:47to let people in to try and grow the
54:49economy, you promise them things, they
54:51actually add fuel to the fire, things
54:53get worse, and it becomes a real
54:55problem. Fundamentally, socialism
54:57destroys progress. It limits everyone's
55:01upside by giving everyone a floor. And
55:03the limiting of the upside is what leads
55:05to these sorts of problems where no one
55:06feels like they can progress. And as
55:08soon as the quality collapse says
55:10because you've taken out all the
55:11incentives in the system, everyone
55:13starts to feel incredible discontent.
55:15France, I think, is one of the first to
55:17go, but if we don't right the ship in
55:19all these other countries in the West, I
55:22think we might find ourselves falling
55:23down um the same hole. And I'm very
55:26concerned about that. I think we need to
55:27observe this for what it is and be
55:30really thoughtful about it with our
55:31politics becoming, unfortunately,
55:34a little more disturbed uh going into
55:36this midterm and over the next two
55:37years, I worry quite a bit.
55:39>> What do you specifically think should be
55:41done?
55:41>> Cut spending, cut deficit to GDP, cut
55:44programs
55:45>> Okay, but hold on. Let me Let me just
55:46push back. Yeah. Let me just push back
55:48for 1 second. So,
55:50you know,
55:51what do we think is at the root cause of
55:52this? Let's just focus on the young
55:54people. Like the young people, they're
55:56not optimistic about the future.
55:58>> Yep.
55:59>> Right? They have
56:01a lot of difficulties
56:03financially, career-oriented, as well as
56:06social. I mean, like I think their
56:07social progression has been stalled
56:09relative to other generations.
56:12You have these things that are like
56:15these swords of Damocles hanging over
56:17them, like pensions. I'm sure that they
56:19don't understand it completely, but you
56:21know, if they took a second to realize
56:23that at some point these pension things
56:24come due for their parents and their
56:27grandparents,
56:28that's going to be on them. So, how do
56:31we set it right for them? What do we do?
56:34>> I think four things are the primary
56:36cause of this that could be directly
56:38addressed. It's
56:40education, housing, healthcare, and then
56:42the retirement system.
56:44Every one of them, there's probably a
56:46multi-year solution. Education, I think
56:48we need to wind down the federal student
56:49loan program. I've said this many times,
56:52and I think in the process we've got to
56:53eliminate a lot of the student debt
56:55that's outstanding. I think it's a
56:56necessity.
56:57>> So, cancel the underwriting at the
56:58federal level, but also cancel the
57:00existing debt of the folks that have it.
57:02>> There's a complicated mechanism that I
57:03think's needed for the deletion of
57:05student debt, perhaps in bankruptcy or
57:07something like that, to go back on that
57:08that law that that made it impossible.
57:11>> I just on this one just to add to this,
57:13I'm extremely sympathetic to canceling
57:15student debt. I don't think it should be
57:16I think it should have always been
57:17dischargeable through bankruptcy,
57:20but I I'm quite I'm quite supportive of
57:22this idea.
57:22>> has it has to happen with deleting the
57:24federal student loan program. What will
57:25happen then is you'll right-size all the
57:27institutions and their tuition costs,
57:28because then all the institutions will
57:29start to fit to the market. It'll take
57:31about 10 years. It won't happen
57:33overnight. So, we have to have a firm
57:35commitment to seeing it happen over the
57:36next decade. In housing, we've got to do
57:38something similar with housing loans
57:40that keep inflating the cost of homes
57:41and deal with the fact that a good chunk
57:43of the middle class
57:44>> but go go through your list. Go through
57:46your list cuz I have a I have I have one
57:48critical question at the end of your
57:50list cuz your list makes sense on the
57:51surface.
57:52>> Then I think that we need to make
57:53defined benefit retirement programs,
57:56including the Social Security Trust
57:58Fund, illegal. I think all defined
58:01benefit programs need to be transitioned
58:03to defined contribution programs over,
58:05call it a decade period. In the process,
58:07everyone gets, like whatever you're owed
58:09in Social Security gets deposited in
58:12effectively an IRA or a Trump account or
58:14whatever they're called for you today.
58:15And all of Social Security is deleted,
58:17and now everyone owns equity.
58:19>> Superannuation funds, like we talked
58:21about in Australia. We just transfer to
58:23that. What would the pushback on that
58:24be, Freeberg?
58:25>> Because people lose a guaranteed income.
58:27But rather than do it one time, what you
58:28would do is you would transition 10% of
58:30all your Social Security amount owed to
58:32you per year for the next 10 years. So,
58:34you do a 10-year transition where
58:35everyone gets their money.
58:37>> out the buying into the market.
58:39>> Yeah, great.
58:39>> Yeah, and then people can choose a 40/60
58:41plan or an 80% bonds plan or whatever
58:43plan they want depending on their age.
58:45>> Do health care and kind of do that.
58:47>> Health care is the hardest, but I think
58:48housing health care is way more
58:50complicated, but I think it's
58:51fundamentally got to get pushed to
58:52corporations that employ people where
58:54they've got to go into a market for
58:56insurance, and that forces the market
58:57for insurance to force prices and
58:59services down. But there's two things
59:01that get in the way.
59:01>> housing
59:02>> Wait, and yeah, so So, HIPAA and housing
59:05regulations are the two things that get
59:06in the way of price coming down in
59:07health care and housing.
59:10So, you got to address those, which are
59:11much more complicated because in the
59:12case of housing, it's it's state and
59:14local.
59:15>> So, let me just ask a question.
59:17So, when you look at France and Italy
59:18and you look at these riots,
59:20I think it's important to observe that
59:22they have, theoretically,
59:24two of the four roughly solved, which is
59:26they give you free education and free
59:28health care.
59:29>> Okay? Do you know how cheap it is to go
59:31to college, Chamath, in in France?
59:35>> I don't. I don't.
59:35>> 300 euros a year and costs the state
59:37like 12,000. So, they are in college and
59:40graduate school. When I was in Paris
59:42this summer, I had somebody explaining
59:43to me the resentment here. They consider
59:45it like
59:47I'm trying to think of the equivalent we
59:49might consider here in terms of
59:50entitlements. They believe they should
59:51be able to go to college and graduate
59:53school forever.
59:54>> I'm asking you, Jason, I'm asking I'm
59:56asking both of you is
59:58you bring up these four
1:00:00issues. One is a vagary of the United
1:00:02States, meaning the student debt issue
1:00:04is purely an anomalous thing that exists
1:00:07in the US that does not exist in some of
1:00:09those Western countries where we're
1:00:10seeing riots.
1:00:12The second thing that we brought up,
1:00:13which is health care, is roughly free
1:00:15over there. The third thing we brought
1:00:16up, which is education, is roughly free
1:00:18over there. The fourth thing we brought
1:00:20up, which is, you know, defined benefit
1:00:22versus defined contribution,
1:00:24again, the the system there is a problem
1:00:26because it's the same system that we
1:00:28have effectively and you have these
1:00:29pension overhangs. But, I guess what I'm
1:00:30saying is
1:00:32there
1:00:33two of the four are roughly solved. One
1:00:35of the four doesn't exist. And yet,
1:00:37they're still rioting. So, it's not
1:00:39solved.
1:00:40>> No, no, no. I can I can actually Hold
1:00:41on. I can answer this, I think. You
1:00:43know, the
1:00:44the original riots, when it was
1:00:45students, were
1:00:47banal, you know, maybe throwing a couple
1:00:50of bottles here kind of riots. The
1:00:52students weren't looting. The looting
1:00:54occurred by
1:00:56uh illegal immigrants uh from Northern
1:00:59Africa and uh a lot of resentment from
1:01:02Islamic Muslim immigrants in France who
1:01:05have not been allowed to integrate into
1:01:07society and in some cases are second and
1:01:09third generation Parisians who have
1:01:11either not integrated into society or
1:01:14part of this is also on the native
1:01:16French who have not integrated into
1:01:18society. If those people weren't there,
1:01:20Chamath,
1:01:21they weren't there, we wouldn't be
1:01:22having this conversation. We would just
1:01:23be talking about run-of-the-mill
1:01:26uh you know, people who just want a
1:01:27slightly better classroom, slightly
1:01:29better teacher ratio to students. This
1:01:32is because of the uh unmitigated
1:01:34disaster of immigration to Italy, to
1:01:38Spain, and to France. And people have a
1:01:40hard time just saying it like it is, but
1:01:43that's what it is. And that's why, you
1:01:45know, here I think we stopped it by
1:01:47closing the southern border.
1:01:49Uh and there they have a different type
1:01:51of immigration. It's it's people who
1:01:53don't want to integrate. If you take the
1:01:54average Mexican who comes to America,
1:01:56they're more than happy to learn
1:01:57English, they're more than happy to
1:01:58integrate into American society and our
1:02:01customs. It is not the same in the UK,
1:02:04Italy, and obviously France, as we've
1:02:06seen.
1:02:07It's an immigration issue.
1:02:08>> But Chamath, I don't I don't know if
1:02:10it's actually solved, because I think
1:02:11the time to get a doctor's appointment
1:02:13is weeks, the quality of the education
1:02:15is poor, you're you're sitting in a
1:02:17classroom with 60 kids. That's what the
1:02:19kids are complaining about. There's
1:02:21obviously a lot of other
1:02:22kind of motivations in those riots that
1:02:24are happening, but there's a fundamental
1:02:26feeling of oppression that I think
1:02:29weighs heavily on this entire
1:02:31population. And the oppression is
1:02:33there's a higher caste that is doing
1:02:35better than we are, and it is not fair
1:02:37because we were promised that we would
1:02:39do well. Pull up this video, Nick, that
1:02:41I sent. I think this is a view of some
1:02:43of the anger.
1:02:45Watch this.
1:02:46>> The Chinese have become too rich in
1:02:47France.
1:02:52>> Too rich?
1:02:55And that's not fair?
1:02:57>> Well, I think it's not fair.
1:02:58They're now in the category of people
1:03:00with nice clothes and big cars.
1:03:02>> So, this was a response interview to why
1:03:05the rioters were going after
1:03:07Chinese-owned stores and destroying them
1:03:09in Paris. So, they're specifically
1:03:11targeting Chinese-owned stores, breaking
1:03:13the stores, burning them down, because
1:03:15they think that these immigrants have
1:03:16become too successful in France and they
1:03:19don't deserve that success. And I think
1:03:22that that is what this weight of
1:03:24socialism, this weight of a giant
1:03:25government
1:03:27does to everyone, which is if anyone is
1:03:28seen to be ahead of the suffering that
1:03:31we are all dealing with, they must be
1:03:33the problem. They are bad. They are the
1:03:35issue. And I think that there's this
1:03:37kind of fundamental rooting on the free
1:03:39Palestine movement,
1:03:41on, you know, the the patriarchy. It's
1:03:44all rooted in the same sort of
1:03:45principle, which is they don't realize
1:03:47that the reason that they're feeling
1:03:48this way is because the government is
1:03:50already trying to do too much and
1:03:52they're not able to do it well because
1:03:53the government is the worst place to
1:03:56provide services to the population.
1:03:58>> Well, and in fact
1:03:59>> see they see any degree of success and
1:04:01it's sort of like what we're seeing with
1:04:02the populist movement in the US, hating
1:04:04tech, hating billionaires, and you know,
1:04:06it's it's very common across the West
1:04:08and it's going to continue to grow,
1:04:10unfortunately.
1:04:11>> Well, and reports where they were
1:04:12specifically targeting Chinese and
1:04:14Jewish businesses and specifically
1:04:16avoiding, this is when it moved into the
1:04:17full-scale riots and terrorism. And
1:04:19avoiding Muslim businesses.
1:04:21>> That's right.
1:04:21>> And I I you know, I
1:04:23I I got to say like thank God for the
1:04:24second amendment. I don't think we would
1:04:26have these type of riots here on a
1:04:28sustained basis. If you remember the LA
1:04:30riots and there was a there's a term
1:04:32called rooftop Koreans. And if you
1:04:35remember, the Koreans just literally
1:04:37went on top of their businesses and
1:04:38defended in
1:04:40uh K-Town in in in LA, they defended
1:04:42their businesses. And when you see these
1:04:43poor business owners not being able to
1:04:45defend their businesses and they're
1:04:46beaten
1:04:47>> Yeah.
1:04:48>> and you just think, my God, that poor
1:04:50small business owner, they put their
1:04:51entire lives into this business. You
1:04:53know it's not easy to run a small
1:04:54business. And then they're being beaten
1:04:55and watching their lives work burn down,
1:04:58that's unacceptable. I I honestly think
1:05:00you got to rethink the first the the
1:05:03second amendment as well. And I know
1:05:05I'll get massive blowback, but
1:05:07you don't you don't see this in Texas
1:05:08and LA.
1:05:09>> Let Let give you how I try to think
1:05:11about what the end game of this looks
1:05:13like.
1:05:14>> Yeah. I mean, basically what you guys
1:05:15are saying is independent of who is in
1:05:17charge, it could be left, it could be
1:05:19right, and independent of what the facts
1:05:21are, you basically showed a chart here
1:05:22that shows that France spends more than
1:05:24every other communist government in the
1:05:27history of the world.
1:05:30These things are a disaster.
1:05:31And there's part of it here where people
1:05:34probably think that this isn't true, and
1:05:37they want to see even more state
1:05:38intervention to make things better. They
1:05:40themselves know that they can't solve
1:05:41the problem, and so the only person to
1:05:43turn to is the government. Because it's
1:05:45not like you're going to turn to the
1:05:46person on your left or your right.
1:05:47>> That's right.
1:05:48>> Okay. So
1:05:50>> you go to the government, but at some
1:05:51point they're going to realize that
1:05:53chart is true. And they're going to see
1:05:54that enough times on their television or
1:05:57on Tik Tok and X and Facebook and
1:05:59Instagram, where they're going to be
1:06:01like, wait a minute, I had no idea we're
1:06:02spending twice as much as communist
1:06:04China. I thought the whole point was
1:06:06that we should be spending more like
1:06:07communist China or like communist Russia
1:06:09because this socialism thing, you know,
1:06:12is the better version of communism. It's
1:06:14not going to work.
1:06:15So then who can really fix this?
1:06:19That's sort of where I go to, because
1:06:20independent of what you think is right,
1:06:23who can fix this?
1:06:24>> You need neo-Nazis.
1:06:26>> and I No, I think that you go I go back
1:06:28to two things.
1:06:30One which is in our control,
1:06:32and one that is in the control of
1:06:36the large pools of money and, you know,
1:06:38what I kind of euphemistically call
1:06:39these bond vigilantes.
1:06:42The bond vigilantes can break the back
1:06:43of the French government.
1:06:45And they can do that economically, and
1:06:47they can do that by essentially making
1:06:49their debt unfinanceable.
1:06:51The knock-on effect of that decision
1:06:55would be where you bring long rates so
1:06:57far up that they can't finance their
1:06:59debt.
1:07:00And it essentially causes an economic
1:07:02reset inside of France. And in that, you
1:07:05have to have a grand bargain where you
1:07:07can try to reset these things.
1:07:09The problem with that is that every time
1:07:12that it's happened in the past, somebody
1:07:14accrues more power and somebody thinks
1:07:16about their own ability to then have
1:07:18influence over the next end number of
1:07:20years after that reset. And it's never
1:07:22about resetting in the right way.
1:07:26The right way would cause you to look at
1:07:28the pension system to your guys' point,
1:07:30you know, enforce a bunch of rules
1:07:33around, you know, the student to teacher
1:07:34ratios in classrooms, enforce a bunch of
1:07:36rules around wait times at hospitals,
1:07:38enforce a bunch of rules around, you
1:07:40know, how fast and or how slow you can
1:07:43get building permits and what you can
1:07:44see in rent adjusted inflation, yada
1:07:46yada. You could do all of these things
1:07:48if you had the political will.
1:07:49Unfortunately, that's not how it's going
1:07:50to go down. It's going to go down by who
1:07:52is in power, who wants to basically pin
1:07:54the blame on somebody else and then have
1:07:56that power for the next five or 10 years
1:07:57coming out of a reset. Because in the
1:08:00middle of a reset, what happens is you
1:08:02just give away all your rights.
1:08:04So, that's not the solution. So, then
1:08:06what's the solution? Then we're only
1:08:07left with, okay, how do we all solve
1:08:08this? And you're back to the whole like
1:08:12is there like a panacea? And the only
1:08:14panacea that nobody wants to believe is,
1:08:16you know, we say it cuz we believe it,
1:08:18like AI abundance and all of this
1:08:20beautiful stuff.
1:08:22It just sounds like meshugas. Like it's
1:08:24like, I don't know, if I were if I was
1:08:25anybody in France, I'd be like, AI? What
1:08:27are you talking about? A bunch of
1:08:29oligarchs from Silicon Valley? No, thank
1:08:31you.
1:08:32So, it's almost this intractable
1:08:34problem. It's like the irresistible
1:08:36force meeting the immovable object. And
1:08:39I am completely at a loss
1:08:41to
1:08:43simulate in my brain
1:08:46an outcome here. I don't see a
1:08:48reasonable end game.
1:08:49And so, I'm completely confused and
1:08:52completely without Yeah, I mean, look,
1:08:54it depends how far the collapse goes.
1:08:56You could have this sort of social
1:08:58collapse that you see in some countries
1:09:01that involve rioting and murder and all
1:09:03these sorts of things or you end up
1:09:05having what is more of like an economic
1:09:06collapse which is sort of like
1:09:09Argentina and Brazil and others have
1:09:12gone through Greece more recently
1:09:16where the ability to service the debt
1:09:18because the government got too big
1:09:21forces this kind of restructuring of the
1:09:23economy and yes there are still massive
1:09:24protests there is a still an incredible
1:09:27loss of net worth there is an incredible
1:09:30inflation effect effectively that wipes
1:09:32out asset values for people
1:09:34and you end up having kind of a
1:09:35resetting economic resetting without the
1:09:38great social resetting. So you know
1:09:40which path does he choose? I think we
1:09:43have to do something. I don't think we
1:09:45can run a higher deficit to GDP than
1:09:48France in the United States with the
1:09:50same debt to GDP as France and only
1:09:52modestly higher GDP growth. I think we
1:09:55are fundamentally going to need to have
1:09:56some degree of reset in the United
1:09:58States. The question is do we want it to
1:10:00be a massive social reset or do the left
1:10:03and the right want to put on their hard
1:10:04caps and say okay we got to go in the
1:10:06mines and do the work and figure out how
1:10:08to cut this from seven and a half a year
1:10:09to four a year and do it in a way that
1:10:11doesn't destroy the economy. And I think
1:10:13that there's a couple of solutions that
1:10:16will allow that to happen like you know
1:10:18you basically can deflate parts of the
1:10:19economy like the education system the
1:10:21higher education system by taking away
1:10:23federal student loans. You can have a
1:10:25hit to housing which will drop the net
1:10:27worth of the middle class by 10 to 20%
1:10:29by taking the federal home loan system
1:10:31and reducing the underwriting that's
1:10:33happening reduce the volume that's
1:10:34pushing through because we've used it
1:10:35largely to keep prices propped up high
1:10:38to keep homes net worth high so that the
1:10:40middle class remains highly solvent and
1:10:42feel wealthy. The problem is it's now
1:10:44inaccessible to the young people so they
1:10:46can't buy a home. So we're now sort of
1:10:48hitting the head here something's got to
1:10:50give. So if you want to have the social
1:10:51unrest
1:10:52you don't move on that. But if you want
1:10:54to allow them to feel like they have
1:10:55access you're going to have to take a
1:10:57hit. Middle class, upper class, wealthy,
1:10:59etc. in the US are going to have to take
1:11:00a hit on asset value and transfer it
1:11:03more kind of carefully
1:11:05to the emerging folks. And then you've
1:11:06got to fix this pension system
1:11:08which is
1:11:10either overloaded or underloaded. In
1:11:11both cases, it's really bad.
1:11:14And this is why when ERISA happened, all
1:11:16of the private companies in the United
1:11:17States got rid of a defined benefit
1:11:20programs. The only defined benefit
1:11:22programs in the United States that still
1:11:23exist are inside of governments because
1:11:26unionized government employees said we
1:11:27want to keep defined benefits. And if
1:11:29they'd moved to defined contribution,
1:11:31this is what's so sad about it. If they
1:11:32had moved to defined contribution
1:11:34programs, including social security,
1:11:36they would have all had so much more
1:11:38money. Everyone would have been
1:11:40wealthier.
1:11:40>> even have to wonder about this. You can
1:11:41literally look at what's happened to
1:11:43Australia and how happy people are that
1:11:45they're forced to be in the market.
1:11:47>> And Norway, yeah.
1:11:48>> And Norway. They're just like, if you're
1:11:50if you're forced to do that. And then we
1:11:51had this incredible thing this week
1:11:53which dovetails with this, which is Elon
1:11:55said he's going to give every American
1:11:57child, which I think is defined as under
1:11:5918, a share of SpaceX worth 150, 160,
1:12:03170 dollars it's been trading at, which
1:12:05is a 10 billion dollar gift, which I
1:12:07believe will be amongst the largest
1:12:09gifts in the history of charity ever
1:12:11done. And he said he's just getting
1:12:12started with that. There's nothing that
1:12:14says this is not going to
1:12:16become
1:12:18you know, a bit of a contagion amongst
1:12:21the affluent to give a large portion of
1:12:24their wealth to
1:12:26or even a small portion of their overall
1:12:28wealth, but giving it directly to kids
1:12:30who can't touch it for 18 years. And if
1:12:32we made it part of their superannuation
1:12:33fund, maybe we could have invest America
1:12:37up to 18 years old for college, and then
1:12:39we could have an invest America for your
1:12:41retirement. I mean, it would actually if
1:12:42you bifurcated that and you gave one
1:12:44share to each, that could be even more
1:12:46accretive, right?
1:12:47>> Jason, what did you think about your
1:12:48friend Kara Swisher's
1:12:50>> No longer friends. Kara and I have not
1:12:52been friends for a long time.
1:12:54I I think she is one of the
1:12:56>> Can you just describe for the audience
1:12:57what her criticism was of Elon and your
1:12:59reaction to it?
1:13:01>> she had multiple videos. One of them is
1:13:03it's only one share
1:13:05and that they should just give money
1:13:08directly to people. And you know, it's
1:13:11obviously what the reason that's
1:13:13incredibly stupid is you should give any
1:13:15possible gift that people are willing to
1:13:17give and forcing people to hold on to it
1:13:20makes them part of capitalism. And Kara
1:13:22Swisher makes $10 million a year from
1:13:24her podcast. She talks about it all the
1:13:25time. She is of the 0.1 or 0.01%
1:13:29she should give half of her podcasting
1:13:31revenue if she's so virtuous to invest
1:13:33America accounts. And she's also saying
1:13:35they should be putting Elon in jail
1:13:38and that the Democrats are going to do
1:13:39that if they win. So there's like kind
1:13:41of a a doubling there, but she's um
1:13:44you know, she's a Karen and she's a
1:13:45hypocrite.
1:13:46>> Yeah. Can I say something about this?
1:13:47Nick, can you pull up that image of that
1:13:49woman
1:13:50whose store was destroyed in France
1:13:52who's a well-known activist? You guys
1:13:54see this this this image
1:13:56>> I did see this and I don't know that
1:13:58it's true or not, but the story here is
1:14:01>> Yeah, the story is that this woman is a
1:14:02well-known activist in France who for a
1:14:04long time was very vocal about opening
1:14:06the borders and letting the immigrants
1:14:07come in and taking care of the
1:14:09immigrants. And she was angry and upset
1:14:11and yelling at people in the street
1:14:12while they were destroying her store
1:14:14saying, "Do you know who I am?
1:14:15I am an activist who advocated for you
1:14:18all." And they didn't give a
1:14:21They went in and destroyed her store and
1:14:23the look on her face which was like,
1:14:24"What the hell happened?" is meant to be
1:14:27this moment that sort of captures the
1:14:29zeitgeist of what some people call
1:14:32suicidal empathy,
1:14:34but I think it's much more about being
1:14:37call it
1:14:38an armchair do-gooder.
1:14:41An armchair do-gooder is someone who
1:14:42kind of sits at home
1:14:44in a nice home in the suburbs
1:14:47or wherever they live and they say
1:14:50let all the people in, take care of all
1:14:52the people.
1:14:53But as soon as the people show up to
1:14:55their front step and says, "Can I come
1:14:58and stay in your house?" the answer is
1:15:00no, of course not.
1:15:02>> Yeah.
1:15:02>> Someone else has to deal with it.
1:15:04Someone else has to hold you in their
1:15:06home. Someone else has to house you.
1:15:07Someone else has to feed you. Someone
1:15:09else has to give you a job because I
1:15:10can't house you. I can't feed you. I
1:15:13can't give you a job. And I think that
1:15:16this is what has become inherent
1:15:17throughout the West, which is the
1:15:19armchair do-gooder that says we should
1:15:21do all these good things.
1:15:23But when it comes home to roost, when
1:15:25the cost of doing those good things
1:15:26comes home to roost they're shocked.
1:15:29>> Yeah.
1:15:29>> And they're all in the face and they're
1:15:30like, "Holy shit." And this This, by the
1:15:33way, we see it in the US.
1:15:34>> We had a term for it, by the way.
1:15:36>> great videos on Twitter, by the way,
1:15:38where these people go around They go to
1:15:39suburban homes and they're like, "Hey,
1:15:41did you vote for so so-and-so bill to
1:15:43have unrestricted immigration?" They
1:15:44say, "Yes." They're like, "Okay, here's
1:15:4515 guys that want to come live in your
1:15:47house." She's And the people are like,
1:15:48"No, I I They can't live in my house.
1:15:50What do you What do you mean?"
1:15:50>> This was called In another era, we
1:15:52called this limousine liberals and it
1:15:55was manifested when they used
1:15:59which was quite offensive
1:16:00but quite effective when they used
1:16:02immigrants and bust them to Martha's
1:16:04Vineyard or other places and were like,
1:16:06"Here's your immigrants. You said you
1:16:07wanted them."
1:16:08>> What Rob Henderson called luxury
1:16:09beliefs.
1:16:10>> You know, luxury beliefs, limousine
1:16:12liberals. You know, if you were to look
1:16:14at the problems in France, they're not
1:16:16completely intractable.
1:16:18A lot of the complaints are, I think,
1:16:20you know, wait times at a hospital.
1:16:21Well, you have free health care. Oh, my,
1:16:24you know, the student ratio. Well,
1:16:25you're only paying 300 euros a year for
1:16:27your college. If you want to go from 60
1:16:29to to one down to 30 to one, okay,
1:16:32you're going to need to charge you 2,000
1:16:34euros a year.
1:16:35>> you don't. No, you don't. You don't need
1:16:36to charge them more. This is the point.
1:16:38The point is when the bureaucracy
1:16:39inflates and the quality of the service
1:16:42goes down.
1:16:43An intelligent person would look at that
1:16:45inflated bureaucracy and say, this is
1:16:47inversely correlated to what I expected.
1:16:50So, really what should happen is people
1:16:52have the intelligence to say, hold on a
1:16:53second, I'm now spending 57 cents of
1:16:57every gross productive dollar on a
1:16:59government that now has increased my
1:17:02wait times and decreased the student
1:17:04teacher ratio.
1:17:05>> there's something not going to be
1:17:06efficient. So,
1:17:07not a sufficient as
1:17:08>> Exactly right. That's right.
1:17:10>> But, if you the second part of what I
1:17:12was going to say here is, you know,
1:17:13there's austerity measures which have
1:17:15been happening in Europe for a long
1:17:16time. But, if you have an immigration
1:17:19rate where you're importing 1.5% of your
1:17:21population a year and your GDP is not
1:17:24growing, that might be a signal to you
1:17:26that maybe you have to change gears. And
1:17:27if the people who you're immigrating
1:17:29into the country are not productive or
1:17:31not productive yet, maybe their kids
1:17:33will be productive or their kids' kids
1:17:34will be productive, but right now
1:17:36they're a drain on society, you can only
1:17:38accept so many people and then there's
1:17:40another nuance point under that, which
1:17:41is if you cannot you cannot accept
1:17:43people who do not make a commitment to
1:17:45assimilate into that country's culture.
1:17:49Period. Full stop. You can have a very
1:17:51small number of people, whether it's
1:17:52Japan, the United States, or France, you
1:17:54can only have so many people join the
1:17:55country who don't speak the native
1:17:57language because that becomes an
1:17:58incredible burden. Now, you have to
1:18:00create an entire second education
1:18:02system, health care system where people
1:18:03speak Spanish, where people speak
1:18:05whatever, you know, French, German, and
1:18:07whatever it happens to be. And and
1:18:09that's what's taxing these
1:18:10>> I think it's one in 19 are from northern
1:18:12Africa and like one in nine is an
1:18:14immigrant. And they have un
1:18:16mitigated immigration in Europe. That's
1:18:20the the other piece to this problem. If
1:18:22they didn't have that for the past 20
1:18:23years and they weren't adding so many
1:18:26low-value immigrants as opposed to
1:18:28high-value immigrants, high-value being
1:18:30people with graduate degrees who are
1:18:31investing in companies versus people who
1:18:33are taking bottom rung jobs, and you
1:18:35started looking at it
1:18:37against the GDP of the country, against
1:18:39the unemployment, against where does
1:18:41this immigrant fit in what jobs we need?
1:18:44So, do we need this type of worker?
1:18:46Construction worker? Yes, okay, we
1:18:47should add more of those cuz we have a
1:18:49deficit, but they're not doing it
1:18:51intelligently, and that speaks to this
1:18:52woman running her small shop. She wants
1:18:55to allow all these people to come there,
1:18:57but she's not looking at the the balance
1:18:59sheet, and the balance sheet doesn't
1:19:01support it. I'll simplify it down. I
1:19:02think it's an economic issue, yeah.
1:19:04>> Just to restate what you said slightly
1:19:06differently,
1:19:07if you're providing social services to
1:19:10an immigrant population, then that's not
1:19:13necessarily a productive immigrant
1:19:15population in terms of being able to
1:19:17contribute to growing the economy and
1:19:18supporting
1:19:19>> Yes.
1:19:19>> civil society. Social services
1:19:23that are requisite for an immigrant
1:19:25population should indicate to you that
1:19:26that's probably not the immigrant
1:19:28population you want to bring in.
1:19:29>> This has nothing to do
1:19:31>> with the countries or the character of
1:19:33the people
1:19:34>> of skin,
1:19:35>> It's simply a statement that if you're
1:19:36paying for social services to support an
1:19:39immigrant population, something's wrong,
1:19:41and it's the wrong population that you
1:19:43want. But, if a population of immigrants
1:19:45come in and they're productive, meaning
1:19:47they can get jobs and they can work, and
1:19:49they can support themselves, that means
1:19:52they're adding something positive to the
1:19:54economy, to the civil society, and that
1:19:56is the sort of selection criteria that
1:19:58needs to take place. So, that's a simple
1:19:59framework for determining whether or not
1:20:01you should do this, and an easy way to
1:20:02solve it is get rid of social services
1:20:05for immigrants. As soon as you get rid
1:20:06of social services for immigrants, then
1:20:08you self-select into the right
1:20:10population that's going to come in. The
1:20:11second thing is Jason, just to put a
1:20:13point on this, it's not about
1:20:15assimilating to society,
1:20:17it's about sharing the values of what
1:20:19makes the United States the United
1:20:21States.
1:20:21>> Well, that's what I mean.
1:20:22>> The value of self-sufficiency, I don't
1:20:24care about your religious beliefs, I
1:20:26don't care about your movie interests, I
1:20:27don't care about your first language, I
1:20:30don't care about your sports interests.
1:20:32The the multiculturalism of the United
1:20:34States actually adds an incredible
1:20:35amount of value. But the values of what
1:20:38do we actually care about? What are the
1:20:39principles that we all share as a civil
1:20:41society of the United States or a civil
1:20:44society of France is a screening
1:20:46criteria that seems to have been thrown
1:20:47out the window as we've allowed the
1:20:49immigrant flocks to kind of take over
1:20:51the West. Yeah.
1:20:51>> freedom of speech? Do you believe in the
1:20:53Second Amendment? Do you believe in the
1:20:54core benefits of America? You pull
1:20:55yourself up by your
1:20:56>> women's rights? Do you believe that
1:20:58women are equal to men?
1:20:59>> Yeah.
1:20:59>> As a simple one, right? Do you believe
1:21:02that you should work for your money as
1:21:03opposed to take it from the government?
1:21:05Do you believe in your individual
1:21:07rights? Do you believe in the legal
1:21:08>> in the legal system versus your church's
1:21:10legal system?
1:21:11>> Is your church paramount to the
1:21:12government and do you in in the laws of
1:21:14the state? And do you believe that other
1:21:16people have rights that you don't get to
1:21:19trample on? Those are the sorts of
1:21:21things that that are in many cases
1:21:23counter to the cultural, religious, and
1:21:25other traditional beliefs of certain
1:21:27populations that end up causing the
1:21:29damage in the West and I think that it's
1:21:31been kind of completely lost in the
1:21:33process.
1:21:33>> You know what I thought when I heard
1:21:34that kid talking about targeting these
1:21:36Chinese people? It's like weirdly
1:21:40reminiscent of like a a more bottoms-up
1:21:43maybe poorly disorganized but a
1:21:46bottoms-up version of Kristallnacht
1:21:47where
1:21:48>> Yes.
1:21:49>> you had obviously the looting of Jewish
1:21:52businesses and synagogues
1:21:54in 1938.
1:21:56>> Cuz they were doing better.
1:21:57>> Which were state-sponsored and you know,
1:21:59they had like some idiotic pretext but
1:22:01ultimately it was because they were
1:22:02jealous of the Jews. And now here you
1:22:04see folks that are essentially like
1:22:07burning and pillaging businesses and at
1:22:09least some narrow percentage of them
1:22:11were doing it effectively to target a
1:22:13cohort of people that they are jealous
1:22:16of and they are jealous of exactly what
1:22:17you said because I don't think the
1:22:19Chinese
1:22:20you know, emigrate to a country and have
1:22:22some of these long-tail views. They're
1:22:24like, "Get me in the game, coach. Tap me
1:22:26in. I'm ready to go.
1:22:27>> They're entrepreneurs and they want to
1:22:29do better.
1:22:30>> They're industrial. They're super hard
1:22:31working.
1:22:31>> Exactly. And so the the contrast is
1:22:33quite stark where you see one immigrant
1:22:36essentially saying "A s'entendre pas
1:22:37juste" is basically his way of saying
1:22:39"It's not fair." Which is like kind of
1:22:41like a petulant child.
1:22:42>> Yeah.
1:22:43>> You know, so on the one hand you have
1:22:44petulant children on the other hand you
1:22:45have hard working immigrants. They're
1:22:47both immigrants. It's the craziest
1:22:49dichotomy. I mean, I just seeing it in
1:22:51real life
1:22:51>> Immigrant on immigrant violence because
1:22:53one is working harder, has a better
1:22:55plan, is applying themself
1:22:58and probably working seven seven days a
1:23:01week, 12 hours a day at that business to
1:23:03make it successful.
1:23:04>> Her her specific story is even more
1:23:08>> I was talking about the Chinese one.
1:23:09Yeah. Not the
1:23:10>> Yeah, the Chinese woman. Cuz they went
1:23:11and they actually told her story and she
1:23:13you know, she had worked there you know,
1:23:14she had worked essentially in
1:23:15restaurants for 34 years to earn the
1:23:17money to open that restaurant.
1:23:18>> Yeah. I grew up in the restaurant
1:23:20business. Let me tell you, it's the
1:23:21hardest thing on a family is to run a
1:23:23business because when somebody doesn't
1:23:25show up for work or you don't have
1:23:28the money, you all just have to work. My
1:23:30family, we all worked at the restaurant.
1:23:32My mom My mom made the desserts at our
1:23:35restaurant because she wanted to save
1:23:37money. And she was a nurse, not a pastry
1:23:38chef. Yeah. And you know, and and my
1:23:41brothers and I, we worked at the
1:23:43restaurant since we were six or seven
1:23:45years old. We set the tables. We
1:23:46polished the silverware. We were working
1:23:48in the kitchen at 10 years old
1:23:51and we were still living week to
1:23:53week.
1:23:54And then you come burn the business
1:23:55down. You come burn my dad's business
1:23:57down? We're going to unload. My dad
1:23:59was packed was packing, you know? Like
1:24:01he was strapped. And you you do this in
1:24:04America? You do this in Texas?
1:24:06You're going to get shot. And I I when I
1:24:08saw those poor Chinese people on the
1:24:10balcony, you know, and their lives are
1:24:12at risk and their whole livelihood, I
1:24:14just like man, I wish these people had
1:24:16guns. I wish these people had guns to
1:24:17could defend their private property.
1:24:20It is just crazy that they could just be
1:24:22run over like this. And they were going
1:24:24to get killed. There's absolutely a
1:24:27non-zero chance that that mob when mobs
1:24:30get into that heightened state, they
1:24:32will murder people. They will beat them
1:24:34to death. And this poor other shop owner
1:24:36was getting kicked and beaten just
1:24:38trying to defend his little tiny
1:24:39business, little tiny store. If you've
1:24:40ever been to Paris, little tiny stores,
1:24:43beautiful bespoke stores trying to make
1:24:45a living in something they're passionate
1:24:46about. It's just absolutely
1:24:47heartbreaking. And you know, it's it's
1:24:50>> And I'll just say this, the bond
1:24:51vigilantes need to show up here and draw
1:24:53a hard line in the sand. I think they
1:24:54need to
1:24:55break the back of the French government.
1:24:58>> Okay, there it is. There's the mandate.
1:25:00>> And you need to do it severely and you
1:25:01need to do it promptly.
1:25:02>> 6-7% would do what? 6-7% on these bonds?
1:25:05>> I think you need to take France to like
1:25:068% and just break them.
1:25:08>> Yeah. And then all of a sudden austerity
1:25:10measures, immigration, all this stuff is
1:25:12going to be taken a lot seriously.
1:25:13>> At least it'll all be on the table and
1:25:15they'll have to sort through. They will
1:25:17have to sort out
1:25:18what they're willing to give up. But
1:25:20right now, if you can skate through this
1:25:22moment,
1:25:23then these things continue to fester and
1:25:26it is a tinderbox that at some point,
1:25:28again, we know, again, just to use the
1:25:31analogy of Kristallnacht, we know what
1:25:33the trail of breadcrumbs
1:25:35can lead to in terms of fomenting the
1:25:37societal rage. You cannot allow that to
1:25:40happen. And so, the government has a
1:25:42responsibility now to nip this in the
1:25:45bud in a structural way.
1:25:47And I think the people have a
1:25:48responsibility to understand the numeric
1:25:52reality in which they are living. They
1:25:54are living in a state that is spending
1:25:56more than any communist utopia that they
1:25:59hoped to achieve.
1:26:01Yet the outcomes are broken, which means
1:26:03the system is broken and you need to
1:26:06reset that. And the government will not
1:26:08reset it on its own and so the bond
1:26:10vigilantes will have to do it for them.
1:26:12This kind of stuff, if it continues to
1:26:14compound, is extremely extremely
1:26:16dangerous.
1:26:17>> Yeah, super dangerous.
1:26:20And but we had our little flashes here.
1:26:21So it's it's we're definitely not immune
1:26:24to it. We're we're a larger group of 50
1:26:26states. So and they're they're at by the
1:26:28way, I think France is almost at 56% is
1:26:31their top tax rate.
1:26:35Maybe it's 55, 56. The only place that's
1:26:37higher I think in Europe is Denmark,
1:26:39which is like a 60% top tax rate.
1:26:41>> It doesn't matter. Take it to 90%.
1:26:44Nothing will get better, folks.
1:26:46Do whatever you want.
1:26:47>> If the government provides a service,
1:26:49there's going to be no innovation and
1:26:50there's going to be no and there's going
1:26:52to be no efficiency. It's just it is
1:26:53what it is.
1:26:54>> an individual human incentive to do
1:26:57well,
1:26:59nothing gets better.
1:27:01>> All right, capitalism at its core.
1:27:03And you know, the thing is Macron was
1:27:04trying to get capitalism going there as
1:27:06well. It was like a big part of
1:27:08his
1:27:08>> click on this image here or this link.
1:27:10>> What are we looking at here?
1:27:11>> I mean, this is sort of another
1:27:13indicator of why GDP growth is
1:27:15staggered.
1:27:16But this is tax revenue as a percentage
1:27:18of GDP. So France also tops the list on
1:27:20tax revenue as a percentage of GDP at
1:27:2246%.
1:27:24So every dollar of GDP generated by the
1:27:27country
1:27:28is recollected as government tax
1:27:31revenue.
1:27:32>> Obviously two two drivers here.
1:27:35High tax rates or low GDP, high GDP and
1:27:39>> No, it's just the tax cuz the tax rate's
1:27:40always a function of
1:27:42income. So
1:27:43>> point is if you put this in the other
1:27:44one together, you have a confiscatory
1:27:46government
1:27:47that represents the overwhelming amount
1:27:50of the economy where you have incredibly
1:27:52deleterious outcomes. That is the
1:27:54connect the dots and you can't hide.
1:27:57>> This is the question for the United
1:27:58States is which direction do you want to
1:27:59go? Do you want to go up the chart or
1:28:00down the chart on both charts that we
1:28:02looked at? Which way do you want to go?
1:28:04Because the quality of life goes down as
1:28:06you move up the chart for all people and
1:28:09the ability to progress gets destroyed
1:28:11as you move up the chart. And as you
1:28:13move down the chart, each individual's
1:28:15ability to progress in life improves.
1:28:19You have to do it on your own, you have
1:28:21to have more agency, but your capacity
1:28:23to improve your life improves. As tax
1:28:26revenue goes down as a percentage of GDP
1:28:28and as government spending as a
1:28:29percentage of GDP goes down, your
1:28:31opportunity as an individual, the basis
1:28:34of the foundation of this great country,
1:28:36improves. And that is really the North
1:28:38Star that we should all be talking about
1:28:40is how do we get back there? How do we
1:28:42continue to move back in that direction
1:28:43in the chart? And there's very practical
1:28:45things. There's very simple things to
1:28:46do. Like we talked about in the
1:28:48>> Hold on. Hold the chart up again and
1:28:49I'll I'll just give you a little bit of
1:28:50a challenge here. I think most people in
1:28:52America, especially young people, don't
1:28:54believe what you're saying. If you look
1:28:55at the US here,
1:28:57uh
1:28:57ahead of the US, Australia, Argentina,
1:29:00South Korea, Japan, Brazil, Canada, UK,
1:29:03Germany, Italy, France,
1:29:05most people want to live in those
1:29:06countries above us. Most Americans who
1:29:10are complaining now, certainly want to
1:29:12live in one of those countries and have
1:29:13health care and want to have cheaper and
1:29:15affordable college. Now housing is still
1:29:18screwed
1:29:18>> I disagree.
1:29:19>> Oh, you disagree that most Americans
1:29:21don't want universal health care?
1:29:22>> No, no, no. Most Americans don't want to
1:29:24live in a country above us on this list.
1:29:26Absolutely not.
1:29:26>> Well, they certainly don't want to go
1:29:28below us to South Africa, Russia,
1:29:30Turkey, China, Mexico, India, Indonesia,
1:29:32or the rest of them.
1:29:33>> I've been to I've been to every single
1:29:34country on this list above us. I've
1:29:36spent a lot of time there. I've been a
1:29:37citizen of one of them and I would not
1:29:40choose to be in any country above them.
1:29:43>> Yeah.
1:29:43>> Jason, do you know what youth
1:29:44unemployment is in Brazil?
1:29:46>> Uh Brazil, I do not know. I'm going to
1:29:48guess 20%.
1:29:49>> Youth unemployment in Brazil is around
1:29:5115, 16%.
1:29:52>> Yeah, that's that's right in the street.
1:29:54>> And and it's and it's rising. And it's
1:29:56rising. It's almost at 20%. US is still
1:29:59fortunately below 10%.
1:30:01>> Yeah, 8 9% now.
1:30:02>> And the same is true in many of these
1:30:04other countries. Like there just isn't a
1:30:05growing economy that can support a
1:30:08youthful employment system in many of
1:30:10these countries. So, I don't know if I
1:30:12would agree with your statement most
1:30:14people want to go live in those other
1:30:15countries because if they wanted to,
1:30:16they could use a credit card, buy a
1:30:17plane ticket, and go do that.
1:30:19>> I think most people want what those
1:30:21countries deliver. Most people
1:30:23>> They they want what those what those
1:30:24countries say they would be willing to
1:30:26pay more tax for it.
1:30:27>> Or they'd certainly be willing for us to
1:30:28pay more tax to achieve it.
1:30:30>> It's not in full scope. The full scope
1:30:32of what those countries deliver is not
1:30:34just an acute no-fee trip to the
1:30:37doctor's office. It's that people want
1:30:39to have the ability to do that and also
1:30:41have a job and also have the ability to
1:30:44progress. And to have all of these
1:30:45things, fundamentally, the best system
1:30:48is a smaller government system because
1:30:50the more the government grows, the less
1:30:52of that pie of things you will actually
1:30:54get. It gets worse and worse.
1:30:56>> I'm in line with your argument. I'm just
1:30:58saying I'm just saying that chart does
1:30:59not make a great That chart is not a
1:31:01good piece of evidence in this argument.
1:31:03I think most people are not
1:31:05entrepreneurial and most people want to
1:31:07pay more taxes.
1:31:08No, but I mean that I think more
1:31:11freedoms, less tax is what we all want.
1:31:13I think a large percentage of America
1:31:15wants us to pay more taxes and get more
1:31:17services. And they believe those
1:31:19countries right above us want more and
1:31:22will give them more services. Therefore,
1:31:23they don't mind waiting an extra couple
1:31:26of hours or or months to get health care
1:31:29if they get health care cuz they
1:31:30currently get no health care or they get
1:31:32bankrupted by health care. That that's
1:31:33how dysfunctional our health care system
1:31:35is in the United States. So, yeah, I
1:31:37think most Americans would disagree with
1:31:39us looking at that chart. That's just my
1:31:41belief.
1:31:42But we we're we're here to sharp iron
1:31:43sharp iron sharpens iron. I think we
1:31:46have to keep working on this value prop.
1:31:48And I think this is where I don't know
1:31:49if you guys saw my tweet this week about
1:31:51how Trump President Trump could save the
1:31:53second term. And a big part of it I
1:31:55think is he has to
1:31:57talk about housing as much as he's
1:31:59talking about data centers. I think
1:32:02raising the minimum wage is a nice
1:32:03gesture since it only impacts 10 15
1:32:05million people, the federal minimum wage
1:32:06in this country. And there are things we
1:32:08can do to
1:32:11make it apparent to the people in the
1:32:13bottom half of the country that we care
1:32:15about them and that the president is
1:32:17signing executive orders for them as
1:32:19well as for us. And us I mean the tech
1:32:21industry and and large businesses. And I
1:32:24think that's that's the big problem of
1:32:25why the president is just so lowly his
1:32:28his ratings are so low, his approval so
1:32:30low and perhaps we'll see on November
1:32:323rd what happens with this election.
1:32:34That's a big part of the the discontent
1:32:36right now is people are frustrated that
1:32:39they're not even President Trump doesn't
1:32:41even talk about them anymore. He doesn't
1:32:42talk about minimum wage, doesn't talk
1:32:44about housing, doesn't talk about
1:32:46inflation. He just says it's all going
1:32:47to be fine.
1:32:48And we're living in the golden age. But
1:32:50then he's more than willing to talk
1:32:51about our issues, data centers, crypto,
1:32:54etc. And the messaging in this country
1:32:57is really far off. It's really far off.
1:32:59And until we get somebody in that office
1:33:02who is willing to talk about the bottom
1:33:04half of the country and 50/50, you know,
1:33:07hey, we're going to talk about big tech,
1:33:08we're going to talk about uh lowering
1:33:10taxes, but we're also going to talk
1:33:11about raising uh the minimum wage. We're
1:33:13going to talk about a a plan to build
1:33:16more affordable housing as opposed to
1:33:17just more data centers. That's the
1:33:19disconnect
1:33:20>> Well, hold on. Hold on. Hold on. Hold
1:33:21on.
1:33:22>> You can break it apart any way you like.
1:33:24>> If you actually read what he tweets,
1:33:27I think he connects the dots on all of
1:33:29these things. I think that he
1:33:32has
1:33:34acted in a way that one should when you
1:33:36realize that
1:33:38the AI buildout is the disproportionate
1:33:41percentage of all GDP. It is basically
1:33:43all of it. And it doesn't just pull
1:33:46through stock performance for seven
1:33:48companies.
1:33:49There's been a bunch of data now. It
1:33:50pulls through massively increased wages
1:33:52for middle class and skilled labor. And
1:33:55then he's been telling people that if
1:33:56you're going to build these data
1:33:57centers, do it in a way where you lower
1:33:59the electricity cost, do it fully behind
1:34:01the meter, invest in the local
1:34:03communities. That also has positive
1:34:04impact. So, you know, he's trying to
1:34:07make sure he lands the plane on the US
1:34:09economy.
1:34:10>> I agree he's doing that.
1:34:12>> Whether you agree with it or not, the US
1:34:14economy is entirely and completely
1:34:16levered to and correlated with the AI
1:34:18trade.
1:34:19>> Yeah.
1:34:20I I agree that he is doing a great job
1:34:23in that. Like better than any president
1:34:25I've ever seen in my lifetime in terms
1:34:27of getting the economy rocking.
1:34:30The public, the overwhelming number of
1:34:33Americans now do not agree with what you
1:34:35just said. They believe that he is not
1:34:38does not care about the issues he care
1:34:39about. That's why he's at his historic
1:34:41low in terms of approval rating. So,
1:34:44there is a messaging problem here. So, I
1:34:47I'm not saying I should be the press
1:34:48secretary here, but somebody's got to be
1:34:50a press secretary here based on
1:34:52communications. First of all, I think
1:34:54sucks out of this administration.
1:34:56>> Let me say something about the press
1:34:57secretary. I think just for one second,
1:34:59just for a small deviation.
1:35:02We have not spent enough time actually
1:35:04acknowledging what more people should
1:35:06acknowledge more often, which is
1:35:08Karoline Leavitt is an absolute goat.
1:35:11She set the bar so high. She's so
1:35:14talented. She's for being so young, it's
1:35:17unbelievably impressive. It's almost an
1:35:19impossible job to follow her. Let me
1:35:20just say that because I don't think
1:35:21people say that enough. Second, I will
1:35:24actually go out on a limb and say this.
1:35:25I think that you are an incredibly
1:35:27dogged bulldog. I think that if you got
1:35:29to know the president,
1:35:31you would be besties with him. I I
1:35:33really legitimately think you would, and
1:35:35I think you would be an incredible press
1:35:36secretary. Now, the problem is
1:35:38the problem is that you have stupid
1:35:40ideas. And as long as and as long as you
1:35:43can avoid representing your stupid
1:35:46ideas.
1:35:46>> You think the minimum wage is a stupid
1:35:47idea? I don't think so.
1:35:48>> No, no, no. No, no, no.
1:35:49You you you It's a terrible idea. I
1:35:51agree with you. The federal minimum wage
1:35:52is not. Look at the studies.
1:35:54>> You are the epitome of an armchair
1:35:55do-gooder.
1:35:56Okay?
1:35:57And and if you could if you could quell
1:35:59that instinct and actually be
1:36:02you know, his vessel, you would be
1:36:04incredibly effective because you are
1:36:05dogged, you're great on your feet,
1:36:07you're super funny.
1:36:08Uh you'd be incredible.
1:36:10>> People are now People are now pitching
1:36:11me to do it.
1:36:12>> I don't think you'll be able to manifest
1:36:14this one, but you are you are incredible
1:36:16at
1:36:16>> I mean, we were number one the last 2
1:36:17weeks in a row, so I would not
1:36:19underestimate my manifesting powers. I
1:36:21do think Free Bird, if you if you look
1:36:23at the studies, I I think you're dealing
1:36:25with old data and maybe a little bit of
1:36:27dogged libertarian principles, which I
1:36:28too shared. If you look at what's
1:36:30happened
1:36:31>> dogged libertarian principles
1:36:33>> No, I Listen, [laughter] I I'm I'm I I I
1:36:36share some of those libertarian ones
1:36:37where I'm like, the free market will
1:36:38work it out. It's only 15 million people
1:36:40in this country who make under uh $15 an
1:36:42hour and who are impacted by the federal
1:36:44minimum wage, which has not moved in
1:36:45decades.
1:36:47I If the president came out and just
1:36:49said, "Hey, we're going to increase the
1:36:50minimum wage by $1 federal minimum wage,
1:36:52which is basically doesn't count for
1:36:54half the states cuz half the states have
1:36:55their own minimum." And he just says,
1:36:57"It would just be like a good faith
1:36:58thing to do. Hey, we're going to do add
1:37:00a dollar a year for 7 years. We're going
1:37:02to get it to $15." What happens is all
1:37:04that money gets recirculated and it just
1:37:06creates more monetary velocity in the
1:37:09country. It all gets spent at the
1:37:10company. So, it's a tragedy of the
1:37:12commons. You know, one person doesn't
1:37:14want to pay this and another person
1:37:15doesn't. Look at what happened in
1:37:16Australia. Look at the Nordics. When
1:37:18they raised
1:37:19>> Jason Jason
1:37:20Listen, California passed Assembly Bill
1:37:231228 2024 and increased the minimum wage
1:37:25in California
1:37:26>> To $20. It was too too fast. I I I'm not
1:37:29advocating from 7 to 20.
1:37:31>> So, when they did that, we lost 18,000
1:37:34fast food jobs
1:37:35in the state. And the reason is Just
1:37:37think about this. But just just But just
1:37:39think about this. The reason is
1:37:41The reason is that the fast food
1:37:43industry sells food to consumers.
1:37:46Consumers want cheap fast food. So, when
1:37:49you increase the labor rate against the
1:37:52market's demand,
1:37:53>> Yeah.
1:37:54>> then the consumer says, "I'm not paying
1:37:56$8 for a burger. I can't afford it. I
1:37:58could only afford $4." So, there's
1:38:00actually no one to pay the wages of
1:38:04those fast food workers. So, number one,
1:38:06the consumers lose because they've now
1:38:07lost their local fast food joint because
1:38:09they can't afford the new fast food. And
1:38:11number two is the employees who were
1:38:14voluntarily
1:38:15willingly working there for 18 bucks or
1:38:1816 bucks or 12 bucks, whatever the
1:38:19number is, suddenly lost their job.
1:38:21>> Yeah, so let me
1:38:22>> many ways as you want to slice this,
1:38:24this is not a dogged libertarian
1:38:25principle. This is just a very common
1:38:27sense view on the world.
1:38:28>> yeah,
1:38:28>> Which is if if the buyers are there and
1:38:30the buyers are saying, "Hey, I'm willing
1:38:32to pay more." And the workers are
1:38:34willing to say, "I need to get paid more
1:38:35to work here." The market finds a happy
1:38:38medium. But, when the government
1:38:40intervenes, it blows up the market. It
1:38:42destroys the supply. It destroys the
1:38:44demand. And people lose jobs. And that's
1:38:46what we see time and again when the
1:38:48government intervenes in market
1:38:49dynamics.
1:38:49>> Here's where you're wrong.
1:38:51Number one, I never said go to $20 an
1:38:53hour. I said go at $1 per year for the
1:38:55next seven, which will get us to 15 in
1:38:58seven years, which would would put us
1:38:59out in 2032. So, you're comparing the
1:39:02most radical program, California's,
1:39:04which I don't agree with, to the my
1:39:06common sense one. Put that now going to
1:39:08the second point. It has added
1:39:11essentially 25 cents to each meal at
1:39:14those places when you do this. And it
1:39:16makes them get rid of the jobs that are
1:39:18going to be replaced anyway, like
1:39:19cashiers. But, we still have the same
1:39:22unemployment rate in these places.
1:39:24People went to other productive jobs
1:39:26when those jobs became automated. An
1:39:27argument you and I have made here many
1:39:29times, which is people go find work. So,
1:39:32you can't have it both ways, Mr.
1:39:33Freeberg.
1:39:34>> I still I disagree.
1:39:35>> people will find more productive work.
1:39:37People do find Those people who lost
1:39:39their jobs found better work in
1:39:41statistically here in California or in
1:39:43>> to be market prognosticator where you're
1:39:45trying to determine what is the right
1:39:47price for the market. And the more you
1:39:50do that, the more you are going to
1:39:52destroy the market. Because
1:39:54fundamentally that $1 or whatever the
1:39:55thing that you do to rationalize the
1:39:57government's intervention is going to
1:39:59have some effect on some set of the
1:40:01market that is going to tip over the
1:40:03point where it no longer makes economic
1:40:05sense for the buyers to pay the premium
1:40:07price and some number of those sellers
1:40:10will get out of the market and the thing
1:40:11goes away.
1:40:12>> This would be true. This would be true
1:40:14if
1:40:15you uh applied this in an unconstrained
1:40:17fashion. In a very constrained fashion,
1:40:20um raising the bottom, all it does it
1:40:23practically is those people go spend the
1:40:25money with the same companies they're
1:40:27working for. The Amazon delivery person
1:40:29or the, you know, UPS delivery person,
1:40:32those people already have $20 an hour
1:40:33jobs, but the people who are doing those
1:40:35jobs, they just spend it at the same
1:40:36companies. So in reality, when you force
1:40:40everybody at the same time to hit a
1:40:41minimum labor uh dollar amount, and we
1:40:44already have it by the way for this very
1:40:46reason, it just gets recirculated. It's
1:40:48actually good for uh companies and we
1:40:50saw it in Australia, New Zealand, in uh
1:40:53the Nordics. As they raised it, all the
1:40:55companies did better because their
1:40:57customers had more money to spend and
1:40:59people were happy. And in a society, you
1:41:01don't just want to do things for
1:41:02economic reasons. You also want to do it
1:41:05to make people have a better life. And I
1:41:07think that's what raising the minimum
1:41:08wage does. It makes an esprit de corps.
1:41:10It makes everybody's in it together. And
1:41:12the president of the United States is
1:41:13looking at that lowest paid worker and
1:41:15saying, "You know what? We're going to
1:41:16ask Amazon, we're going to ask
1:41:17McDonald's, whoever it is, to just do a
1:41:19little bit better for but 10 million
1:41:22Americans. It has no effect. It's only
1:41:2410 million Americans." That's why I
1:41:26think it's a no-brainer. It's a
1:41:28no-brainer.
1:41:29>> We're not going to agree on this. I
1:41:30think an alternative would be to drop
1:41:32taxes for the lower bracket and give
1:41:34people a zero tax rate.
1:41:35>> Okay. Zero tax rate up to
1:41:38That that would solve everything.
1:41:39>> And when you do that, you don't have to
1:41:40raise the minimum wage. You can give
1:41:42those people that money in their pocket
1:41:43because they don't have to pay money out
1:41:45of their their pocket.
1:41:46>> Zero out the taxes.
1:41:47>> And you can do other things like credits
1:41:49for them. So, there's a lot of ways,
1:41:51Jason, to kind of get to the bottom
1:41:53>> Donald J. Trump
1:41:54>> just finish. Let me just finish. There's
1:41:54a bunch of ways to solve that problem of
1:41:57giving people access to capital and
1:41:59allowing them to progress in a
1:42:00meaningful way without forcing your way
1:42:02into the market, which I think
1:42:05always has adverse consequences. And you
1:42:07can do that through tax policy and
1:42:09incentives and other structures. But,
1:42:10let's agree to disagree and move on.
1:42:12>> a minimum wage. So, what we're debating
1:42:14here is what should it be. All right,
1:42:15everybody. We'll see you next time on
1:42:16the literal number one podcast in the
1:42:19world. Bye-bye.
1:42:21>> Bye-bye.
1:42:24>> We'll let your winners ride.
1:42:26>> [music]
1:42:26>> Rain Man David Sacks
1:42:31And as I said, we open sourced [music]
1:42:32it to the fans and they've just gone
1:42:34crazy with it. Love you, Sacks. You're
1:42:35the queen of
1:42:37>> [singing]
1:42:44>> Besties are gone.
1:42:46That is my dog taking a nice [music]
1:42:48near driveway dump.
1:42:51Oh, man.
1:42:54We should [music] all just get a room
1:42:55and just have a one big huge orgy cuz
1:42:57they're all just useless. It's like this
1:42:58like sexual tension that they just need
1:43:00to release somehow.
1:43:02What? You're a deep V. What? You're a
1:43:05pure freak.
1:43:06What?
1:43:07We need to get Mercy out of the pod. I'm
1:43:09going
1:43:15>> [music]
1:43:17>> I'm going all in.

About this transcript

This page contains the full transcript of AI Cyber Panic, Ken Griffin’s Bubble Warning, The Fall of France? by All-In Podcast, generated from the public captions YouTube serves with the video. The transcript has 20,879 words across 3,280 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to translate the transcript into 100+ languages.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, translation to 100+ languages, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.