Anthropic’s $30B Ramp, Mythos Doomsday, OpenClaw Ankled, Iran War Ceasefire, Israel's Influence

0:00 How many PRs you think are going to get

0:01 pushed to the core structural internet in 100 days?

0:04 What's the overunder number?

0:05 Cuz I'll give you a number.

0:06 You're going to say zero.

0:07 My my answer to that is I'll say like 10,000.

0:10 But it's going to be immediately

0:11 if it prevents your browser history from being

0:14 released to everybody in the world, Chamath,

0:16 that may be something that you're willing to, you know, let 100 days pass on.

0:19 I think you got Chimat's attention when you said browser history.

0:21 What about the dickpicks?

0:26 Chamat is he's going to release them himself.

0:30 We'll let your winners ride.

0:38 We open source it to the fans and they've just gone crazy with it.

0:41 Love you.

0:46 All right, everybody.

0:46 Welcome back to the number one podcast in the world.

0:48 David Freeberg is out this week.

0:50 But in his place, the one, the only, our fifth bestie, Brad Gersonner.

0:56 I mean, why don't you ever give me puts a little namaste in your payday anymore?

1:00 You used to be

1:02 I'm going to bring back the greatest moderator,

1:04 but now it's just kind of You know what?

1:06 These guys beat me up.

1:07 They beat me up and they just beat the the joy out of me doing this program.

1:13 It's because you're a Roana apologist now.

1:16 No, I We'll get into it.

1:17 Okay.

1:18 Save it for the Roana apologist.

1:21 just because I said like, "Hey, they've stopped maxing and they've started

1:26 doing like some logical things." Uh, yeah.

1:29 Okay, here we go.

1:30 It's great to be here.

1:30 Great to be here.

1:31 Good to have you.

1:32 Good to have you here.

1:33 And of course, uh, we have David Saxs is back.

1:37 Everybody wants to hear from David Sax.

1:39 We missed you last week, bestie.

1:41 We didn't beat the joy out of you.

1:42 We just try to beat some of the hot air.

1:44 Turn any any fluff that you can put on the show

1:48 that just involves you talking and saying nothing is that's the stuff we got.

1:53 Turn up.

1:54 Yeah.

1:55 Turn up.

1:55 Okay.

1:55 Yeah, we'll cut it right out.

1:56 Um we'll cut it out and we'll just put a promo in for the syndicate.com.

2:00 Thank you.

2:01 Also with us, Jamalet is here.

2:04 How's your maxing going since last week?

2:06 Did you have a a a maxing full weekend?

2:09 Did you have a good full weekend of just smoking cigars in the back

2:12 deck and not ruminating about all the chaos you've caused in the last 20 years?

2:17 I think I've done generally more good than than not.

2:22 Oh, you have.

2:23 But there's been some chaotic moments.

2:25 Don't think about it.

2:26 You can't, bro.

2:27 You can't have ups without downs, man.

2:29 It's like, what are you there to do?

2:30 Just like plate everybody and be a loser?

2:33 Are you there to be a winner?

2:34 Yes, you're in the arena,

2:36 but have you stopped going to therapy after realizing ruminating?

2:40 What's up with this uh sudden interest in maxing?

2:44 Are you like the clavvicular for maxing?

2:47 No, the world finally caught up with me.

2:49 That's it.

2:49 What do you I mean, I've been maxing this whole time.

2:52 THEY JUST DIDN'T HAVE A name for it, guys.

2:54 Wow.

2:54 Okay.

2:55 Eli's videos are really good.

2:56 I watched two more this week.

2:58 What take us through what's so appealing about not ruminating,

3:02 smoking a cigar, and just living your life?

3:05 Because what he says actually works at every level of society

3:09 and every sort of thing that you may want to achieve.

3:12 Even if you're trying to like climb the rungs,

3:17 you very quickly learn that the more you want something,

3:20 the less you're going to get it.

3:21 And I think that's like his real message is let go,

3:25 live life, and just try stuff or don't try stuff.

3:29 And I think that that detachment is really healthy for people.

3:33 I like it.

3:33 I like it a lot.

3:35 Who's the guy who says this?

3:36 I actually didn't know.

3:37 Elisha Long.

3:38 Well, Eli, I think, is how he goes by.

3:41 But he's fantastic.

3:42 He Mark YouTube channel.

3:44 Mark Andre found him and he's like, "This is this guy is the new guy.

3:49 Modernday philosopher.

3:50 He gives you a road map for how to live your life, right?

3:53 A new age sage.

3:54 What's the name of the guy?

3:55 The character's name from Dune.

3:57 I was into girls books.

4:00 I was dating girls.

4:01 He's the Lison Algib of the modern internet.

4:04 This is why we need Freeberg here is to explain these deep holes.

4:08 All right, listen.

4:09 We got a lot to get to.

4:10 Don't The basic point is build something and don't ruminate.

4:13 Okay, ruminating is just not worth it.

4:15 Just everybody go for it.

4:16 No, just do stuff.

4:17 Stop blathering in your own head.

4:18 Just do stuff.

4:19 Absolutely.

4:20 All right.

4:20 Listen, speaking of doing so, Anthropic is withholding its newest model, Mythos.

4:25 I'm using the Greek uh pronunciation, its newest model, Mythos,

4:29 uh saying it is far too dangerous for any of us to have access to it.

4:33 According to the company,

4:35 the model autonomously found thousands of vulnerabilities,

4:38 including bugs in every major operating system and web browser.

4:42 This uh little study they did included 20 year old

4:46 exploits that had been missed by security audits for decades.

4:49 Uh some examples, they found a 27y old

4:51 vulnerability in OpenBSD used in firewalls and critical infrastructure.

4:55 They found a 16-year-old bug in FFmpeg that was

4:59 missed by automated tools after 5 million scans.

5:03 The Linux kernel, all kinds of uh bugs they found.

5:07 They released a hype video hyping up why

5:10 they were not going to share this model.

5:13 Here's Dario.

5:14 Come on the program anytime, brother.

5:16 But as a side effect of being good at code, it's also good at cyber.

5:19 The model that we're experimenting with is by and large

5:23 as good as a professional human at identifying bugs.

5:28 It's good for us because we can find

5:30 more vulnerabilities sooner and we can fix them.

5:33 It has the ability to chain together vulnerabilities.

5:36 So what this means is you find two vulnerabilities,

5:39 either of which doesn't really get you very much independently,

5:41 but this model is able to create exploits out of three, four,

5:45 sometimes five vulnerabilities that in sequence give

5:48 you some kind of very sophisticated end outcome.

5:50 All right, Brad, uh, by the way, that set they're using there,

5:52 that's the same room those guys play Dungeons and Dragons in every Sunday.

5:56 Brad, you're Brad, you're an investor in this company.

6:01 Is this virtue signaling or is it reality?

6:04 Is this a good move by them to not release this model and be thoughtful,

6:08 give it to a handful of people and just find

6:11 all the bugs it can before releasing it to the public?

6:14 And we've got a lot more issues to discuss.

6:16 I I actually think they deserve a ton of credit

6:18 here and let me walk you through why, right?

6:21 They the company could have just released Mythos,

6:23 broken a lot of core things on the internet.

6:25 Often times in Silicon Valley, we say move fast and break things.

6:28 In this case, it means just releasing

6:30 the model to move further ahead of your competition.

6:33 But here the company realized it would wreak havoc.

6:35 They ran their own vulnerability testing.

6:37 They saw that it would allow offensive hacking

6:40 and people to expose browsers and browser history,

6:43 expose credit cards, you know, on the internet.

6:46 So, you know what I like about this is

6:48 they didn't need government to hold their hand on this.

6:51 We have plenty of government regulations.

6:53 They know it's in the best long-term interest

6:55 of the company and the industry, you know.

6:57 So, they set up Project Glass Wing.

6:59 It's an AIdriven, you know, kind of cyber coalition.

7:03 Apple, Microsoft, Google, Amazon, JP Morgan, 40 of the most important companies.

7:09 And their goal is very simple.

7:10 Let's spend a 100 days use advanced AI to find and to fix and to harden

7:15 these software vulnerabilities before hackers exploit them.

7:19 Now, what I think this represents, Jason, is a threshold that we're crossing.

7:24 Mythos and Spud, which is going to be out from OpenAI any day now,

7:29 which is the first Blackwell trained model at OpenAI.

7:33 They represent the beginning of what I would call AGI models.

7:37 These are models with massive step function improvements and intelligence.

7:41 Um, and they're just too smart to be released immediately, you know.

7:46 And by the way, there was nothing that said

7:47 that every time you you finish a model, you got to immediately release it GA.

7:53 So they set up this idea of sandboxing, building defensive alliances,

7:57 you know, in order to move away from that regime.

8:00 I I think it shows, and Saxon and I have talked about this a lot,

8:03 so I'm interested to hear what he thinks.

8:05 It shows you can trust the industry

8:07 and market forces in coordination with the government.

8:11 They were talking to the government about this.

8:13 But they're not relying on some top- down regulation in order to do this.

8:18 They laid out a blueprint that seems to me

8:20 very pragmatic that now that we're at this threshold,

8:23 we're going to sandbox these things.

8:25 I think that open AAI will end up doing the same thing.

8:28 I think Google will end up doing the same thing.

8:30 It's an aggressive way to keep the RA, you know,

8:33 the pressure on and and win the race

8:35 at AI while making the tradeoffs to protect safety.

8:39 So, you know, I think you're always going to have to make these trade-offs.

8:42 I think in this case,

8:43 it was a great move by Dario and team and I think they deserve a lot of credit.

8:46 Sachs, when you look at this, we had

8:48 Emil Michael on the program a couple weeks ago.

8:50 It might have been four or five weeks ago,

8:52 and we had a very thoughtful discussion about,

8:55 hey, if the government is going to have these tools, you know,

8:58 an anthropic wants to withhold them and, you know,

9:01 what is the proper relationship there, you have to think that the government,

9:06 and I know you don't speak for all parts of the government.

9:09 If you were just going to run through the game theory,

9:11 they must have gone to the government and said, "Listen,

9:13 this thing is so powerful, it can put together two or three hacks,

9:16 create a novel attack vector, and this is incredibly dangerous.

9:20 What if China has it?

9:22 And if this thing is as powerful as Daario says it is,

9:25 then this is an offensive weapon as well for us to take out,

9:29 let's just pick, you know, uh,

9:31 a pressing issue, the North Korea's ballistic missile program.

9:35 This is equivalent the way it's

9:37 being described as the Manhattan project perhaps.

9:40 So what are the chances two-part question for you

9:43 Sax that China already has this and is using it and do you think Daario

9:48 is doing the right thing by regulating themselves?

9:51 I think Anthropic has proven that it's very good at two things.

9:56 One is product releases.

9:58 The second is scaring people.

10:00 And we've seen a pattern in their previous releases of at the same

10:05 time they roll out a new model or new model card, something like that.

10:08 They also roll out some study showing really

10:12 the worst possible implication of where the technology could lead.

10:16 We saw this last year about a year ago.

10:18 They rolled out this blackmail study where

10:21 supposedly the new model could blackmail users.

10:25 There's been a whole bunch of these things.

10:26 Actually, I went back to Grock and I just asked, "Hey,

10:29 give me examples where Antropic has basically used scare

10:32 tactics and it's it's a pattern." Okay, it's a pattern.

10:38 Okay, these guys, I'm not saying it's not sincere,

10:40 but they have a proven pattern of using

10:43 fear as a way to market their new products.

10:46 And if you think back to, again,

10:48 my favorite example is this blackmail study where they prompted

10:53 the model over 200 times to get the result they wanted.

10:56 And that result was was clearly reverse engineered

11:00 and it got them the headlines they wanted.

11:02 And I would say the proof that it's

11:04 reverse engineered is we're now a year later.

11:06 There's a bunch of open- source models out there that have

11:10 the same level of capability that that anthropic model had.

11:14 And have you seen any examples of blackmail in the wild?

11:16 I don't think so.

11:18 So in other words, if that study were true

11:22 in the sense of being a likely outcome of that model,

11:25 I think you would see examples in the wild of that behavior.

11:28 And we haven't seen any of that in the past year.

11:30 Now, let's talk about this specific example with cyber hacking.

11:34 I actually think that this one is more on the legitimate side.

11:39 I mean, look, the reason why I bring

11:40 this up is anytime Anthropic is scaring people,

11:42 you have to ask, is this a tactic?

11:44 Is this part of their Chicken Little routine?

11:47 or is it real?

11:48 You know, are they crying wolf or not?

11:49 I actually would give them credit in this case

11:52 and say this is more on the the real side.

11:54 It just makes sense, right?

11:56 So that as the coding models become more and more capable,

11:59 they're more capable of finding bugs.

12:01 That means they're more capable of finding vulnerabilities.

12:03 And like one of their engineers said,

12:04 that means they're more capable of stringing

12:06 together multiple vulnerabilities and creating an exploit.

12:09 And so I do think that over say the next 6 months we're going

12:12 to have this call it one-time period of catching up where AIdriven cyber is

12:20 going to be able to detect a whole range of of bugs that maybe

12:23 have been dormant over the past 20 years across a wide range of systems.

12:28 And so I do think that there is real risk here.

12:32 And I do think therefore that having this pre-release period makes a lot

12:35 of sense where they're giving the capability to all these software companies

12:40 that have existing code bases to use the tool to detect the vulnerabilities

12:43 for themselves so they can patch

12:45 them before these capabilities are widely available.

12:49 And by the way, it won't just

12:50 be anthropic that makes these capabilities available.

12:53 We know that like let's say the Chinese open source models like Kimmy K2,

12:57 it's about 6 months behind.

12:59 So we have a window here of maybe 6 months where we're

13:03 still in this pre-release period where I think companies that have large code

13:07 bases can get advanced access to this model and uh I guess open

13:13 AI is going to release a similar thing in the next few weeks.

13:16 I do think that every company or IT department

13:19 or CISO that is managing code bases should take this seriously

13:25 and use the next few months to detect any again

13:29 like dormant bugs or vulnerabilities and and roll out patches.

13:34 If everybody does their job and reacts the right way,

13:36 then I do not think it will be

13:38 the doomsday scenario that Anthropic is sort of portraying.

13:41 But it's one of these things where the fear might end up being a good

13:45 thing in order to drive people to in order to drive the correct behavior.

13:50 So sure, I ultimately think this is going to work out fine,

13:53 but you do need everyone to kind of pay attention, use the capabilities,

13:58 fix the bugs, then we're going to get into a big arms race

14:01 between AI being used for cyber offense and AI being used for cyber defense,

14:05 but it'll be a more normal sort of of period.

14:09 Chimath, we have uh Daario and uh you know

14:12 a number of the participants here are taking this super seriously.

14:15 They're making a big statement.

14:17 Zach's very nuanced uh I think take there.

14:20 What's your take on how do these companies have it both ways?

14:24 Hey, this is shouldn't be regulated.

14:26 This should be regulated.

14:27 If this is in fact a cataclysmic, oh my god, they're going to hack everything.

14:33 What if the Chinese have this right now?

14:35 That would speak to more government either coordination, regulation,

14:39 or some kind of relationship between the CIA, the FBI for domestic stuff,

14:45 and these companies because there it is a non-zero

14:48 chance that the Chinese have an equal capability here.

14:51 We're assuming they're behind,

14:52 but who knows what they're doing behind closed doors.

14:55 So, what's your take on this?

14:56 Is it uh The Boy Who Cried Wolf, or is this the real deal?

14:59 Now, I think it's mostly theater.

15:05 Okay.

15:04 In February of 2019, when Daario was still at OpenAI,

15:10 they did the same thing with GPT2.

15:14 That was a 1.5 billion parameter model,

15:17 which sounds like a total fart in the wind in 2026.

15:21 But at that time, this 1.5 billion parameter

15:25 model was supposed to be the end of days.

15:27 And it was supposed to unleash this torrent of spam and misinformation.

15:31 And that was the big bugaboo at the time.

15:33 And so what happened?

15:34 They went through this methodical roll out over six or nine months.

15:37 They started releasing the smaller parameter models and then

15:40 they scaled up to the big 1.5 billion parameter model.

15:44 And at the end of it, it was a huge nothing burger.

15:47 If you actually think that Mythos is capable of doing what it says it can do,

15:52 two things are true.

15:54 One is a very sophisticated hacker can

15:56 probably do those things right now with Opus.

16:00 And two, if these exploits are this easy to find,

16:07 whether you use Opus or whether you use Mythos,

16:09 the reality is you'd have to shut down

16:11 the internet for about 5 years to patch them all.

16:14 So when you see like a large multi- trillion dollar gang, it's a bit of theater.

16:21 Why?

16:21 What do you think they can actually accomplish in 2 months?

16:25 Do you actually think that if there's these vulnerabilities,

16:28 it's all going to get fixed?

16:29 Let's give them six months.

16:31 Let's give them nine months.

16:33 But the reality is that capitalism moves forward,

16:36 the funding needs moves forward,

16:38 and the need for these guys to build adoption moves forward.

16:42 And that's going to supersede what this is.

16:46 So I do think that Sax is right that they have

16:49 figured out a very clever go-to market muscle here and a go

16:54 to market motion that activates hyper attention and hyper usage and so

17:00 I give them tremendous credit and I'll maintain what I've maintained before.

17:04 Anthropic is shooting the lights out right now.

17:07 This is like Steph Curry going bananas from every everywhere on the court.

17:11 These guys are hunking threes.

17:13 It's all in that.

17:14 Okay.

17:15 So huge kudos to Anthropic, but we've seen it before.

17:20 We saw it when these folks were the principal architects

17:24 at OpenAI who are now seeing the same playbook here.

17:27 I think we'll look back and I think what we'll say are these two things.

17:30 One is if we're really going to patch all these security holes,

17:33 we need to shut down the internet for some number of years,

17:36 honestly, literally years.

17:38 And the second is an advanced hacker can probably

17:41 do this today with Opus if they really wanted to.

17:46 Okay.

17:46 Hey Brad, I gota I'll get you in here for the for the last word.

17:49 I I'm going to go with Yeah, maybe they did uh Crywolf before,

17:54 but based on what I see with these models advancing and using them

17:57 and I'm using a lot of the open source ones right now from China.

18:00 I think that this is like code red kind of moment.

18:03 This is Defcon.

18:04 like we should be taking this deadly seriously and I think these companies got

18:08 to coordinate with the CIA and this is

18:10 uh equally a defensive as offensive opportunity.

18:14 Do you think this

18:15 you're asking for the nationalization of AI now?

18:18 No, I actually I I I don't think it should be nationalized.

18:20 Um although I did see people sort of insinuating that.

18:24 I think these companies need to build a group

18:26 Brad that work and coordinate with the CIA.

18:29 I assume that they're already doing this.

18:31 I'm assuming you Emil Michael and uh you know Trump and everybody

18:35 have these people in a room and that they've given the defcon

18:38 and said hey how can our government use this to stop bad

18:42 actors and this is already being coordinated with the CIA and the FBI.

18:46 I am 100% certain of that that Dario went to them and said look what we found.

18:51 This is the real deal.

18:52 I'll give you the last word on this Brad since you're

18:54 an investor in both companies and you know them quite well.

18:56 the frontier model forum which was which was put together in 23 um

19:00 is cooperating on anti- and adversarial

19:03 distillation stuff as we speak right they

19:05 don't want to make it easy on you know so Google and and open

19:09 AAI and and anthropic they're coordinating on this stuff you know there are

19:13 times where I've pushed back on anthropic because I thought it was you

19:16 know perhaps regulatory capture or something else

19:19 this is very different in my mind

19:20 right he could have easily Dario could have easily come out and said

19:23 oh my god we passed a threshold we need to have a government moratorum.

19:27 Remember, even our friend Elon called for a six-month

19:30 moratorium in 2023 because of civilization risk.

19:33 This guy didn't do that.

19:35 Instead, he said, "Okay, what what should we do?

19:37 I'm going to get 40 of the leading companies together.

19:39 We're going to spend a 100 days sandboxing, hardening the systems,

19:43 and then we're we're we're going to keep pushing forward."

19:45 What do you honestly think is going to get accomplished in a 100 days?

19:48 How many PRs you think are going to get

19:50 pushed to the core structural internet in 100 days?

19:52 What's the overunder number?

19:54 Cuz I'll give you a number.

19:55 You're gonna say zero.

19:56 My my answer to that is

19:57 I'll say like 10,000, but it's going to be immediate.

20:00 But if it prevents your browser history

20:02 from being released to everybody in the world, Chimat,

20:04 that may be something that you're willing to, you know, let a 100 days pass on.

20:08 I think you got Chimat's attention when you said browser history.

20:10 What about the dickpicks?

20:14 As Chimat is, he's going to release them himself right now.

20:17 CHIMAT'S LIKE, "HEY, CHINESE HACKERS, HERE ARE MY DICKPICKS.

20:20 Please put them out." Oh my god.

20:23 we have to be out there complimenting

20:24 when they're doing the right things or relying

20:26 on the market rather than running to the nanny state and saying do more of this.

20:29 So this to me was just an example of of a of a good balance.

20:33 I'm sure we're going to have plenty of debates about this in the future.

20:36 But you know this is one I would like to see more of.

20:38 This is why to use your word Jake I tried to have a more

20:41 nuanced take is because we have no choice but to take this seriously.

20:45 Whether it's total theater, whether it's fear-mongering,

20:49 and they do have a pattern around this, we can't take the risk, right?

20:52 And it does logically make sense that as these models

20:57 become more and more capable at coding,

20:58 they're going to get better at cyber.

21:00 And there's going to be that one time

21:02 period where you're moving from preAI to post AAI,

21:05 and you need a patch for that.

21:07 So, my guess is we're going to see a lot of patches over the next few months.

21:10 I think that that will resolve the problem.

21:14 I think this is a case where I'm going to give them the benefit of the doubt.

21:17 I I think that, you know, I've criticized him in the past.

21:21 I think that blackmail study was embarrassing to the level of being a hoax,

21:26 but I think in this case,

21:28 I'm going to give him credit and say that I think that it's legit.

21:31 So, it's not the anthropic hoax.

21:33 This could be legit.

21:34 I, you know, looking at

21:36 we have no choice but to treat it that way.

21:38 Of course.

21:39 Yeah.

21:39 I mean, even if two things could be true at the same time,

21:42 Saxs, they could have used this tactic before.

21:45 It could be performative,

21:46 like the video with the dramatic music in the background.

21:48 It does have a little bit of drama to it,

21:51 and the way they presented it is very dramatic,

21:54 but it does make logical sense that the one company that made the bet

21:59 on code bigger than anybody else would

22:02 be the one who would discover this quickest.

22:04 And you know in a 100 days that's a pretty

22:06 good um that's a pretty big advantage versus the hackers.

22:10 But let me think one more point there Jimat

22:13 the most important thing that people haven't

22:15 talked about here is the amount of code

22:17 being pushed right now because of these tools is 10x 100x in most organizations.

22:22 So we need to have this type of security embedded

22:25 in these new coding tools to do it in real time.

22:29 That's the opportunity.

22:30 There should be real time correcting of this.

22:33 If this was real, they picked the wrong companies.

22:36 Meaning, there are energy companies, folks that control nuclear reactors.

22:42 There are airplane companies that are flying hundreds of thousands of people

22:46 in essentially manufactured missiles of like streaming

22:51 gas going at 500 miles an hour.

22:53 None of those companies were the ones that were included in this.

22:57 And so I think if you really thought that this was end of days,

23:02 at a minimum we can agree maybe we should have expanded the circle a touch.

23:07 Well, maybe those are customers of the ones they're including here.

23:10 Anyway, uh this is a really important story.

23:12 We'll obviously track it in the coming weeks

23:14 to see what turns out to be reality.

23:17 And uh Daario, do come on the program at some point.

23:19 Hey uh Brad, will you get Dario to come on the program?

23:21 I've invited him like three times.

23:22 I got his phone number.

23:23 He's ghosted me.

23:24 I don't know why.

23:25 Wait, he he's ignored you?

23:26 I get I literally got an introduction from the number like

23:29 one of the number one venture capitalists in the world.

23:31 He's on the cap table very early.

23:33 He just won't respond.

23:34 I don't know why.

23:35 I would tell you Daario's podcast with Dwarkish

23:38 who I think is an excellent podcaster.

23:40 I've listened to that three or four times taken notes every time.

23:43 It is a really exceptional piece really exceptional piece of work by by by them.

23:49 All right, let's keep moving.

23:49 We got a lot on the job.

23:50 You may once again be tarred with your affiliation with us.

23:54 Poor you.

23:55 I mean, I don't care.

23:56 Literally, I I've got friends on both sides of the aisle.

23:59 I have friends of course you do.

24:01 Even JCAL.

24:02 Even JCAL has friends everywhere.

24:04 Let me ask Brad a question here just while we're on the topic of anthropic.

24:07 There was a really interesting story or tweet I

24:10 guess you could say by the founder of OpenClaw that Peter.

24:15 Peter.

24:15 Yeah.

24:15 What's his name?

24:16 Peter Steinber.

24:17 Steinberger.

24:17 Steinberger.

24:18 Steinberger.

24:18 Yeah.

24:19 renowned coder created openclaw which is kind of the thing

24:23 that launched this whole agent era now you I guess you could

24:26 say any event he said that anthropic was cutting off his access

24:30 to was it was to claw is that the next topic

24:34 this is on the docket it's a little bit nuanced everybody using openclaw would

24:39 take their $200 a month subscription

24:41 to anthropic which was essentially like a people

24:44 were using more tokens and it's an average the people from openclaw it

24:48 is very verbose and those people are 100x the usage of the average subscriber.

24:54 So he said you can't use your 200, you have to use the API.

24:58 You move from the $200 plan to the API, add a zero to your token use.

25:02 So or more.

25:03 And so they essentially anchored Open Claw

25:08 and then 10 days later or less they released

25:11 or announced their new agent technology which is

25:14 according to them a safer better version of OpenClaw.

25:17 So, hey, all fair in love and war and they

25:20 have basically shot a huge cannon across the bow of openclaw.

25:26 Wait, can you just explain that exactly?

25:27 So, so I think you're right that they systematically copied feature by feature

25:32 of open claw, incorporated that into clawed

25:34 and then the coupross was basically cutting off open oxygen.

25:39 Can you just explain exactly what they did?

25:41 Okay, very simply, when you buy a subscription to these services,

25:45 they have blended your usage across many users.

25:49 So there's, you know,

25:50 nine out of 10 users use less than the tokens

25:53 they're paying for and the top 10% use much more.

25:56 When OpenClaw became a phenomenon, the number one open source project in history

26:01 on GitHub with all of this usage, people went crazy.

26:04 And you heard me talking about how crazy I went for it.

26:07 those people with the $200 subscriptions

26:09 were using $2,000 $20,000 worth of tokens.

26:12 So they said you can no longer use your subscription to, you know,

26:16 either your professional or enterprise subscription at $200

26:20 and plug that into your open claw.

26:22 You now have to go to the API and pay per usage.

26:25 So no more like unlimited.

26:28 If you use Anthropic's own agent harness, are you part of the bundled flat rate?

26:33 You can assume that that's what they'll do,

26:35 which if you were thinking on an antirust

26:37 level might be token dumping or price dumping.

26:40 I'm not saying like I'm ratting them.

26:42 No, it's like bundling, isn't it?

26:44 Well, price dumping or bundling.

26:45 When you price something under the market price in antitrust,

26:48 that would be price dumping, right?

26:50 And if you were to bundle, it would be like the bundling issue.

26:54 Critically important.

26:55 You can use openclaw via claw API and every company

26:58 has a right to set the price for its products.

27:00 It's just saying that you were for under their current regime,

27:04 they were selling dollars for 10 cents via OpenClaw because these were such

27:08 power users and now they're just saying we have to price this rationally,

27:11 but we're happy to have you guys use the API.

27:13 So, okay.

27:14 Okay.

27:14 But Brad, when you use the OpenClaw competitor that Anthropic now offers,

27:21 correct?

27:20 Are they subsidizing that?

27:22 Are you paying?

27:22 We don't know yet because it's in closed beta.

27:24 So in other words, what I'm saying is if they charge for API usage, right,

27:29 their own first party agent harness or system,

27:32 then that would be apples to apples.

27:34 But if

27:35 if they end up charging the bundled flat rate, let's say, for their stuff,

27:40 but then charge the metered rate for third party stuff,

27:44 you could make a bundling argument.

27:46 Sure.

27:46 Sure.

27:47 And you could say it's anti-competitive assuming

27:49 that Anthropic has dominant market share in coding,

27:52 which I think most people would say they do at this point.

27:54 And assuming that it's the same product, I mean,

27:57 the reason most enterprises will probably use the Anthropic uh version

28:02 of this agentic product is because it

28:05 meets all of your security parameters, right?

28:07 So, Altimter runs, you know, a lot of stuff on Enthropic.

28:11 They're already integrated within our our data warehouse,

28:14 our data lake, things of that nature.

28:16 So just letting openclaw loose on the uh altimeter you know data

28:20 set would not be wise and so it's a different fundamental product.

28:24 No I get that and I think that anthropic has a huge

28:27 advantage let's say cloning open claw and just building it into claude.

28:31 I'm not denying that to me that would be the reason why they

28:34 don't need to do price discrimination is

28:36 because there's already a very good reason

28:39 to use the let's call it the bundled offering on a featured basis.

28:42 But the question I'm specifically asking is

28:44 whether they're giving themselves a price advantage because

28:48 I think Brad is giving the the most generous interpretation.

28:51 You're taking a more cynical one.

28:52 I'm with you, Saxs.

28:53 I'm 100% on the cynical side.

28:55 Open Claw is so powerful.

28:57 It's got so much momentum that not only is anthropic trying to ankle it.

29:03 I believe when Sam Waltman bought it,

29:04 it was uh and he didn't buy OpenClaw itself, he hired Aqua hired Peter.

29:09 I believe it was to subvert the open- source project to get Peter's next

29:13 set of genius ideas inside of OpenAI as opposed to letting them go there.

29:18 People are going to say I'm a conspiracy theorist,

29:20 but this is the number one focus and let me just give

29:23 you a list of who is trying to kill OpenClaw/compete with them.

29:27 Obviously, you have Anthropic, but also Perplexity Computer launched.

29:32 It's awesome.

29:32 I've been using it.

29:34 Anthropic has this clawed managed agents.

29:37 They dropped that on Wednesday, April 8th.

29:39 uh yesterday uh today's Thursday when we tape you you guys listen on Fridays

29:43 and then you have Hermes agent that was

29:46 released on February 25th that's also open

29:48 source and very good so that's in the open source camp Alibab is coming out

29:52 with one that's going to be based on their Quinn model then you have Elon

29:56 who said he's got something called rock computer coming out of macro hard which

30:01 is a play on words for Microsoft in addition to that Amazon and Apple are

30:05 preparing uh new releases of their uh

30:09 maxing assistants Alexa and Siri that will be

30:11 less in this new version and then nothing out of uh SAT and Microsoft yet.

30:17 So the number one goal I believe in the large language

30:21 model frontier model space is to kill this open source product.

30:26 No, I mean come on like why they're building multi-functioning agents

30:30 that can move from answering questions to actually doing something for you.

30:35 like you got to do that because that's what consumers and enterprises wants.

30:38 It doesn't mean that it's about killing OpenClaw.

30:41 just this is an obvious thing right to do it

30:44 but this is a giant movement to stop it because

30:47 this is the equivalent of having an open-source Android like player

30:51 in the market and that could be incredibly disruptive these I

30:54 believe open source is going to win the day on the large

30:56 language models and take 90% of the token usage and I

30:59 think the entire frontier model space could be undercut by open

31:02 source and I think they realize that SLMs the the smaller

31:06 language models that are verticalized now that will run on you know,

31:10 desktops and laptops and is even starting to run on the top ones.

31:14 That is their biggest competitive threat and I hope it happens.

31:17 All due respect to your investments, Brad,

31:19 I think this technology and the interface is uh you know, he placed bets,

31:23 but I I think it's imperative that the agent level,

31:26 which is essentially your entire life, you don't give that to Anthropic.

31:30 You don't give that to OpenAI.

31:32 That's your entire business, your entire life.

31:34 It is foolish for you, Brad, to give your entire business and all

31:38 the knowledge you have to anthropic through that.

31:40 unless you're just doing it to boost

31:41 your um your your investment in those companies.

31:43 But I would be very concerned if I was you with putting all

31:47 of your knowledge that you've earned over

31:49 a lifetime into any of these large language models.

31:52 All right, Jake, let me ask you.

31:53 Can I ask a question?

31:54 Thank you for that impassioned monologue.

31:57 Um actually, I want to ask my TED talk.

31:59 I Yes, thank you for that TED talk.

32:01 Um I have a yes no question for each of you.

32:06 Do you believe that anthropic has dominant market share in coding?

32:09 right now?

32:10 Yes.

32:11 No, no.

32:14 In in coding, yes,

32:15 they had the lead, but not that they had the lead, but not dominating.

32:18 I think it's a trillion dollar market,

32:19 and these guys have less than 10% of it today.

32:22 So, it's hard to make a case that

32:24 What percent of coding tokens do you think

32:26 that anthropic is providing the market right now?

32:29 Greater than 50%.

32:30 Yeah, that's true.

32:31 Okay, that's called dominant market share.

32:33 Uh, I don't know about that.

32:35 More than 50% of the market.

32:36 You got to look at what you got to look at what the TAM is.

32:39 with the Tan, right?

32:41 There are a lot of people who provide, you know,

32:44 that that are in this tiebreaker before we move on to the next.

32:48 I'm not saying it's a permanent condition,

32:50 but if you're telling me that today

32:52 Anthropic is delivering over half of the coding tokens,

32:57 that's clearly a dominant position in the market for coding.

32:59 It's an early market.

33:00 It could change, but

33:01 if I were representing them, David, I would say nine months ago,

33:04 everybody t called us uh, you know, out of the game.

33:08 We were being destroyed by open AI in three months.

33:10 Now people are saying we have dominant market position.

33:13 This is the fastest changing most competitive market in the world.

33:17 I think it would be very hardressed to walk into, you know,

33:20 some district court make the case that these guys

33:22 have somehow already formed a monopoly against Amazon,

33:25 Google, Microsoft, Open AI, etc.

33:29 Well, I'm not saying it's a it's already a permanent monopoly,

33:31 but I am just asking about market share.

33:34 And I do think you guys all agree that Shimov, go ahead.

33:37 They probably have 50 to 60% market share because

33:40 I think codeex is actually quite broadly used as well.

33:45 But that belies the more important point which is AI

33:49 enabled coding I think is still 5% of the broad market.

33:53 So it's kind of a nothing burger.

33:54 Yes, they're leading but they're leading in something that isn't that big yet.

33:58 Now you would say how could it not be big?

34:01 And what I would say is because most of the stuff

34:03 that's being written is still white sheet denovo code.

34:08 And I think the ugly truth is I don't care what model you have,

34:12 but the long horizon ability for any of these models to actually

34:16 build enterprisegrade software is still shiit

34:22 And that's the actual lived experience.

34:24 Not for me, but when I call on our customers,

34:28 half a trillion dollar banks, hundred billion dollar insurance companies,

34:31 none of these guys are like, "Wow,

34:33 it just works out of the box." It doesn't work.

34:36 So, most of it is still handtuned.

34:39 So, until I can honestly tell you that we can

34:42 point a model at this with the right guard rails,

34:46 which I can't today, what I would say is it's a small market

34:50 that will become large as these models become better.

34:54 But we are in the world where we have

34:57 50 years of accumulated tech debt as a world.

35:01 And I suspect when you enumerate the number of lines that that represents,

35:04 it's hundreds of trillions of lines of just

35:07 pretty marginal mediocre code to bad code.

35:11 On top of that, we have all these legacy languages.

35:14 I'll tell you one of our customers,

35:16 they have to go and get 60year-old pensioners

35:18 to come into the office to interpret cope.

35:21 No, I'm not joking.

35:22 This is a snowball for trend.

35:24 This is a hundred billion dollar a year

35:26 revenue company and that's how they solve these problems.

35:30 It's not opus just solves it.

35:32 So I I would just keep in mind that most of the tech debt

35:36 in the world that exists 99% of it is still poorly addressed by these models.

35:41 We are untying this Gordian knot.

35:44 It's going to take decades to do it right.

35:46 So all the breathlessness about all this other stuff,

35:48 I really think it's not where the money is.

35:50 It's not the big time stuff.

35:51 And you can tell me, "Oh yeah, it's going to be the future." And I would say,

35:55 "Tell this business that's a hundred billion dollars a year

35:58 of revenue and 50 million billing relationships that all of a sudden

36:01 you're going to open claw your way to a solution." It's

36:05 Not to say that you can't have a great chief of staff,

36:07 and not to say you can't do some useful stuff and trickery and, you know,

36:11 have a good knowledge base.

36:12 I'd like that, too.

36:14 But the core things that your lived experience sits on today

36:19 is a mess of tech debt that will get very slowly replaced.

36:22 And that's just the reality of life.

36:25 And there are competitors that are extremely disruptive.

36:28 I'll tell you about one.

36:29 We talked about Bit Tensor Tao on this program

36:31 a couple weeks ago when we had the um Jensen interview.

36:34 You brought it up actually Chimath.

36:35 There's a there's a project that's subnet 62.

36:38 It's called Ridges AI.

36:40 And what they're doing is a competitor that is

36:43 not only open- source but anybody can contribute to it.

36:47 They spent about a million dollars in tow like rewards and in 45 days they

36:51 hit 80% of what Claude 4 is and they did that in under 45 days.

36:56 The way that works is they give rewards for people who and they can

37:00 do this anonymously make that coding product

37:03 which is like codeex or claude code better.

37:06 that flywheel is racing right now with participation in the same way Bitcoin is.

37:11 So you're going to see a lot of open- source and these crypto open-source

37:16 combinations and uh anybody who's not investigated

37:20 this, I highly recommend you investigate this.

37:23 I do think you're right about one specific thing.

37:25 I would put zero, literally the probability

37:28 zero of any important company worth anything

37:31 more than a dollar having and outsourcing

37:34 their production code to an open source project.

37:36 That'll never happen.

37:37 However, what will happen though is when you look

37:40 at the cost of training this 10 trillion parameter model

37:45 on Blackwell and when you look in the future let's

37:49 just say in six or nine months that a 15

37:52 or 20 trillion perm model is going to get

37:54 trained on Vera Rubin I think Jason where you

37:57 are right I have zero and just to be

37:59 clear I have no investments in this at all I'm to be so super clear

38:04 I'm just observing because another project other than Bit

38:07 Tensor that someone brought up to me is Venice.

38:09 The concept of opensource training and orchestration is

38:14 a hugely disruptive idea which is the complete

38:18 orthogonal attack vector to this idea that you

38:21 have to raise tens and tens of billions

38:23 of dollars to train your models because if the capital markets run out of 10

38:28 and 20 billion dollar checks to give people

38:31 the only solution is to be totally distributed.

38:34 I tend to agree with you Jason that there is going

38:36 to be at some point a very successful open source project for pre-training.

38:42 Absolutely.

38:43 Will there never ever be an open- source way where a real company that has

38:47 any skin in the game says here guys

38:50 re-engineer my codebase as an open source project.

38:52 Never going to happen.

38:53 Yeah, I I think the coding tools will.

38:55 And if you look at the history of open source, Brad,

38:57 you actually I think had a lot of bets in this space.

39:00 Linux, Kubernetes, Apache, Postgress, like Terraform,

39:04 like these open source projects are deep inside of enterprises.

39:07 Deep.

39:08 And we're sitting here 15, 20 years ago, the same argument was made.

39:12 Nobody will ever adopt these inside the enterprise.

39:14 You got to go with Oracle, whatever.

39:16 And fair enough, many people do.

39:18 But I think this is this $29 ridges um subscription to do this versus 200.

39:24 It's starting to take hold inside of startups.

39:28 And that's where I always look at the tip of the spear.

39:30 Startups love to, you know, use open source products.

39:33 I think this could be the next big thing.

39:35 But listen, I I I invest in things that have a 90% chance of going to zero.

39:40 So do your own research.

39:41 No crying in the casino.

39:43 Can I just make a a final few points?

39:46 So just just quickly so number one is with respect to this market

39:49 for code or code tokens whatever you want to call it

39:53 it might be 5% today meaning 5% of the codes

39:56 AI generated versus human generated I think it's going

39:59 to 95% I mean I bet any amount of money

40:02 on that the only question is when probably over

40:04 the next few years so that's point number one

40:06 point number two is it's possible that if you're

40:10 the early leader in coding as a AI model company

40:15 let's say you have 50 to 60% of market share.

40:17 You have the most developers using it.

40:19 Therefore, you have the most access to code bases.

40:22 You might get the most training tokens.

40:25 There is a potential flywheel there where you

40:28 can see the early market leader consolidating its

40:30 lead because it's generating the most code tokens

40:33 and it's getting access to the most existing code.

40:36 Now, I'm not saying for sure that's going to happen.

40:38 is possible that the other guys catch up,

40:40 but I think there is a possibility of a flywheel there and strong,

40:44 I guess you'd call it data scale effects, things like that.

40:47 So, I do believe that the market for coding tokens could be monopolized.

40:52 Third, Anthropic's revenue run rate,

40:55 as based on what I can tell and what's been publicly released,

40:58 is the fastest growing revenue run rate at scale that I think we've ever seen.

41:03 Uh, we perfect segue.

41:04 It's the next story.

41:05 Okay, maybe

41:06 pull up the the tweets.

41:08 But this thing is ramping at a rate we've never seen before.

41:12 We can get into that in a second.

41:13 But just one last final point

41:15 is I think it's pretty clear that where we go from here

41:18 is agents and coding gives you a huge step up on agents because

41:24 you know one of the main things that agents need to do

41:26 is is write code to be able to enable them to complete tasks.

41:30 Correct.

41:30 And so if it is the case that coding is this huge market that's going to be

41:37 dominated by one or two companies and then

41:40 that leads to another huge market which is agents.

41:44 My point is just I think all these companies need to behave in a very clean way

41:49 and not engage in tactics that later the government

41:52 might say you know what that was anti-competitive.

41:54 Everyone should just I think play fair.

41:56 Do not engage in discrimination against other people's products.

42:00 engage in fair pricing.

42:02 I'm not accusing anyone of breaking any of the rules,

42:04 but what I'm saying is that eventually

42:06 the government's going to look at this market

42:08 with the benefit of 2020 hindsight and I

42:11 think everyone should just basically, you know, keep it

42:14 keep your nose clean.

42:15 Keep it tight.

42:16 Keep it tight.

42:17 Keep it tight.

42:17 Tight is right.

42:18 I think is an excellent point.

42:20 Let's talk about the revenue ramp of Anthropic.

42:24 This is just unprecedented.

42:26 Anthropic's revenue run rate has topped 30 billion with a B.

42:31 Early 2023, they turned on revenue.

42:33 They started charging for API access.

42:35 End of 2024, they're at a billion dollar run rate.

42:38 February 25, they launched Claw Code.

42:41 That was the starter pistol.

42:42 Mid 2025, $4 billion run rate.

42:44 End of 2025, $9 billion run rate.

42:48 Just a couple of months later in April, $30 billion run rate.

42:52 Yes, that's right.

42:53 Triple.

42:54 Uh and the way they did this is enterprise

42:57 uh customers are a major part of the spend.

43:00 Dario announced a couple of months ago that there's

43:02 over a thousand enterprises paying over 1 million annually.

43:06 This is truly mindboggling when you think about it

43:10 because those are the most coveted customers in the world.

43:13 These are the big fish that you

43:14 just uh when people are running enterprise software,

43:18 they they dream Slack dreamed of getting these million-dollar customers.

43:21 Uh Salesforce dreams of getting these million-dollar customers.

43:24 Brad, you're an investor.

43:25 I guess uh Sam famously on BG2 asked you

43:28 to sell your uh OpenAI stock back to him.

43:32 You didn't.

43:32 You demired, but you're an investor in both.

43:36 How shocking is it to you to place both of those bets

43:40 and then see one of them come from so far behind?

43:43 You know, Chat GPT has 900 million users.

43:46 I don't know if they've they've passed a billion officially yet,

43:49 but they are the Verb, right?

43:50 They're the Uber.

43:51 They're the Xerox.

43:52 They're the Polaroid of AI, but they didn't go after the enterprise.

43:58 Daario made that and Daario worked.

44:00 He was the co-founder of OpenAI.

44:02 He left and according to the New Yorker

44:04 story that came out from Ronan Farrell this week,

44:06 he was basically left because of his disgust in working with Sam Alman.

44:13 Your thoughts?

44:14 Well, you know, before we go down the OpenAI rabbit hole,

44:17 let's just really contextualize like what's going on here.

44:20 You know, check I I I have this additional chart.

44:22 you showed one, you know, they added 4 billion of revenue in January,

44:26 7 billion in February, 11 billion of annualized run rates, um,

44:30 or 10 or 11 billion in March, just to put in perspective,

44:34 that's data bricks plus Palanteer combined

44:37 that they added in a single month, right?

44:39 So we started with everybody at the start of the year ringing

44:43 their hands including you know Gurley and others saying we're in a big bubble

44:47 asking whether the AI revenues would show up to justify all of this investment

44:51 and bam you have the largest revenue explosion in the history of technology.

44:56 So the company's plans were to end the year

44:59 at about a $30 billion ex exit run rate.

45:02 They got there by the end of March

45:05 right and I suspect that it's continuing in April.

45:08 So you have to ask what's going on and what's the big

45:10 so what the first thing for me is that model and product capability

45:15 just hit this threshold we talked about earlier near AGI whatever the hell

45:18 you want to call it and everybody like alimter said damn this is

45:22 so good I have to have it this is no longer about my IT

45:26 budget this is about labor augmentation and labor replacement and by the way

45:30 co-work is growing even faster than Claude go at the same stage

45:36 of development So what it showed is we have a near infinite TAM.

45:41 It turns out that the TAM for intelligence

45:43 is radically different than anything that we've seen before.

45:47 And I think the best example of this, right?

45:50 This is millions of self-interested parties, consumers,

45:54 enterprises, a thousand now over a million dollars.

45:58 Right?

45:58 It's not that there was some great go

46:00 to market and anthropic that all of a sudden,

46:02 you know, they snuck up and blew everybody away.

46:04 No, it was companies demanding the product.

46:07 They're getting throttled on the product.

46:08 Why?

46:09 Because it's so good.

46:10 It makes them better at their business.

46:12 We are all self-interested actors.

46:14 And when millions of those people are all making the same decision,

46:18 there's a huge tell.

46:19 And the tell here is that the TAM is

46:21 as big as Daario and Sam and others have been saying.

46:25 We knew intelligence was going to scale on the exponential.

46:28 The question was whether revenue will scale on the exponential,

46:31 and that's what we're seeing.

46:32 And remember, they're doing this with only 1 1/2 to 2 gawatt of compute, right?

46:38 These guys are massively compute constrained.

46:41 They're each going to be adding 3 GW of compute this year.

46:44 And so that will unlock they would be growing even faster.

46:48 But for that, and then Jason, to your point about the open source models that we

46:52 all want to be a part of this solution, I've talked to a lot of big companies,

46:57 65 to 70% of their token consumption is open-source model, right?

47:01 are these cheap Chinese and other tokens.

47:04 So these revenue ramps are happening while

47:07 the world is already using open source.

47:09 This is not frontier only.

47:11 This is Frontier plus open source.

47:13 We're going to see massive token optimization over the course of the year.

47:17 But what happens on this Jebans paradox is

47:20 the co the unit costs right of intelligence is plummeting.

47:24 Not the cost of tokens.

47:26 The unit cost of intelligence is plummeting because

47:28 the capabilities of these models is so much better.

47:31 I look at what it does for Altimeter day in and day out.

47:34 I talked to a major uh company yesterday.

47:37 They're on a run rate to do a hundred million

47:39 of token consumption this year on about $5 billion in opex.

47:43 They think that we're now nearing peak employment in their company,

47:47 but that their token their intelligence consumption,

47:50 okay, let's not call it token consumption, right?

47:52 because tokens may go up a lot,

47:54 but their intelligence consumption is going to go up, you know, a lot.

47:58 So, I would leave you with this.

48:01 We're early to Chimas's point.

48:03 We have low penetration of the global 2000.

48:06 We have low penetration of the use cases.

48:08 We have low penetration of of within the use cases that they're already using.

48:14 And the models are only getting better.

48:15 So I think when you look out toward the end of the year,

48:18 I would not be shocked if you see Anthropic

48:22 exiting this year at 80 to 100 billion in revenue.

48:26 And by the way, doing it at the same time that OpenAI, who is also on the wave,

48:30 they'll be releasing an incredible model in the next imminently.

48:34 They're going to be on that wave and you're

48:36 going to see an inflection in their revenues as well.

48:38 Okay, Chimath, question one has been answered.

48:42 The question of hey, does this stuff actually have utility?

48:45 that went from a question mark to an exclamation point.

48:47 Of course, it's got utility.

48:48 People are getting value from it.

48:49 And it might be variable.

48:50 Some people get more value than others.

48:51 Number two, the revenue ramp was a big question.

48:54 Now, that's turned into an exclamation point.

48:56 The final piece of the puzzle that you've

48:58 brought up many times is can this be profitable?

49:00 And these companies are burning through a large amount of cash.

49:05 So, what is your take on when these companies can get out of the J curve?

49:09 We talked about this, I think, three episodes ago.

49:11 I estimated like we're going to be looking

49:13 at $4500 billion in investment into these data

49:16 centers at a minimum and then they have

49:19 to climb out of that to get to profitability.

49:22 So what are your thoughts on these becoming profitable companies?

49:26 Do you remember the investor that published this list Jason where he put all

49:33 of the terms you talk about when one

49:36 of the terms you can't talk about is profit.

49:38 It's a list where it's like if you can't talk about free cash flow,

49:41 you talk about IBIDA.

49:42 When you can't talk about IBIDA, you talk about margina.

49:47 When you can't talk about that, you talk about revenue.

49:49 And then when you can't talk about revenue, you talk about gross revenue

49:55 bookings.

49:54 So you can kind of figure out, I think, where we are in any part of any cycle

50:02 by just indexing into what does everybody talk about.

50:06 I think where we are is we are between gross revenue and net revenue.

50:11 That's where the discussion is.

50:14 Okay.

50:14 There was another article I think today in I think maybe it

50:17 was the information that tried to categorize

50:20 and distinguish that anthropic presents gross,

50:24 open AI presents net.

50:26 They're different.

50:27 We don't know what the various take rates are.

50:30 So they're saying that there's a difference.

50:32 If it's not true, there's been no clarity provided by these companies.

50:35 So, at a minimum, you have this confusion where there's the breathless talk.

50:40 Then there's people that don't even know the difference between

50:42 actual recognized revenue and run rate revenue and how to multi.

50:46 I mean, so we're definitely there, okay?

50:48 We can quibble about the details,

50:50 but we are not at the place where people are like, "Oh, here's your steadystate,

50:53 you know, free cash flow margin,

50:54 and here's what your EBA does." We're never we're we're years from that.

50:58 They're gonna have token maxing IBA like IB at the Wii.

51:03 The thing that we need to understand is

51:04 how gross margin negative is this revenue growth.

51:07 We don't know that and at least we don't as outsiders.

51:11 Brad might know.

51:12 Brad may know.

51:13 I I I I will tell you think about this.

51:16 What are their big cost inputs?

51:17 The number one cost input is the cost of compute.

51:20 Cost of compute.

51:21 Right?

51:22 I just told you they only have a gigawatt and a half of compute.

51:24 and they have that gigawatt and a half of compute whether they

51:27 have a billion in revenue or whether they have 80 billion in revenue.

51:31 So you might actually expect to see

51:33 these companies their gross margins are exploding higher

51:36 like the fastest increase in gross margins

51:38 I've probably seen out of any technology company.

51:41 So this is not gross margin negative you're saying?

51:43 No definitely not gross margin negative.

51:45 And what I would tell you so that they must be hugely profitable then

51:48 well you may see accidental why I call it accidental profitability.

51:53 They may not be able to spend this revenue fast enough chamath on compute.

51:57 And remember it's only 2500 people.

52:00 Google crossed this revenue threshold when they had 120,000 people.

52:05 These guys have 2500 people.

52:07 So the only thing you can really spend money on, right, is compute.

52:10 And they can't stand up the compute fast enough.

52:13 But none of this foots to me then to be honest

52:15 because if you were on a threshold of 90% plus gross margin,

52:20 I'm not saying it's there.

52:22 I'm not saying it's 90% plus.

52:23 I'm just saying it's gone from meaningfully negative

52:26 18 months ago to, you know, very very positive.

52:30 I've seen rumored out there 50% is what you're saying.

52:33 The trend is there.

52:34 Let me just say this.

52:37 I think if you're an incumbent, you want the cost of compute to go down.

52:42 I think if you're not an incumbent, so specifically, who do I mean?

52:45 Meta, Google, and SpaceX.

52:51 I think those three people who have all three of them,

52:54 well, sorry, Meta and Google have a fortress balance sheet.

52:57 I think by the end of June, SpaceX will also have a fortress balance sheet.

53:02 What they will want to do is they will want to make

53:04 this a compute problem because they will

53:06 control the the conditions on the field.

53:08 You already see this today.

53:11 Yeah.

53:10 Meta's models today, what people's general reviews are it's okay,

53:15 but the one thing that people say is it's incredibly performant.

53:18 The model quality is okay, but the performance is great,

53:21 which speaks to Meta's huge advantage.

53:23 They have a massive compute infrastructure.

53:24 So if you're if you're not open AI and anthropic,

53:28 they'll want to make this a capital problem because then they can win it.

53:31 If you're anthropic and open AI,

53:33 you want this thing to be as efficient as possible.

53:36 I think where we are is very much in the early innings.

53:39 And we're bumbling around talking about gross margins and you know revenues.

53:43 We are not at profitability.

53:44 And what is true for Facebook and what was true

53:47 for Google was irrespective of where they got to a billion.

53:51 Who g cares?

53:53 They were profitable by year three and they never looked back.

53:57 I was there.

53:58 I remember it was glorious.

54:00 The the cost the cost of building uh you know AI totally

54:05 stipulate is radically higher than the cost

54:07 of building retrieval at Google, right?

54:09 Like it's just a fundamentally more expensive problem.

54:12 But I will tell you that there's a lot

54:14 of FUD out there about negative gross margins.

54:16 I mean Jason, you started the segment

54:18 by saying they're burning through large amounts of cash.

54:20 I think people are going to be shocked at the burn

54:23 how low the burn levels are at these companies.

54:26 Anthropic or Open AI.

54:27 Yes.

54:27 And and I would say at Open AI as well like they're if they're on you know if

54:31 they do $50 billion this year again just look

54:33 at the number of people they have revenue per people.

54:36 It's pretty low and the inference cost is plummeting.

54:38 Inference cost is down by 90% year-over-year.

54:42 And so just finally I want to make respond to this point

54:45 about gross versus net uh this this tweet that Chimath was referencing.

54:50 Okay, so there's a certain percentage,

54:51 a smallalish percentage of Anthropics revenue,

54:54 right, that they distribute through the hyperscalers

54:56 and like a lot of arrangements, whether it's Snowflake or Data Bricks or others,

55:00 you pay a commission, right, uh on on that.

55:03 I will just tell you that you're talking

55:05 singledigit percentage of total revenue of these companies.

55:08 So the gross versus net thing isn't what's being reported.

55:11 like the apples for apples is pretty easy and if you want to be conservative

55:14 on it take down Anthropic's revenue by you

55:17 know five to 10% which you know again I

55:19 don't I think it's better to gross up OpenAI's revenue but any way you do it

55:23 I just think it's a distraction from what's

55:25 really what's really going on here happy to

55:27 s you have any thoughts on this uh massive revenue ramp

55:31 yeah I mean I want to go back to a point

55:33 that Brad made because I think it was just really important

55:36 and I want to just underline it consider where we were

55:39 at the beginning of the year and What everybody was saying is

55:42 that AI was a big bubble and the evidence they would

55:46 point to was the fact that hundreds of billions of dollars was

55:50 going into capex that needed to be spent on these data centers

55:54 and there was no evidence of significant revenue to justify that spend.

55:58 Where was the ROI?

55:59 By the way, as an aside,

56:01 the same doomers who were saying that AI was

56:03 in a bubble were also the ones who were saying

56:05 that AI was so powerful it's going to put us

56:07 all out of work and it's going to, you know, take over from humanity.

56:11 I mean, in other words,

56:12 they couldn't decide if AI was too powerful or not powerful enough.

56:16 But putting aside that contradiction,

56:18 they clearly were making this case that AI was this big bubble

56:22 and that there'd be no payoff or justification

56:26 for this massive capex that's being spent.

56:29 And I think we're starting to see here there is justification for it.

56:32 Uh we're seeing it just in this one vertical of AI which is coding.

56:37 We're again seeing the fastest revenue growth in history.

56:40 It's utterly unprecedented.

56:42 And this is just one category or vertical of AI.

56:46 We know that agents are coming next and the enterprise

56:50 adoption of that is going to be absolutely massive.

56:53 So, I guess what I'm saying is that this is early

56:56 proof for I think the thing that makes Silicon Valley special,

57:00 which is we're willing to basically bet on things that just

57:05 intuitively on a gut level we know are the next big thing.

57:09 We're not that spreadsheet driven.

57:10 Actually, Silicon Valley believes that if you build it,

57:13 they will come and is willing to finance that build out.

57:16 And that's basically what's been happening.

57:18 Again, just the top four hyperscalers,

57:20 $350 billion of expected capex this year on its way,

57:24 I think Jensen said 1 trillion by 2030.

57:27 So, Silicon Valley, whether it's big companies, whether it's founders,

57:30 they're always willing to bet on this next big thing.

57:33 They're not like Wall Street.

57:34 They don't need, you know, specialist to tell them where to go.

57:38 They know where the technology is going and they make their bets based on that.

57:42 And I think that there is going to be a big payoff for this.

57:46 And I think it's the thing that's going to make

57:48 our economy and the United States in general remain

57:51 extremely dynamic and in the lead on this thing is

57:54 that we are willing to make those kinds of bets.

57:57 And I think it's going to pay off big time.

57:59 Yeah, clearly.

58:00 Hey, um Brad, you didn't answer my question

58:03 about the vibes over at OpenAI versus Quad.

58:07 Open AI is um I wouldn't say reeling but there's a lot

58:11 of hand ringing going on a lot of employees leaving a lot of people

58:15 who are wondering like is our strategy the winning strategy of like consumer

58:20 first they shut down Sora you know unwinding the Disney deal and really

58:25 trying to get the company focused and it's kind of like I mean

58:28 listen the New Yorker story was a bit of a rehash so I

58:30 don't think we have to go into the blowby-blow because we covered here

58:33 three years ago but the truth is a lot of the great founders,

58:38 co-founders of OpenAI and a lot of the great

58:41 contributors are now at Anthropic and other large language models.

58:46 And in the secondary market, OpenAI is trading lower than the last valuation.

58:51 And Anthropic is trading significantly above the $380 billion.

58:57 So maybe talk a little bit about this competition,

58:59 this Microsoft versus Apple, this Google versus Facebook.

59:03 Well, let's let's start with immense credit where credit is due.

59:07 Anthropic was literally counted out of the game last year.

59:10 Y, right?

59:11 And here they come over the last 12 months

59:13 and and and they've kicked OpenAI's ass over the last 90 days, right?

59:17 And what did Anthropic do?

59:19 Anthropic made choices.

59:20 No multimodal, no video, no hardware, no chips, no building data centers.

59:25 They said, "We're just going to focus on coding and co-work.

59:28 We think that is the path to AGI

59:30 and and and and ASI." They executed their butts off.

59:33 They took the lead.

59:34 2500 people tight pulling on the ore in the same direction.

59:40 But I think you would be seriously foolish to count out open AI, right?

59:44 And I think we're we're we're at peak open AI FUD.

59:47 And I'll tell you, it starts with great researchers and great models.

59:50 And I think when you see the Spud model, they're about ready to release.

59:53 I think it's going to be an excellent model.

59:56 Shows that they're firmly on the wave.

59:59 Um, if you look at what's going on with Codeex,

1:00:02 incredible ramp on Codeex, fastest ramping model with 5.4,

1:00:06 I think 5.5 or Spud, whatever we're going to call,

1:00:08 it's going to be an even faster ramp.

1:00:10 Have you seen Spud?

1:00:11 Have you used it?

1:00:12 Have you gotten a preview?

1:00:13 People are using Spud, right?

1:00:15 So, it it is being previewed and so

1:00:18 So, you're talking to people who've used it and what are they telling you?

1:00:21 They're telling us that it's an incredible model on par with Mythos, right?

1:00:25 and that it's a a very usable model in terms of um how it's packaged.

1:00:30 I will say that back to David's point now this is

1:00:34 the most important point I think anybody can take away here.

1:00:38 This is not zero sum.

1:00:40 The TAM of intelligence is dramatically larger than any TAM we've

1:00:45 ever seen in our investing careers over the last two decades.

1:00:48 Right?

1:00:49 And if you're on the wave, which Open AAI is,

1:00:52 you are going to be selling into the world's biggest TAM,

1:00:55 they are going to build a very big company.

1:00:57 I'm a buyer of the shares today.

1:00:59 Notwithstanding all of the vibes that you describe,

1:01:02 I think these companies are firmly on the wave.

1:01:05 They are jarred.

1:01:07 They are sitting there saying, "What did we do wrong?

1:01:09 And how do we get our mojo back?" They want to compete.

1:01:12 It is embarrassing to people on the research

1:01:14 team and the product team over there.

1:01:15 So, I'm not saying there's not a real awakening occurring there,

1:01:19 but I think that's what the case is.

1:01:21 And by the way, to Chamas's point, do not count out Meta, right?

1:01:25 I think Meta is absolutely in this game.

1:01:27 Google is absolutely in this game.

1:01:28 Elon is absolutely in this game.

1:01:30 And if you're

1:01:31 got some stuff dropping shortly that's going to be very impressive.

1:01:34 If you're on team America,

1:01:35 the fact that we have five frontier models competing against each

1:01:39 other and David made sure they

1:01:41 weren't throttled by excessive government regulation.

1:01:44 We have mythos come out.

1:01:46 It's a self-imposed safe harbor, you know, to harden our system.

1:01:50 It wasn't a call for moratoriums or getting the government involved.

1:01:53 We have the type of competition that's causing us

1:01:56 to accelerate our lead against the rest of the world.

1:01:59 We can't take our eye off the prize.

1:02:00 We got to stop adversarial distillation and we need

1:02:03 to make sure that we're distributing our products around the world.

1:02:06 But I view this as really good for team America.

1:02:10 Well said.

1:02:10 And here is your poly market IPOs before 2027.

1:02:14 Obviously SpaceX at 95% uh Cerebrus at 94% and uh hey number five

1:02:22 on this list 51% chance that Anthropic goes out before the end of the year.

1:02:26 44% chance that OpenAI comes out before then.

1:02:30 All right here is the closing market cap

1:02:34 for Anthropic on Poly Market only $158,000 in volume.

1:02:39 So, Chimath, when you put in 400K, you're going to really tilt this market.

1:02:44 78% chance that it's above 600 billion, 19% chance that it doesn't go out.

1:02:50 So, it's looking like this will be a decent investment for you.

1:02:54 Brad, what valuation did you get into Anthropic at?

1:02:57 We first invested in I believe it was the uh 30 or $150 billion round.

1:03:04 So, this will be a 7x 5x for Altimeter L, please.

1:03:08 Congratulations.

1:03:08 I mean, no, listen.

1:03:09 I I I again, there are lots of people who were there before us

1:03:12 and who are on the board and who are going to do much better than that.

1:03:15 What' you put in?

1:03:16 50.

1:03:17 What' you put in?

1:03:18 No, we've got billions in both companies.

1:03:20 Uh billions in both companies.

1:03:22 Oh my lord.

1:03:24 I think there's this existential thing going on in venture today.

1:03:28 David could talk about it as well.

1:03:29 I mean people can't they're extraordinarily nervous about you look

1:03:34 at the IGV stock index down 30% year to date

1:03:38 down 5% today all software stocks plummeting right venture

1:03:43 capitalists are terrified to invest money in anything other than

1:03:48 these frontier models and things like SpaceX or military modernization

1:03:52 finding something that's out of harm's way of AI right

1:03:56 where you can count on the terminal value to Chamas

1:03:59 insights over the last few weeks is very difficult to do.

1:04:02 That's why you see this crowding.

1:04:03 So, we've taken a barbell approach, right?

1:04:06 We've got a lot in what we think are the most

1:04:08 important companies that are on the frontier and then we're

1:04:10 betting with on really small teams that we think have

1:04:13 very defensible businesses in a world of uh you know, AGI.

1:04:17 But it's what happens to all these enterprise software companies?

1:04:20 Do they become PE takeouts?

1:04:22 Do they get consolidated?

1:04:24 um or do they just have to adopt

1:04:26 these AI technologies and and and solve this problem

1:04:30 of hey the frontier model is just going

1:04:32 to solve for whatever these niche software companies do.

1:04:36 I think the market's probably being a little too pessimistic

1:04:39 with respect to at least some of these software companies.

1:04:42 I mean, obviously, there's going to be big differences

1:04:44 in the quality of the modes of these companies.

1:04:48 And so, look, software is going to be a lot cheaper and easier to generate,

1:04:53 but I'm not sure that was the competitive advantage of a lot of these companies.

1:04:57 So, there's probably a little bit of the baby

1:04:59 being thrown out with the bathwater right now,

1:05:00 and there probably are some value buys in enterprise software.

1:05:04 I think the interesting question here and we've

1:05:07 been talking about this for a couple of years

1:05:08 in the pod is just where you see the AI

1:05:12 value capture being in terms of layer of the stack.

1:05:15 Remember where we started it was really just the chip

1:05:18 layer of the stack was where all the value capture was.

1:05:20 It was basically Nvidia was the first company

1:05:22 to be worth multiple trillions of dollars because of AI.

1:05:26 And for a while it looked like that's

1:05:28 where all the value capture was going to be

1:05:30 because OpenAI for example was losing so much

1:05:32 money and Anthropic wasn't on the radar as much.

1:05:35 Now we're seeing wait a second um you know it's not

1:05:38 just the chip companies it's also the hyperscalers are now benefiting

1:05:42 and now we're seeing at the model layer it looks like

1:05:45 Enthropic and Open AI they're all going to be huge beneficiaries.

1:05:48 I think the next question is at the application layer of the stack.

1:05:52 Okay.

1:05:52 Well, now does all that value capture just get eaten

1:05:54 by the model companies or are there applications that get turbocharged?

1:05:59 I guess you could say that Palunteer is already one of them, right?

1:06:02 It's an application company that's

1:06:04 been turbocharged by these model capabilities.

1:06:07 Who else will be a big beneficiary?

1:06:09 Is it again, is it all going to be at the model

1:06:11 layer or will you see an explosion of value at the application layer?

1:06:16 I'm hoping obviously that it'll be at all layers of the stack.

1:06:19 PC beneficiaries.

1:06:20 But to me, that's a really interesting question right now.

1:06:23 Yeah.

1:06:23 What happens to Salesforce, HubSpot, you know, Oracle, right down the line?

1:06:27 David, uh, Chimati, your thoughts here, uh,

1:06:29 on the the layers here and where the value is captured.

1:06:34 It's too early to tell.

1:06:36 Too early to tell, right?

1:06:36 And energy we kind of put into sort of data center as well,

1:06:40 but that's obviously been a clear winner.

1:06:42 Little housekeeping here.

1:06:43 Liquidity, put a little Tiffany in here.

1:06:45 uh producer Nick D is sold out.

1:06:49 There's a wait list of hundreds of people, but it is what it is, folks.

1:06:52 If you snooze, you lose and top tier speakers are coming.

1:06:56 Uh it's going to be great.

1:06:58 We'll get a an update from But I think Brad,

1:07:00 you're going to be joining us again.

1:07:01 Yes.

1:07:01 For liquidity.

1:07:02 I have an update.

1:07:03 That's probably not your headliner, though.

1:07:05 I'm probably not your headliner.

1:07:06 No, but you always score so high.

1:07:08 Every event you've spoken at, you've been either number one,

1:07:10 two, I don't think you've ever dropped to three.

1:07:13 Go ahead, Sham.

1:07:14 Make your announcement here.

1:07:17 Nat sent me an article from Wikipedia about

1:07:19 penile links when you guys are talking about breaking news.

1:07:22 Showing me showing me that I'm in the large category.

1:07:25 Top 5%.

1:07:26 She highlighted it.

1:07:27 Top 5%.

1:07:28 Okay.

1:07:28 And that's with Is that with Nano Banana or without?

1:07:32 Is that She just texted dummy.

1:07:35 It's clogged.

1:07:35 My apologies.

1:07:36 Clogged.

1:07:38 Oh.

1:07:38 All right.

1:07:38 This is why Jamath isn't afraid of the cyber is because nothing's going

1:07:42 to come out that's more embarrassing than what he says himself on the box.

1:07:44 He's like Bezos.

1:07:45 When Bezos got hacked, HE'S LIKE, "GUYS, I GOT HACKED."

1:07:49 SO, I saw the agenda for this thing.

1:07:51 It's incredible.

1:07:51 Congrats to you guys.

1:07:52 I mean, like the uh like just the fun of being in Napa,

1:07:56 all the poker, all the the dining experience.

1:07:58 This is five star all the It looks really six-star.

1:08:01 It's a man level because Chimath was, I dare I say, belligerent in his demands.

1:08:09 He said, "This has to be six-star or I will not show up." Jake Al.

1:08:12 I said, "Okay, boss, get to work." And uh, Chimath, what do you got any?

1:08:17 No mids.

1:08:18 This is all elite.

1:08:19 And for the hundreds of people who are on the wait list,

1:08:21 I am sorry, but we have a capacity issue.

1:08:23 We'll try to get you in for next year.

1:08:24 But Chim, give us some updates here.

1:08:26 You have any updates that you want to share?

1:08:28 because you are running programming for liquidity 2026 up in Yon.

1:08:32 Look, it's going really well.

1:08:35 Really excited to hear all of these great folks speak.

1:08:38 I think the next two will release today.

1:08:40 Brad Gersonner and Thomas Leaf of COTU of CO2.

1:08:44 That's a great get.

1:08:46 We also have I think three people confirmed for their best ideas pitch.

1:08:49 Really interesting folks.

1:08:51 They each run between one and six or seven billion awesome

1:08:56 superstar compounders early in their career.

1:08:59 This is a new zone chamat.

1:09:00 It's great.

1:09:00 So right now we have Bill Aman, we have Andre Carpathy, we have Dan Loe,

1:09:06 we have Thomas Lefont, we have Brad Gersner,

1:09:08 we have Sarah Frier and more to come.

1:09:11 We will announce more.

1:09:12 There might be one or two surprises.

1:09:13 Jay Cal and a couple and a couple of surprises.

1:09:17 Yeah, we we don't announce all the speakers.

1:09:18 Jay Cal's got a couple of surprises coming.

1:09:22 And if you didn't get in to liquidity, apologies.

1:09:24 You're on the wait list.

1:09:26 We are going to be hosting the fifth annual all-in

1:09:31 summit in Los Angeles September 13th to the 15th of Sax.

1:09:36 You going to come to that?

1:09:39 Allin.com/events.

1:09:40 Sax, you should come to that.

1:09:41 I've been advised that I can attend business.

1:09:43 I can be in the state for business reasons.

1:09:46 Okay, there you go.

1:09:47 Then we'll see you at liquidity and the summit.

1:09:49 Correct.

1:09:49 That's that's big news.

1:09:50 Now we just got a bunch of Sachs stands who are racing.

1:09:53 Uh and now we're going to get Sachs

1:09:55 at This is what happens every year behind the scenes.

1:09:58 Sachs at the last minute says, "Oh, I have four speakers and I have 72

1:10:02 people who need tickets and then the whole team

1:10:04 has to like do a fire drill 48 hours before the event." Okay, here we go, guys.

1:10:08 We're going to go to the third rail here.

1:10:10 We got to catch up on the Iran war.

1:10:12 Here's the latest.

1:10:14 Two weeks into a ceasefire have started just two

1:10:17 days ago at the taping of this VP JD Vance,

1:10:20 friend of the pod is a and some special consultants Wikoff

1:10:25 and friend of the pod Jared Kushner are headed to Islamabad,

1:10:29 the capital of Pakistan for talks this very weekend.

1:10:33 So while you're listening to this event,

1:10:34 they are going to be working on the peace deal.

1:10:36 Easter Sunday, Trump posted a truth stating, "Open the straight,

1:10:41 you crazy bastards, or you're going to be living in hell.

1:10:44 Just watch." Praise be to Allah.

1:10:46 On Tuesday morning, Trump posted uh a another threat on social media.

1:10:50 A whole civilization will die tonight.

1:10:53 Never to be brought back again.

1:10:54 I don't want that to happen, but it probably will.

1:10:57 Tweets were obviously discussed uh a lot over the last week.

1:11:00 He gave him an 8:00 p.m.

1:11:02 deadline.

1:11:03 At 6:30 p.m.

1:11:04 POTUS announced on Truth Social that he had agreed.

1:11:08 President Trump had agreed to a two-week ceasefire if Iran opens the straight.

1:11:13 He also said, "Hey, listen.

1:11:14 We got the straight.

1:11:15 Maybe there'll be a toll booth, but we'll take the majority of the toll

1:11:18 and we'll split it with Iran." Here's the quote.

1:11:20 We received a 10-point proposal from Iran,

1:11:23 and we believe it's a workable it is a workable basis on which to negotiate.

1:11:29 And apparently Netanyahu took the ceasefire to mean

1:11:32 level Lebanon dropping 160 bombs in 10 minutes yesterday.

1:11:37 Saxs, uh, you were out last week.

1:11:38 Everybody wants to know your position on the war.

1:11:41 I'll hand it off to you.

1:11:42 What are your thoughts on how on the two

1:11:44 ceasefire and everything that's occurred up until this point?

1:11:47 Well, look, I have to preface what I'm about to say,

1:11:50 which is I'm not part of the foreign policy team at the White House.

1:11:55 And the last time I commented on the war on this show,

1:11:58 it somehow made international headlines that Trump advisor says XYZ.

1:12:05 And I'm not a Trump adviser on this issue.

1:12:08 I think that'd be a fair headline to write

1:12:09 if it was a technology issue, but this is not.

1:12:12 So whatever I say is just my personal opinion,

1:12:15 but then the media is going to somehow portray it or attribute it

1:12:19 to the White House or try and create an issue out of it.

1:12:21 So, I feel like I'm limited in what I can say except

1:12:24 that to say that I think it's terrific that we have the ceasefire.

1:12:29 I think it's great that there's going to be

1:12:31 this meeting in Islamabad to hammer it out.

1:12:36 And I think what the president's accomplished so

1:12:38 far with the ceasefire is it's a great

1:12:40 thing because what happens with these wars is they take on a life of their own,

1:12:45 meaning they tend to go up the escalation ladder, right?

1:12:48 And there's a lot of podcasts that are discussing the so-called escalation

1:12:51 trap and supposedly there are stages to this based on historical patterns.

1:12:55 And so I think it's actually very hard to pull

1:12:57 out of these things and I give the president tremendous

1:13:00 credit for negotiating the ceasefire that we've achieved so far

1:13:03 and then sending the team to hopefully work this out.

1:13:07 Brad, actually my first trip to the Middle East was when you

1:13:09 and I uh maybe four years ago when Thank you for taking me.

1:13:12 What is your take on where we're at here?

1:13:14 I think we're just wrapped up week six of this and we're going into week seven.

1:13:18 First, on March 4th, I tweeted the Trump doctrine in Iran.

1:13:21 Massively destroy all military capa capabilities.

1:13:24 Kill the people building lethal weapons to use against us and get out.

1:13:29 Reserve the right to do it again if needed.

1:13:30 Zero efforts to build Misonian democracy.

1:13:33 Iran's going to have to build what comes next.

1:13:35 And I think what the market has said right if you look back at last year

1:13:40 on tariffs Jason the top tobottom draw down was

1:13:43 about 15% on the NASDAQ intraday is down 22%.

1:13:48 Okay, the draw down in this period over Iran was

1:13:52 only down about 5 to 7% on S&P and NASDAQ, right?

1:13:56 So, the market has said, listen, re trust Trump at his words.

1:14:01 He said he's not going to get into an entangled war here.

1:14:04 I think he terrifies the hell out of people with his tweets about,

1:14:07 you know, destroying civilization and all this other stuff.

1:14:10 But I think people, even though they don't like to hear it,

1:14:13 they've resolved for themselves that when he says he's going to get out,

1:14:16 he will in fact get out.

1:14:17 Of course, there was a lot of hand ringing,

1:14:19 but if you look at the markets today,

1:14:21 we basically bounced all the way back from where

1:14:24 we were pre-Iran on both the S&P and the NASDAQ.

1:14:28 If in fact we land the plane,

1:14:30 if JD lands the plane, and by the way, on Lebanon, yes,

1:14:32 they were bombing yesterday, but Netanyahu has now said that you're going

1:14:36 to have direct government talks between Israel and Lebanon.

1:14:38 So, if the if we land the plane on these two things,

1:14:42 I think it's off to the races in the market.

1:14:44 By the way, while everybody's focused on Iran, stay tuned.

1:14:47 I think we're getting close to a deal on Ukraine, Russia, right?

1:14:51 Venezuela is, you know, kind of going seemingly very well.

1:14:55 I think there's also going to be news on Cuba.

1:14:57 You could envision a world there's risk to the downside.

1:15:00 Certainly, I will stipulate,

1:15:02 but you also have to pay attention to the risk to the upside.

1:15:05 If you land the plane on those things heading into America 250 July 4th,

1:15:09 the market could really take off.

1:15:11 All right.

1:15:11 Well, let's uh maybe uplevel this a little bit

1:15:14 and talk about why we're in this war to begin with.

1:15:17 And that's the big discussion amongst both sides of the aisle.

1:15:20 On Tuesday, the New York Times dropped an inside the room piece

1:15:24 on how President Trump made the decision according to this report, if it's true.

1:15:30 I know some people don't uh subscribe to the New

1:15:32 York Times anymore or think it's fake news,

1:15:34 but how Trump decided to basically follow Netanyahu into this war.

1:15:38 On February 11th, Netanyahu met with Trump at the White

1:15:41 House where he gave him a four-part pitch on attacking Iran.

1:15:44 JD Vance, according to the story, if it's true, disclaimer, disclaimer,

1:15:48 warned Trump that the war could cause

1:15:50 regional chaos and break apart Trump's MAGA 2.0,

1:15:54 the Trump 2.0 coalition we talked about here, the big tent.

1:15:57 And that's turned out actually to be true.

1:15:59 There's been a bunch of hand ringing from Megan,

1:16:01 Kelly, Tucker, Carlson, right on down the line.

1:16:04 Rubio was anti-regime change, but he was largely ambivalent,

1:16:08 according to this story about the bombing campaign.

1:16:10 Susie Wilds, chief of staff,

1:16:12 said she had concerns about gas prices before the midterms.

1:16:15 Pretty good uh advice there.

1:16:17 And General Dan Kaine, chairman of the Joint Chiefs of Staff,

1:16:21 said this of Netanyahu's pitch.

1:16:23 Quote, "Sir, this is, in my experience,

1:16:25 standard operating procedure for the Israelis.

1:16:27 They oversell and their plans are not always well-developed.

1:16:30 They know they need us and that's why they're hard selling.

1:16:33 If you put this together with Rubio's walked

1:16:35 back comments at the start of the war, we knew, this is quote from Rubio,

1:16:41 we knew there was going to be an Israeli action.

1:16:45 We knew that would precipitate an attack against

1:16:48 American forces and that's why we did it.

1:16:51 I had Josh Shapiro on the All-In interview show

1:16:53 and um uh he talked a lot about this.

1:16:58 There is a big underpinning here, Chimath,

1:17:00 that the United States foreign policy is being driven by Netanyahu.

1:17:05 Every Jewish American person I've talked

1:17:08 to feels Netanyahu is not doing Jewish American

1:17:11 and Jewish the Jewish diaspora any favors here by his approach to these wars.

1:17:15 What are your thoughts on why we got into this and how we get out of it?

1:17:23 I mean, the person that decides is the president of the United States.

1:17:26 some foreign leader isn't getting to call the shots in the United States.

1:17:32 I think very practically speaking,

1:17:35 the markets are effectively pricing in that this was

1:17:40 a small blip for whatever people think.

1:17:44 That's just what the best prediction market that we have is telling us.

1:17:48 I think that's important to acknowledge that we're probably in the endgame here.

1:17:53 And the second thing to acknowledge is if I was Israel,

1:17:56 I would really be concerned that unless I help find an offramp quickly,

1:18:02 the risk that Israel loses America

1:18:05 as a predictably steadfast ally could go down.

1:18:08 And I think that that's problematic for Israel

1:18:11 far more than is problematic for the United States.

1:18:13 So all of that kind of tells me that we will find an offramp.

1:18:18 A because I think economically it makes sense and then B geopolitically I

1:18:22 think Israel will want to make sure

1:18:24 that this doesn't burn a long-standing relationship.

1:18:29 Yeah, that that seems to me to be the major issue

1:18:32 here is Americans basically do not want to be in this war.

1:18:36 Americans do not want a forest policy

1:18:38 being influenced to the extent they believe.

1:18:42 I'm not putting my belief in here.

1:18:43 Americans believe we are being dragged into this by Israel and that Israel

1:18:48 has too much or Netanyahu specifically has far too much influence.

1:18:51 And then people believe the anti-semitism that's occurring here.

1:18:54 Josh Shapiro gave me a lot of push back on this.

1:18:57 Uh but all the Jewish Americans I talked to say

1:18:59 Netanyahu is causing with his actions in Gaza, Lebanon, Iran.

1:19:04 Uh he's gone too far and it's

1:19:06 causing the anti-semitism we're experiencing uh today.

1:19:09 So you can make your own decisions about that.

1:19:11 Any final thoughts here, Brad,

1:19:13 on the American foreign policy being influenced too much by Israel?

1:19:18 No, it's the discussion.

1:19:20 I I mean, listen, um kind of like Sax said earlier,

1:19:25 um I think that we will ultimately be judged by the outcomes, right?

1:19:31 And that everybody is an armchair pundit today on, you know,

1:19:36 uh the the the approach that we're taking in these two different places.

1:19:40 I think we could be on the verge of a massive transformation of the Gulf States.

1:19:46 You went there with me, Jason.

1:19:47 Saudi, Qataris, Kuwaitis, Emiratis.

1:19:50 I've talked to a lot of them this week.

1:19:51 I think they're very hopeful and optimistic.

1:19:53 I think you could bring Iran into the fold.

1:19:56 But listen, I'm an optimist on all of this stuff.

1:19:58 I I just want to remind people, doing nothing in Iran had tremendous risks.

1:20:05 Doing nothing in Venezuela had tremendous risks.

1:20:09 So, it's not as though this was uh, you know,

1:20:13 something that I I I I think wasn't well calculated,

1:20:17 but I think we have to let the cards be

1:20:19 played and and and then let history be the judge.

1:20:22 But I think there's uh a risk in both directions,

1:20:25 but I'm going to remain optimistic.

1:20:26 All right.

1:20:26 You uh said in the Gaza situation,

1:20:29 we should have a wide birth for criticism of Israel and Netanyahu.

1:20:32 What are your thoughts on this belief here

1:20:35 in the United States now in this discussion that Israel is having far too much

1:20:39 influence over the United States foreign policy?

1:20:42 Well, I noticed in my feed today that Naftali Bennett,

1:20:46 who's a major Israeli politician who was a former prime minister,

1:20:49 tweeted polling that showed that Israel was becoming

1:20:54 very unpopular in the US and he was

1:20:56 expressing concern about that and expressing the need

1:21:01 to to basically address that or fix that.

1:21:04 So, I think you're starting to see Israeli politicians raising that as an issue.

1:21:10 And I think that's probably a good thing.

1:21:13 Yeah, there it is.

1:21:14 And it's really cool actually how X

1:21:15 now just automatically translates things from foreign languages,

1:21:19 in this case, Hebrew, and it puts it in your feed.

1:21:21 So, yeah.

1:21:22 So, here's Naftali Bennett, former prime minister,

1:21:24 saying, "This is a serious situation.

1:21:25 There's a lot of work ahead of us to fix everything." Now,

1:21:29 obviously, this is not Netanyahu.

1:21:30 this is one of his um political opponents.

1:21:33 But yeah, I mean this is something for Israel to consider

1:21:37 and think about and I think that they would improve their popularity uh

1:21:42 if they got behind the ceasefire and I have no indication

1:21:45 that they won't but that would certainly be a good place to start.

1:21:48 I have to say just as an aside, this auto translate feature has done more

1:21:54 for understanding across borders than anything I've ever seen.

1:21:58 And it is the most impressive tech feature I I've seen released in years.

1:22:04 Putting AI and large language models aside for people who don't know

1:22:08 what's happening because of Grock being really good at doing auto translate.

1:22:12 They've taken the pockets of the best of what's happening in Japan,

1:22:17 what's happening in Israel, what's happening in France,

1:22:19 and they're surfacing it auto translated.

1:22:21 Then when you reply as an American to somebody in Japan,

1:22:25 they see it autorated as well, which has led to people who don't speak

1:22:29 the same language engaging on X in a very nuanced, fun, interesting way.

1:22:36 And that for as a truth mechanism is just absolutely extraordinary.

1:22:41 I think this is going to have such a profound effect.

1:22:43 Maybe Elon and the X team should get like a Nobel Peace Prize award for this.

1:22:47 I think it's going to change.

1:22:48 I mean, I hate to be hyperbolic, but have you been using this feature, Chamat?

1:22:52 Has it been coming up in your feed?

1:22:53 And which language is up in your feed right now?

1:22:59 English.

1:23:00 Okay.

1:23:00 So, you're not part of the translation thing.

1:23:02 Brad, has this hit your feed yet?

1:23:03 And and which regions are you?

1:23:05 Definitely.

1:23:06 Definitely see it in on the Middle East stuff.

1:23:08 Um, and uh, you know, I've seen on Chinese,

1:23:11 I've seen it on on the Russian Japanese super helpful.

1:23:14 Let me tell you, bass Japanese is a whole another level of beast.

1:23:18 Whoa.

1:23:18 Man, base Japanese makes like Fentes and Alex Jones seem tame.

1:23:25 They're like, look at this group of people.

1:23:27 Insert whatever group of immigrants you like.

1:23:29 And they're like, this is unacceptable behavior.

1:23:32 This is not Japanese culture.

1:23:34 These people need to be get the hell out of Japan.

1:23:36 It is wild, folks.

1:23:39 And if you don't have an X account, you are missing out.

1:23:42 Go to X.com and sign up for this reason

1:23:44 only because you think about the velocity.

1:23:47 Like journalists are not even taking the time

1:23:49 to translate and cover what's going on in those areas.

1:23:51 And this is happening automatically in real time.

1:23:55 So you start thinking about what happened in Ukraine.

1:23:57 If you had people Russia and Ukraine doing

1:24:00 this and having conversations with each other, it would be wild.

1:24:03 You're like a such a good hype man.

1:24:04 The problem is you hype buttered bread the same way you hype a nuclear reactor.

1:24:08 And so it's hard to really tell, you know,

1:24:10 what you're really hyping because your level of excitement,

1:24:13 the intonation is exactly the same.

1:24:15 Yo, man, there's nothing better than a slice of great toast.

1:24:18 I mean, if this is very this in in a way, it is like sliced bread.

1:24:21 It's very simple, but it is so powerful in the experience.

1:24:25 This has been It is true.

1:24:26 X is better today than it's ever been.

1:24:28 And remember, they have 70% fewer employees than

1:24:32 they had the day Elon walked into the building.

1:24:34 And so if there were ever a debate

1:24:37 about this, like, and I remember everybody saying, "Oh, it's going to tip over.

1:24:40 Oh, it's going to be a crappy experience."

1:24:43 The fact of the matter, here's we are a few years later, 70% fewer employees,

1:24:48 and every other company in Silicon Valley is looking at that.

1:24:51 I think for a lot of these tech companies, we've hit peak employment.

1:24:54 We're going to create a tremendous number of new jobs,

1:24:57 but for the existing jobs,

1:24:59 these companies are all realizing they can do more with less.

1:25:02 Nikita Beer just tweeted that they're about to go

1:25:04 ham on these bot accounts that auto reply.

1:25:08 Yes.

1:25:08 Those those literally ruined my feed.

1:25:10 That's why I went to subscriber mode

1:25:12 in my replies and it's it's worked out great.

1:25:14 Yeah.

1:25:14 No, shout out to him and um to Chris

1:25:17 Saka who was in tears at what happened to Twitter.

1:25:19 You It's gonna be okay, Chris.

1:25:21 Sorry.

1:25:22 No more tears.

1:25:24 You only let subscribers respond to your tweets.

1:25:26 I I do 50/50.

1:25:28 Sometimes I'll just let it rip and get chaos.

1:25:30 And then other times I have 2,000 paid subscribers.

1:25:32 I give all the money to charity, like 30 grand a year.

1:25:35 And it's just wonderful to get to know

1:25:36 the same 2,000 people out of my million followers.

1:25:39 It's kind of like having this little subset.

1:25:41 So sometimes I'm like,

1:25:42 I don't have time to deal with a 100red or 200 or 300 replies.

1:25:45 You have a million.

1:25:47 That's incredible.

1:25:48 I mean, it's just I mean, you have two million.

1:25:50 I think Sax must have a million, right?

1:25:52 You have a million, right, Sax?

1:25:52 Only only Brad.

1:25:53 How many you have now?

1:25:54 You're getting popular.

1:25:56 You built a couple.

1:25:57 Got a couple hund.

1:25:59 What's your Oh, your alt cap altca.

1:26:01 I'm at 1.4 million.

1:26:02 What are you at, Jacob?

1:26:04 Have I surpassed you?

1:26:06 I think you have.

1:26:06 I'm like 1.1.

1:26:08 What it cost me to get my real name, Jason?

1:26:11 Uh, I know a guy.

1:26:13 Find out.

1:26:13 You're 1.1.

1:26:14 Yeah, I made it to 1.4.

1:26:15 I don't know how that happened exactly.

1:26:17 Just having the number one podcast in the world.

1:26:20 Uh, another amazing episode of the number one pop and Chimath has two million,

1:26:27 but that's only because he engages he has

1:26:30 just incredible moments of uh engaging with his haters.

1:26:34 Oh my god, the the the replies that Chimath sometimes drops are so great.

1:26:39 I love Chimoth goes I light them up.

1:26:42 I light them up.

1:26:43 He lights them up.

1:26:44 Then you had somebody who was like, "Oh my god,

1:26:46 I was in the casino and you told me to bet black,

1:26:48 so you bet black, so I bet black and I lost my money."

1:26:51 And so you're responsible and then you paid for the kids college.

1:26:54 He has two young girls and so I I funded their college accounts.

1:26:59 I thought that was hilarious.

1:27:00 Just as obviously I'm very happy for him and his two daughters.

1:27:03 I'm even more happy at how much it'll anger

1:27:05 all these other goofball dorks living in their mom's basement.

1:27:10 Yes.

1:27:09 Who'd literally have no take they take no responsibility for their lives.

1:27:14 And uh they should enjoy those Hot Pockets.

1:27:16 By the way, for those folks in their mom's basement,

1:27:18 the Hot Pockets and the Fish Sticks are ready

1:27:20 and you get one more hour of Xbox from mom.

1:27:23 All right, listen.

1:27:24 We missed you, Freeberg, but this is the best episode in two years.

1:27:27 Uh Freeird at the end of the show.

1:27:32 And we will see you all at the liquidity summit except

1:27:35 for the 400 people on the wait list who aren't going to get in.

1:27:37 We got an email from the guys at Athena because we were just Oh my god.

1:27:42 the they they're they're going to hire like 500 new Athena assistants.

1:27:46 Yes, they had a thousand people after last

1:27:49 week when we mentioned how much we love Athena.

1:27:51 Go to Athena.com.

1:27:52 But that's amazing.

1:27:53 Those are like 500 hardworking men and women who are like working

1:27:57 in the Philippines.

1:27:59 Sax have great jobs.

1:28:00 Sax, I'm going to get you a couple Athena assistants as a birthday present.

1:28:03 That's what I'm going to get.

1:28:04 You're going to love this, Sax.

1:28:06 H Athena assistants are the best.

1:28:08 Congratulations to my friends over there.

1:28:09 All right, everybody.

1:28:10 We'll see you next time.

1:28:11 Love you boys on tonight favorite podcast.

1:28:19 Let your winners ride.

1:28:23 Rain and we open source it to the fans and they've just gone crazy with it.

1:28:30 Love you queen of winners.

1:28:39 Besties are gone.

1:28:41 That is my dog taking out your driveways.

1:28:46 Oh man, my appetiter will meet.

1:28:49 We should all just get a room and just

1:28:50 have one big huge orgy cuz they're all just useless.

1:28:53 It's like this like sexual tension that we just need to release somehow.

1:29:02 That's going to be good.

1:29:02 We need to get merch.

1:29:12 I'm going all in.