0:00
[Music] Heat. Heat. Heat. Heat. N. [Music] [Music] Heat. Hey, Heat. Heat. Heat. [Music] [Music] Heat. Heat. N.
Search this video
2:19
[Music] [Music] [Applause] [Music] We came to this [Music] world to reach the stars.
3:14
We came to this [Music] world to see our [Music] future.
3:23
We came to this [Music] world to reach the stars to shape our future.
3:40
We came to feel the new [Music] [Music] [Applause] [Music] You [Music] [Music] know, [Music] [Applause] [Music] feeling it.
4:27
[Music] You're watching TBPN.
4:48
Today is Monday, April 21st, 2025.
4:51
We are live from the temple of technology, the fortress of finance, the capital of capital, the institute of iron, the hall of hypertrophy.
5:01
Uh we wanted to kick it off with a post from my good friend Sahil Bloom.
5:06
He's been on the show before as we highlighted his book, but this one hit hard, especially today.
5:10
He says, "Here's the truth.
5:12
If you're half in, you're actually all out.
5:14
Even 90% in gets you nowhere.
5:18
There's something magical in the last little bit simply because so few are willing to do it.
5:24
That's where you unlock new levels to the game.
5:26
And it does not take talent, just courage.
5:30
Well, you're certainly being courageous today, John. Thank you. Congratulations. Yeah, big news. I'm all in on TBPN. I am entirely in. I'm completely in TBPN. Uh full-time.
5:39
The announcement went out today.
5:42
Uh fantastic experience as an entrepreneur in residence at Founders Fund.
5:46
It was a great twoyear two year three month run. Incredible run.
5:50
Uh fantastic team over there.
5:53
Uh they're really going to be suffering.
5:54
They're they're they're just kicking off a new $4 billion fund.
5:56
It's going to be rough, but you know, I think they'll be I think they'll be okay.
6:00
The management fees on a $4 billion fund should maybe be able to find a replacement. Yep. Yep. For sure. For sure.
6:07
Uh but honestly, I mean, fantastic team over there. Founder mode. Founder mode.
6:12
Uh super close with everyone.
6:12
And of course, uh, we have Delian as a regular on the show already. Uh, Trey is coming back.
6:19
He's going to be one of our first in-person guests soon.
6:21
We're getting Scott Nolan talking about general matter soon.
6:25
So, uh, obviously the teal bucks will still be flowing.
6:29
We're not getting paid directly by Founders Fund anymore, but we are getting paid by RAMP, our sponsor, and one of a major founder Fund position. That's right.
6:37
So, it's all one hand washes the other over here in Tech Media.
6:40
Remember, we are 100% corporatebacked.
6:42
We'll never ask you for a dime.
6:44
And uh yeah, so if you're a company and you want to sponsor us, hit us up because we we love our sponsors.
6:52
Anyway, speaking of sponsors, uh the market's in turmoil.
6:54
If you're trying to get in on the action, head over to public.
6:57
com investing for those who take it seriously.
6:59
They have multiasset investing, industryleading yields, and they're trusted by millions.
7:02
Uh the big news today, stock market is mocked.
7:07
Absolutely mogged as the trade war escalates.
7:09
Global markets tumbled amid renewed USChina trade tensions and an unprecedented clash between President Trump and the Federal Reserve.
7:17
Uh now there's a rumor that uh President Trump threatened further tariffs on China and even raised the possibility of firing Fed Chair Jerome Powell, alarming markets.
7:27
I'm sure you would have to have a very long moment of silence for Jerome if that were happen.
7:32
Um, apparently the the the the rumor uh going around in the media is that uh the administration reportedly explored options to fire Powell after Trump lambasted the Fed, spurring fears about Fed independence.
7:48
And so, yeah, the Fed is supposed to be independent.
7:49
We definitely we don't we don't comment on politics, but it really is funny to, you know, come out uh with this tariff strategy and then be like looking for somebody to fire who's important and like fire the guy who's like, "What did I do?" Yep.
8:02
And so the Fed has the ability to lower interest rates when they're doing, I think, new auctions.
8:09
Uh but they aren't always doing new auctions.
8:11
We talked about this with a buddy of ours uh who said that the Fed actually has not that much control over the long end of the yield curve and they try and do different things to yeah the 10-year is like the really important signal.
8:23
That's the one that people try to finesse it.
8:24
You can try and finesse it.
8:25
It's going to do its thing.
8:26
It's going to kind of do its thing and that's the that's the economic indicator that a lot of people have been tracking.
8:32
In fact, uh some Trump uh appointees had said that lowering the 10-year was a goal for this administration and the opposite happened.
8:42
The 10-year rose uh and that means, you know, potentially higher higher mortgage rates, less affordability, maybe house prices come down because interest rates go up.
8:51
Sometimes that happens, but in general, it can be it can lead to a lot of turmoil economically.
8:54
And it's not where it it wasn't the stated position of the of the of the administration.
8:59
And so now they're maybe looking for someone new to step in to take over and potentially work towards lowering that interest rate.
9:07
But there's only so much you can do.
9:08
It is the S&P 500 and the Dow fell sharply with US stocks down over 10% year to date. Moment of silence. Moment of silence.
9:16
This moment of silence is brought to you by RAM. Go to ramp. com.
9:20
uh tech shares uh led those losses after companies like Nvidia uh export restrictions to China would quote unquote chisel billions off their sales.
9:31
We talked about this last week.
9:33
The NASDAQ today alone is down three and a half%.
9:35
It's it's a big I mean it's not 5% like it was a couple weeks ago, but it's not great.
9:40
Uh they're calling it the sell America trade that picked up uh back picked back up on Monday.
9:46
We've been saying this bull market for short selling bull market for coining but I would prefer people don't create a coinage around selling America.
9:53
No live here and I like it but uh it was the trade over the last few weeks.
9:59
Uh there's been a flight to safety.
10:01
Gold has been on a record run uh which has accelerate accelerated amid the uncertainty.
10:07
Uh people are just looking to avoid volatility.
10:09
Bitcoin has also this is the interesting thing you wanted to talk about um decoupling. Yeah.
10:17
I I mean it's been like a back and forth narrative because Tyler Winklevos posted uh hey for the first time Bitcoin's not trading in lock step with the market.
10:24
Uh so the market sold off and Bitcoin didn't sell off as much.
10:27
Normally they're highly highly correlated which is something that was theorized not to be the case if if Bitcoin really does become this uh this asset of last reserve, this store of value, this go this digital gold.
10:39
Um but now after looking at the market from uh and of course he posted that and immediately Bitcoin moved in lock step with the uh which is kind of just like it it's it feels like if you post something like very uh convictionled like that you're just going to get destroyed by whatever the market does.
10:57
I think uh who's the the Barial sports guy?
10:59
Dave Port was posting about Bitcoin.
11:02
A lot of people you know decoupling.
11:04
A lot of people said, "Oh, could be uh, you know, a bad uh sort of bad signal that Dave Porto is bullish."
11:11
Um, but Bitcoin's down 7% year to date, but today it is up almost 2%.
11:17
Interesting, which is pretty meaningful.
11:18
You wouldn't expect that given how much the NASDAQ has sold off.
11:19
Typically, they move in lock step.
11:21
So, uh, it'll be interesting to keep, uh, uh, to to follow that story.
11:25
We're going to do a whole crypto day on TVPN, bring on a bunch of founders and investors and just general crypto folks and try and get to the bottom of like what is cryptocurrency?
11:34
That's the question I want to answer.
11:36
Like I I know it has something to do with like numbers and money.
11:39
But you've never seen it. I've never seen it.
11:41
So, I want to bring on some experts.
11:42
We're going to bring on some people that run public companies, some people that are billionaires, multi-billionaires, and we'll ask them like what is this whole crypto thing about?
11:50
Uh, and that should be very informative. Yeah.
11:51
For for me, on a more serious note, interested to see where crypto goes from here, right?
11:57
There's been a cool down in the meme coin market. Yep.
12:00
The pump fund chaos seems to have subsided.
12:03
Yeah, we've seen that twice now.
12:05
We saw this with OpenC, you know, was doing, I forget, you know, hundreds of millions of of net income a year.
12:12
I think at its peak, Pump uh Fund was doing the same thing.
12:15
Uh OpenC's volume, you know, fell off a cliff, unfortunately.
12:20
Um, we'll see what happens.
12:22
Uh, they were in my YC batch.
12:26
We should get them on the show.
12:26
I'd love to know kind of like they have a lot of money.
12:29
They have a lot of talent.
12:30
They're they're clearly planning their next act.
12:32
I wonder what that would be.
12:32
I think that, you know, a lot of people have written them off, but that might be wrong.
12:37
Maybe they'll come back and do something else.
12:39
There's there's there's clearly opportunity in crypto if you have a lot of money and a lot of, you know, hardcore developers to build something.
12:46
You can build something entirely new.
12:47
Possibility that they need to completely reinvent themselves. I agree.
12:50
But that can also be a trap.
12:52
Sometimes the second that you know a company reinvents themselves, the the thing that they were originally working on rips again.
12:58
Y uh so who knows what happens with NFTTS. Yeah.
13:01
I mean NFTs specifically were always one of those things where I was I was willing to to buy the narrative of like this is a digital art house.
13:11
This could be the next Sbees, but their valuation was like 10 times 10 times what Sibbees was or something like that very quickly.
13:18
And so the idea of just being like a museum like I think at the height like NFTTS were the volume was higher than the fine art market well hundreds of years of of experience.
13:30
So so two reasons for that.
13:32
One same thing happens in the art world where if you discover an artist and you buy up their works before they y get into ses and these big auctions and um galleries etc.
13:42
You can buy incredible pieces of art for thousands you know singledigit thousands.
13:47
And then the second, the reason for the volume is that it's very hard to have a, you know, $200,000 piece of art change hands six times in a day.
13:57
But that was happening with Crypto Punks.
13:59
That was happening with, you know, Board Apes and um many, many, many, many other projects. So, oh well.
14:06
Well, let's move on to uh the legacy of Pope Francis.
14:08
Uh he passed away, I believe, this morning.
14:10
Uh in fact he was he was so sudden that the Wall Street Journal the print edition uh does not mention his passing away.
14:19
It mentions that yesterday uh thousands in St.
14:24
Peter Square on Sunday were treated to a surprise Pope mobile trip drawing cheers and applause as Pope Francis continues his recovery from double pneumonia.
14:32
He also met with pres vice president JD Vance.
14:34
And this was getting me emotional because um like it's it's pretty I mean I I I I don't know that this is the case, but it feels like like they did this because he knew that he was not maybe not going to make it very much longer.
14:46
And so uh it was like his last opportunity to go and see everyone and and uh it's just it's just brave. I don't know. It's it's beautiful.
14:54
Um anyway, uh the Wall Street Journal has an has an uh an obituary uh of course.
15:01
Um, and I think we should read through a little bit of it and give some background on who this man was.
15:03
So, he was elected the 20 266th pope in 2013 and it marked a series of firsts.
15:10
He was the first Jesuit pope and as an Argentine, the first from outside Europe.
15:15
His legacy is Pope Francis who died on Easter Monday uh at age 88 was disappointing even on the priorities he set for his papacy.
15:22
Pope Francis was known for urging concern for the poor in the best Christian tradition.
15:26
He called for a clergy of shepherds who have the smell of their sheep. Interesting metaphor.
15:32
Uh that is priests and nuns who shared the suffering of their neighbors.
15:36
He made support for the weakest among us the rhetorical centerpiece of his papacy.
15:40
He brought a public informality and openness to the Vatican.
15:45
Alas, Pope Francis believed ideologies that kept that keep the poor in poverty.
15:49
One of those earthly dogmas is radical environmentalism, which isn't about keeping the earth clean for human beings, but keeping the earth itself and treating man as the enemy. Interesting.
16:01
So, the Wall Street Journal is kind of coming for him a little bit on this, saying he shouldn't have been as much of an environmentalist as he was, which is interesting how how much the vibe has shifted.
16:09
And we'll see where this goes the next with the next pope.
16:11
In one of his writings as pope, uh, he cited air conditioning as an example of harmful habits of consumption that will lead to mankind's self-destruction.
16:20
He didn't seem to realize that escaping poverty requires greater energy consumption, which is something we've we've touched on with a lot of the founders on the show.
16:29
Obviously, Augustus Dico is is religious, but then also says that, you know, it is it is our job as as stewards to terraform and make the make make the planet uh safer and healthier for all.
16:40
There's a lot of evidence that if you uh enable humans to live in uh cooler climates, even just indoors, that can, you know, be uh quite positive, increase prosperity and also health. Yeah. Yeah.
16:53
Uh his papacy was marked by anti-Americanism uh and not merely against Donald Trump.
16:58
He seemed to believe that Latin America is poor because the United States is rich.
17:01
That's a recipe for stagnation and despair because the real reasons so many in Latin America languish in poverty are at home.
17:07
lack of rule of law, business, government collusion, protectionism, and other barriers to human flourishing.
17:13
Some attribute his hostility to free markets to his Latin American background.
17:16
Born in Buenosaris, Pope Francis at a young age was made the provincial the provincial superior for the Jesuit order in Argentina during the time of the military huna.
17:24
This was a hard line to walk and some of some of his order accused him of unfairly being too friendly with the regime.
17:31
Argentina for much of his life was dominated by Peronism, a brand of left-wing populism named for Argentine President Juan Peron.
17:39
Uh when he looked around, he saw corruption and the rich doing very well as their fellow countrymen languished in poverty.
17:45
Perhaps it was undeniable that he confused Argentinian Argentina's corporatism with capitalism, which is a common mistake.
17:52
It makes it makes a lot of sense.
17:54
Um, less forgivable was his deal with Beijing as the pope that gave the Communist Party influence in the choice of bishops.
18:00
Conditions for Catholics in China have worsened though the Vatican has renewed the cowtow several times.
18:06
The Vatican has stayed silent on the plight of publisher Jimmy Lie who is China's best known imprisoned Catholic.
18:12
Unlike his two immediate predecessors, John Paul II and and Benedict, Pope Francis was from the progressive wing of his church.
18:18
He punished uh traditionalist bishops who disagreed with his direction and he has populated the cardinal ranks with fellow progressives.
18:25
The irony is that progress progressivism is most popular in places like Europe where the Sunday pews are empty.
18:31
The church is thriving in Africa among younger Orthodox Catholics in the west uh and among younger Orthodox Catholics in the west looking for meaning in life beyond material consumption.
18:41
The cardinal who will choose the pope's successor will determine which future they want for the church and the world's 1. 3 billion Catholics. Fascinating.
18:48
Um yeah, interesting to see the Wall Street Journal kind of wrestle with his legacy in uh in a in a tasteful but uh critical way. That was interesting.
18:57
The one thing that he did is he got the Pope mobile to go electric. Saw the electric.
19:03
That's a G an electric G Wagon. Electric G Wagon.
19:08
So, it's the car that he was in yesterday for uh I had no idea because there is a new there is a new G Wagon that's electric but they they using the G580 base really which makes a lot of sense. You're in crowds.
19:22
Let's not you know emit let's flex on everyone with our G Wagons.
19:24
Is that what you're saying?
19:26
Uh that wasn't what I was saying but uh being like that makes a lot of sense.
19:31
Um but uh anyways and of course there were a lot of memes cuz JD JD uh saw him right before had a visit yesterday.
19:40
Again, it makes sense if the pope is sick, you want to go visit him before he's sick.
19:43
Anyone saying anything else is uh is being a little silly in my opinion. Put on the tinfoil hat. Put on the tinfoil hat.
19:49
Anyway, uh there was an interesting article, an an op-ed by Roland Frier in the Wall Street Journal that I wanted to go through.
19:56
The economics of polarization.
19:58
Uh are you familiar with Roland Frier? No. Fascinating.
20:00
So, he's an economics professor at Harvard.
20:03
uh very outspoken, very uh he's been a little controversial.
20:07
There's like a there's a back and forth. He was kind of attacked.
20:10
Um but he came up with a bunch of behavioral economics kind of foundational research.
20:16
One of the most controversial things was that you should pay kids to do their homework.
20:19
So basically like create an economy for children. Yeah.
20:24
But but he but he studied it and found that uh it was like one of the greatest motivators was just paying kids to to to do work. Makes so much sense.
20:32
Why would you not want to drill into a child's head that if you work hard, you will be rewarded? Yeah, totally.
20:40
Because that's sort of the lesson of life. Yeah. Yeah. No, 100% 100%.
20:44
My uh this started, you know, mowing lawns, picking weeds around. Yeah.
20:49
I was listening to Palmer Lucky on uh on Tetetrogrammaton with Rick Rubin yesterday and he was saying that he he had uh he he built like the largest collection of VR headsets and then he also built a a massive gaming rig with six monitors and top tier uh computer hardware.
21:06
Um and then that's what inspired him to go into VR and start Oculus.
21:11
And I was just and he was like, "Yeah, and I paid for all this like doing like little odd jobs and like side hustles as like a kid."
21:15
And I was like, "Yeah, but I'm running the math."
21:17
And I'm like, if you're spending if you're spending like $40,000, like you're pretty good at those side hustles, Palmer.
21:23
Like like you were working hard.
21:25
Like it's very different. Like white hat hacking. Totally like that.
21:28
I don't even know if it's that.
21:30
I think I just had like like true grind side for like not just mowing lawns, but like mowing a lot of lawns because most kids, you know, if they have like a side job, it's like 200 bucks a month or something.
21:39
It's not it's not enough to get you to like a $20,000 investment in like a gaming PC very quickly.
21:47
Like for me it was mostly like make some money and then go buy like a single video game for 50 bucks or some candy. Some candy. Exactly. This is the usual stuff.
21:55
But seemed like Palmer really got the economic flywheel going early.
21:57
But it makes sense because like he's very entrepreneurial.
22:00
It would make sense that he he'd be making a lot of money.
22:03
Anyway, uh Roland Frier uh just published some research on the economics of polarization that I think is fascinating.
22:09
Um there's this big question about, you know, America seems more polarized than ever. what is driving that?
22:16
And uh the the the the subhead here is people tend to interpret ambiguous information as confirming whatever they believed to begin with.
22:26
And so uh we'll we'll read some of this article and and go through it.
22:28
But it's uh it's interesting to see like where else will this take place and where else can this potentially be exploited or or or mitigated depending on your your goals, I suppose.
22:39
Um, so nothing throws America's divisions into stark relief quite like having Donald Trump in the White House. But Mr.
22:46
Trump is an effective uh but Mr.
22:47
Trump is an effect of polarization as much as a cause.
22:53
We've been growing apart politically for decades.
22:55
And so he gives some data to kind of back up this claim that uh the political division in America is not driven necessarily by Trump.
23:01
Trump is like a product of it.
23:03
Uh so he says and a catalyst in many ways because he's totally extreme. Yes. Right. Yes, totally.
23:10
Um, and so nowadays 85% of Democrats, but only 30% of Republicans think the government should ensure that everyone has healthare.
23:19
So healthcare used to be a little bit more bipartisan.
23:23
That it the the gap has grown 24 points during the last two decades.
23:26
Uh, then there's some other stats here.
23:29
The government has too much power.
23:30
When they poll people on that, 73% of Republicans say yes, the government has too much power.
23:36
only 31% of Democrats say uh the government has too much power.
23:40
And it's a 51 point shift.
23:41
Uh in fact, during the George W.
23:44
Bush administration, Democrats were more likely to say that the government has too much power.
23:48
Um and then there the the split on whether abortion should widen 30 points. Yep.
23:53
The other team has too much power. Yep. Yep. Which I agree with. Yes. Yes. Yeah.
23:58
Uh 100% of people agree that they don't have enough power, I think.
24:04
Um and then uh on global warming the split is 33 points.
24:08
On abortion it's 30 points.
24:10
And so uh lawmakers have also become more polarized over the last 50 years.
24:14
Uh and there's even polarization over why this happened.
24:20
Some data suggests Republican politicians pulled to the right.
24:22
But conservative notes that the but conservatives note that the government has moved sharply to the left.
24:26
And so, uh, 52% of Americans said the Democrats had moved too far left, while 35% said the Republicans had moved too far right.
24:35
And so, uh, everyone is saying, "Oh, the Overton window is shifting. I haven't changed.
24:40
Everyone else has changed."
24:42
It's a very common refrain.
24:42
Uh, uh, even seemingly nonpartisan measures such as the consumer sentiment index reveal uh, polarization.
24:50
Republicans had more positive views than Democrats about their economic se situation during the first Trump term.
24:56
And then this flipped in 2021.
24:58
So they were like, "Oh yeah, like I'm doing pretty well economically. Like I got a job.
25:02
Inflation's not too bad."
25:03
And then in 2021 they're like they're they're like, "Oh, like it's terrible."
25:06
Of course, like COVID's in there.
25:07
There was really inflation. There was stimulus.
25:09
All sorts of stuff happened.
25:10
But um of course people are more likely to report that like, you know, I'm doing better if my guy is in the big in the big house in the White House.
25:18
Uh how can two people observe the same information and come away with starkly different conclusions?
25:21
And why do views on factual questions such as the cause of global warming or the strength of the economy break down so neatly on ideological lines?
25:28
And so Roland Frier tells this anecdote about uh his wife and how this this like inspired him to run this test essentially.
25:39
So he says his wife is a great driver, but she blows the horn way too much for his taste, which is hilarious.
25:45
Um said any slight, perceived or real, and you get a loud honk if she is behind the wheel.
25:50
One morning when we were commuting, a car pul pulled past her on the highway and veered just slightly our way so that its tires drifted into our lane. She honked.
25:58
I tried to reason with her.
26:01
His driving was within the usual margin of error.
26:05
What an economist thing to say.
26:05
It's like this is this is totally within one standard deviation of driving abilities.
26:10
Like you shouldn't you shouldn't be honking.
26:14
Can be a lot in certain circumstances.
26:16
So her response, she says, "I've kept myself from many, many accidents by being a proactive honker."
26:20
And so they observed, he says, "We observe the same incident, but we drew opposite conclusions, and each became more convinced we'd been right all along.
26:29
Is this consistent with rational thought, and could it explain why Americans have become so polarized?"
26:35
As soon as she dropped me off on campus, I ran to my office to tell a fellow economist this anomaly I had observed. Was my wife irrational? Was I?
26:42
or did we need to think about inference and decision-making a bit differently?
26:46
I first confided in Matthew Jackson who specializes in social networks.
26:49
He seemed as perplexed as I was because he knows my wife.
26:53
He offered up several interpretations that would make her seem more rational.
26:56
Finally, he relented and it became one of the guiding examples for us to think differently about how humans process information when there is uncertainty.
27:03
In the simplest version of the model they deployed, imagine that the truth is either A or B.
27:07
Climate change either is or is not caused by human activity.
27:12
There's no gray area in between.
27:14
The death penalty either deters crime or it doesn't.
27:16
No one really knows the truth, but we start with a prior belief about how plausible A or B seem.
27:23
Each person observes a series of signals, information that s that suggests the truth might be A or B.
27:27
Some signals are ambiguous and come off as AB rather than A or B.
27:34
If you were fully rational, able to set aside prior beliefs, you'd store the information in sequence.
27:40
A B AB AB A A A A B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B B like that.
27:45
So, so when there's a confusing signal, you just think it's both.
27:47
But and and if you add that up, that sequence that he describes, it's three points for A, three points for B and three ambiguous signals, like the signal doesn't doesn't lean one way or the other.
27:57
But if you tend to align unclear evidence with your previous expectation, you would come away thinking your original instincts were right because you'd you'd count all the AB signals as A if you're on like team A originally.
28:11
And so now you think the evidence falls on your side by a two to one margin.
28:16
So you would count up six A's.
28:18
You would ignore the three B's in the ABS and you would have three Bs.
28:21
So you'd be say, hey, six to three, I'm seeing a pattern here.
28:25
Further observations of the world entrench this view rather than correcting it because future ambiguous ambiguous signals will have the same skew.
28:32
Our our main mathematical result demonstrates that if a large enough share of experiences are open to interpretation, maybe the guy who drifts into your lane until you honk is an example of the horn saving lives.
28:44
Or maybe you're an average he's an average driver who never posed a threat.
28:47
Then two agents who have differing prior beliefs who see the exact same sequence of events can often end up polarized with one person being absolutely sure of A and the other of B.
28:54
And so they they uh ran an experiment online with 600 subjects uh modeled on a 1979 paper by Charles G. Lord.
29:03
First pres precipitants were presented with questions about their beliefs on climate change and the death penalty.
29:11
They read a series of summaries of research about each topic.
29:16
After each summary, we asked participants if they thought the summary provided evidence for or against the topic on a 16-point scale.
29:22
After all the summaries were presented, we repeated the initial questions about their belief about on the topic.
29:28
There was a very significant correlation between a subject's prior belief and his interpretation of the evidence.
29:33
More than half of our sample exited our experiment with more extreme beliefs than at the start, even though the evidence presented to them was neutral. That is wild.
29:44
The discouraging implication is that in a world where information is plentiful, people will become more divided, not less. That is true.
29:52
Even if they all see the same information, which they don't, because they can choose between Fox News or MSNBC.
29:57
And it's true even if our widening divisions prove deeply unhealthy for our country.
30:01
And that's why you got to turn off Fox News, you got to turn off MSNBC, and you got to only watch TBPN. That's the only option.
30:09
Um, but there's so much to talk about here, and we promise to validate your pre-existing beliefs. Yes, with data. Absolutely. Absolutely.
30:17
Um I thought this was interesting because uh this goes back to the debate around Facebook.
30:23
So when there was increasing political polarization, it was kind of blamed on Facebook for funneling people into extreme echo chambers, right?
30:33
They called them, what were they called?
30:36
Uh I forget I I I forget the name.
30:39
forget I I I forget the name. there was like some buzzword for this for this like you go down like a rabbit hole on YouTube and you start with like and this was pre algo feeds even in the way they are today there of course there were algorithms that would decide what content to serve you but it wasn't it
30:57
was still heavily based on the social graph is the term bubble filter bubble and so you would search for uh like you you you would start a search like uh you know what to do after college how do I get a job and then you would land on, you know, a Jordan Peterson video that was just, you know, about how to live your life, make your bed, right? But
31:15
But then Jordan Peterson also had conservative views.
31:18
And so you would go from the general life advice into his conservative views, but and then he's not really that extreme, but then there'd be someone else who was recommended who was like just political, less life advice, and a little bit more extreme.
31:32
And then a little bit more extreme.
31:33
Make your bed with an American flag. Yeah. Exactly.
31:37
and and and so people would go down these these filter bubbles and then pretty soon everything they would be seeing was there.
31:42
Um Zuck made the argument that you know we haven't seen political polarization in other countries and where Facebook is very active and so he was saying like maybe this is more of a reflection on America than uh than Facebook as a product because the usage data is really high.
31:57
I think his example was like Malaysia or the Indonesia or something and and their society wasn't as polarized as America.
32:05
Um, but I think it's interesting that that um that that that it's merely it's potentially merely the the explosion of information that creates division and and if you if you're just exposed to if you're literally living in a cave, you're less partisan.
32:20
Like it's kind of it's kind of interesting. Yeah.
32:24
So yeah, I mean I I I don't really know where where all this goes.
32:27
Like does it exonerate Facebook and the social media platforms entirely?
32:31
Um, clearly there are some platforms that are like extremely skewed and biased.
32:35
Like that's kind of like by design almost.
32:37
Uh, but uh, it's it's it's interesting to dig into um, at the very least like you as an individual need to understand if you're interpreting those AB signals, those neut those neutral evidence points as adding to your side.
32:53
I think it also applies to the last few weeks with the tariffs, right?
32:58
People are like, "Oh, treasuries are selling off.
32:59
Oh, gold is gold is ripping.
33:02
Bitcoin is ripping and everybody's seeing sort of the same set of information but applying different beliefs to it.
33:07
Somebody might say uh you know gold is ripping uh you know people uh are you know basically shorting the dollar, they're short the dollar. They're short America. So they're buying gold.
33:21
Yet every financial crisis, for the most part, people tend to buy gold because it's seen as um you know, more stable and and predictable and and a true store of value. Yeah.
33:32
Same thing with with treasury selling off, it's like okay, if you're in a trade war and there's this complex geopolitical, you know, um dynamic uh and uh T bill yields are are ripping because foreign governments are selling them.
33:48
Well, there there could be a lot of other reasons other than purely just sell America.
33:52
America's over, the dollar's done, that kind of thing.
33:55
And so, um, yeah, it's, uh, it's hard to Yeah, it's hard to balance all the different signals.
34:03
Um, but if you're looking to get in on the action with some gold, why don't you buy a gold watch on, go to getel.
34:09
com, download the app, uh, pick up a a Rolex Day Date.
34:13
A lot of people think, "Oh, I should buy gold.
34:17
Yeah, I'm going to go get some gold bars. Yeah, why not?
34:19
Gold something you can protect Philippe instead. That's right.
34:23
Or gold uh Yeah, buy a gold watch. Go to bezel.
34:27
Uh your bezel concierge is available and now uh now to source any watch on the planet. Seriously, any watch. Um not financial advice. Not financial advice.
34:37
But uh speaking of things that are somewhat related to fashion, uh robots are having trouble making Nike sneakers.
34:44
This is an interesting story in the context of like re-industrialization.
34:48
Uh Trey Stevens had a has a had a piece on pirate wires all about the importance of robotics and automation in the re-industrialization story that we should go through or talk to him about.
34:59
Um but I thought this was an interesting um interesting discussion of like okay if the tariffs are here to stay and we are trying to boost US manufacturing um what what can we learn from Nike's move to Asia historically?
35:15
And so, um, Trump is betting that the tariff the threat of tariffs and, um, on lowcost countries in Asia will pressure American companies to bring back manufacturing and jobs to the United States.
35:28
U, but high US labor costs mean companies would have to find a way to replace human workers with machines.
35:32
For some industries, that's proved surprisingly difficult.
35:35
Indeed, uh, a years'sl long effort by Nike to shift part of its manufacturing from China, Indonesia, and Vietnam to North America illustrates how tough it is for US brands to wean themselves off of the flexible, low, lowcost contract manufacturers that use armies of laborers to churn out an array of products for consumers.
35:53
products for consumers. Um yeah, I mean one of the interesting things about the the the push in Foxcon where they bring in a million migrant workers is that um you can't really do that with robots like the capex there's no there's no
36:08
fungeible flow of robotics yet that you could say hey yeah you can imagine a world in the future where you bring in a 100,000 on demand robots for you know a a sprint a seasonal sprint but that world also feels pretty far off 10 15 years away, right? It's hard to imagine it happening in
36:28
It's hard to imagine it happening in three years, right?
36:30
Or on the timeline that the tariffs are on, which is the most pressing issue. Yeah. Yeah. Yeah.
36:35
It's it's not something that a company can think like, okay, we're like these tariffs are going to affect our Q3 results, like how do we how do we automate everything?
36:42
So, um this this actually started a decade ago.
36:44
Uh, and it feels like we're still a few decades out from the robotics automation of uh, footwear manufacturing, but back in 2015, Nike poured millions of dollars into an ambitious effort to partly automate what has always been a highly laborintensive industry that's making shoes.
37:01
Uh, at the time, rising labor costs in China and advances in manufacturing techniques such as 3D printing opened the possibility of finding a new way to make shoes that would rely on fewer workers.
37:09
Uh, have you have you ever seen Zeerfeld?
37:11
It's a 3D printed shoe company.
37:14
We should have the CEO on.
37:16
I've met him at a at a party once.
37:17
Um seemed like a good uh uh and I think the company's doing very well.
37:21
Um but they but they've been obviously like small because they're a startup.
37:25
Um but it'll be interesting to see about where where he thinks that will go.
37:29
Um yeah, but this is interesting because even as far back as 2015, Nike wanted was trying to think about bringing production back to North America, localizing it, and uh it hasn't exactly panned out. Yeah. Yeah.
37:44
Uh which puts them in a tough position, right?
37:46
Because they're like, "Hey, we're getting tariffed and we've also been made a good faith effort to do this and it didn't pan out.
37:52
Now here we are kind of stuck between a rock and a hard place." Yeah.
37:57
So, uh the the sho Nike partnered with this company called Flex, an American manufacturer that helped Apple set up a complex factory in Texas to make Mac Pros.
38:06
Um that's the one that there was that picture of Donald Trump cutting the ribbon.
38:10
Um, the goal was to make tens of millions of Nike sneakers at a new high-tech man manufacturing site in Guadalajara, Mexico by 2023.
38:18
The plant would still include thousands of workers, but far fewer than are needed in Asia to make the same number of sneakers.
38:25
If successful, the project could be a model for production in the United States, according to some involved in the effort.
38:33
Uh Nike's competitors also sensed an opportunity to rethink manufacturing built around massive Asian factories where armies of cheap skilled laborers stitch fabrics and glue souls to shoes on hand.
38:44
It feels less modern, more like a Ford Model T production line combined with a Middle Ages cobbler's bench, said Kevin Haley, the executive vice president of innovation at clothing maker Under Armour.
38:55
In 2015, he pledged to use automation to make shoes in Baltimore in a project he called Project Glory. Love it. good name.
39:02
Um, but you know, it's rough.
39:06
Uh, Adidas also got in the action.
39:08
They launched speed factories in Atlanta and Germany with high-tech manufacturing that quickly spit out shoes, heralding a new era in footwear creation.
39:16
Um, they want to move out of China and Vietnam.
39:19
They have the technology to do that differently.
39:21
Uh, said Mike Dennis Flex's then president in 2016.
39:26
Nike's effort was the boldest.
39:29
The company aimed for large-scale automated production in under a decade, which it said would save on labor costs and allow it to deliver new models of shoes to Americans faster.
39:37
Tom Fletcher, who oversaw the project for Flex, came into the effort feeling confident, having just built a highly complex Mac Pro factory for Apple in Austin, Texas.
39:48
You would think that making shoes would be easier than making Mac Pros.
39:50
Uh, and yet they ran into some trouble.
39:52
At the time, Apple had been looking to bring some manufacturing home.
39:55
Flex pushed to rejigger production lines and use automation trying to find as many ways as possible to minimize human interaction.
40:02
That experience came in handy.
40:04
Jord um now it's just interesting.
40:07
Flex has not uh exactly performed despite uh despite the boost u in interest in localized manufacturing.
40:18
Are you looking at the stock?
40:18
Um yeah, they're they're How big is the company?
40:22
very large company already that $30 billion guessing one of uh 30 billion of revenue uh I'm guessing one of the reasons for this is I bet their supply chain uh yep they're they're b they've still got despite you know making efforts to help onshore uh they have
40:40
quite a bit of exposure to uh Asia everyone does these days if you dig deep enough you're going to find some some tariff goods in your supply chain no matter who you Uh, so the machines were supposed to build the upper part of the shoe, knit fabric, add logos, and glue the sole. I
40:56
I I talked to somebody who was doing some uh outsourcing of the upper in China, went over there while the tariffs were there.
41:08
Very chaotic and uh yeah, just like a very a very rough go, big culture clash moment.
41:13
Um, and so these efforts ran into trouble.
41:16
The robots struggled to handle the soft, squishy and stretchy parts that are integral to shoem.
41:20
Shoe fabrics also expand and contract depending on the temperature.
41:24
While in shoe making, no two soles are exactly alike.
41:27
So yeah, I mean if you consider like the benefit of automation on a car production line is that like some of these parts are too heavy for people to lift.
41:36
So you have to use a machine and then it's the same it's the same stamped, you know, metal every single time.
41:41
And uh it's dangerous and it's also just the exact same way. Okay.
41:44
And then there's also welding going on.
41:46
And shoes are much smaller and more fine grained and then obviously um more malleable. Yeah.
41:55
It it everything about making uh electronics is about perfect precision, right?
42:02
It's making sure that you know a a um component has a specific spot within a device and you need to make sure that it goes there that you don't want a human on that. Yeah.
42:11
Yeah, you can imagine that that being harder for a human to do very consistent consistently.
42:17
But if it means like, hey, we need to put um uh you know, this string in in this hole and then that hole and over here uh feels very difficult for a robot to do u you know entirely reliably. Yeah.
42:29
Uh all these things all these things always sound trivial. Yep.
42:34
Hey, we're going to make a robot that makes shoes.
42:36
It sounds so easy honestly. Yeah.
42:38
That doesn't sound it doesn't sound as hard as it is in practice, right?
42:43
That you have Nike doing I think it's I I I would imagine that Crocs are fairly automated because it's kind of just like almost 3D printing. Yeah.
42:51
Pour the the polymer in a mold and you're good.
42:54
Just kind of like, you know, melt it together.
42:56
Um, but for something as complex as a Nike shoe that has a sole, a logo sewn on, there's I mean, even even just lacing a shoe, like that is an incredibly difficult task for a robot, right?
43:08
And they come laced like the you have to poke those through.
43:13
It's hard for me sometimes.
43:14
Shoes don't lace themselves. No.
43:16
Uh, you're trying to do something very precise and then it gets a little colder or warmer and the material and the material changes on you.
43:21
We did not anticipate that.
43:24
As a result, factory production never became as automated as envisioned.
43:29
As shoe production increased, the factory personnel swelled to 5,000, about twice as many as originally planned and costing more than a similar workforce in Vietnam.
43:37
Task after task proved challenging to automate, like the delicate work of gluing soles to the upper part of the shoe.
43:44
If you didn't lay it the right way, there would be a noticeable twist of the shoe, a misalignment that aesthetically means it would fail quality tests.
43:52
A central product was also the huge variety of shoes Nike produces.
43:56
For decades, American consumer companies have given designers nearly unlimited freedom to dream up the coolest products and relied on Asian manufacturers to deliver them.
44:02
And unlike cars or iPhones, shoe models are changing all the time. Yeah.
44:07
I mean, if you think about some of the uh like the how it's made episodes that clearly are heavily automated, it's very much stuff like uh you know, even like food manufacturing, it's like every Cheeto is going to be the same.
44:19
They're just going to go in like slightly differentiz bags and you can just imagine you know all that going down a conveyor belt.
44:25
Little bit trickier um when you're assembling a shoe where you know they're all different sizes.
44:31
Yeah, we talked about things, right?
44:33
Like it's a plastic bag and then something needs to go in it. Yep.
44:37
Fill it up and it moves down.
44:37
It's like far more easy than a shoe which has I bet the average Nike shoe has I mean they sell a lot of themual components, but they sell a lot of them.
44:47
But when you think about the skew complexity, I mean, what are there like 20 different sizes of men's shoes just for a vanilla run if you just want to be able to reach everyone from like a size four to a size 14 or something?
44:59
Like you're you're in a Nike says there's typically 23 parts to a shoe. 23 parts of shoe.
45:05
Each one of those needs to be different for the size quarter quarter over outs tip laces swoosh. Wow.
45:15
It's like a miracle that these even existing. Yeah.
45:19
And then the sizing thing you were talking about, it's like great, you get it working for one size.
45:22
Is it going to work when you need to do the next trip or size? Okay, it's fine. It's no big deal.
45:32
Uh, so automating manufacturing means designing simple products that machines can o undertake over and over.
45:38
Electronics manufacturing uses hard standardized materials, as you mentioned, allowing machines to replicate the same step millions of times.
45:44
You'll have to make sacrifices from how to design the to the complexity of the materials and models you work with, says the former Nike executive who oversaw the project.
45:51
Uh that goes against what the consumer wants.
45:54
They want incre an incredible diversity of product.
45:58
And uh we I mean we see this in cars like the the the Tesla Model 3, Model S, like they look very similar.
46:05
Even the Y and the X look very similar.
46:07
Uh they often come in the same colors. There aren't that many.
46:08
You don't really see that many with like, oh, this one has fender flares.
46:12
This one has wings on it.
46:13
like all that stuff would be completely aftermarket.
46:14
It doesn't come from the factory that way.
46:16
Uh versus, you know, a more a more mature company like Mercedes sells like like a wagon version of the of the E450, right? The E63.
46:26
They also sell convertibles of the C series.
46:29
They sell long wheelbase versions.
46:31
They sell versions, electric versions.
46:33
For example, the uh the Q8 is the same as the Cayenne is the same as the Urus. Yeah. Yeah.
46:42
same power, all Volkswagen.
46:44
So, it's like basically the same car with like a different exterior and a different exhaust and like and consumers want that.
46:50
Consumers want that and and so so I mean Elon's found a way and we'll go into the Tesla story later, but Elon's found a way to convince people in mass at least during the height of the Tesla boom to, you know, say, "Hey, yeah, it's going to look like every other Tesla 3, and you're going to get lost in the parking lot probably, but it has a summon feature.
47:10
It has the best self-driving. It's electric. It's super cheap.
47:14
Found a new way to differentiate. Exactly.
47:15
So, it's like this appliance car.
47:17
It's like it's moving it to the iPhone.
47:18
Nike, if they really wanted to go all in on automation, they should have found a way to make a product that is as ubiquitous as the iPhone with as little standardiz or as much standardization.
47:29
Tesla is just getting hammered down another no 7% but they're suffering from the trade trade war tariffs tariffs but then also the lack of a naturally aspiration also just a broader tech selloff. Yeah. Yeah.
47:44
And uh yeah I mean still no news about a gated manual. Right.
47:46
And so that's got to hurt the stock.
47:48
We're kind of waiting on we're waiting on that.
47:49
Uh that that's the Ferrari is coming back by the way with a with a a manual.
47:53
They're doing a manual V8.
47:57
They did they name it yet? I don't know.
48:00
But uh no, they they all they said was that we are committing to a manual uh IC like no hybrid car.
48:05
They heard the vibe shift.
48:08
Well, they saw what happened with Porsche. Yeah.
48:09
Because Porsche launched the ST which was really great and then the R I believe was was manual as well.
48:15
And so they said like there are certain collectors that just want manual like more analog cars and they're willing to pay for them and it makes sense for Ferrari. They make a ton of Yeah.
48:23
I was trying to think about it.
48:24
I was trying to think what is the desire a highly desirable car that is a hybrid or electric. La Ferrari. LaFerrari.
48:39
I don't think La Ferrari is a hybrid.
48:40
I'm almost positive it is. Is it? Yeah. You're right. You're right. Moged. Moged. Uh 918 Spider. Hybrid. Yeah, the spider too.
48:49
Yeah, the spider too. uh 296 but hybrid I guess I could have uh made it more precise and said in the last few years right um we have in the last few years what uh I mean there's a new 911 that has a hybrid so I think it'll sell pretty well um what else who else uh is the is the latest uh W1 from
49:13
McLaren a yeah I'm sure that is it's got to be hybrid right everything's hybrid now anyway let's get back to Nike Uh, at one point it took the Flex team eight months to figure out how to automate a way to put the Nike swoosh on a shoe only for Nike to move on to a new shoe line for which the method Nike developed no longer worked. And so there's just
49:30
And so there's just this cadence of like new shoes.
49:32
They have to have new new styles.
49:33
Uh, they said it would have been easier to massproduce uncomplicated shoes such as ones with a machinek knit upper part and matched with simple molded bottoms.
49:42
But Nike was unwilling to put limits on its design and expected manufacturers to produce whatever new shoes their teams dreamed up.
49:50
Manufacturing in a lot of ways did not have an equal seat at the table.
49:54
And that's interesting because uh obviously the manufacturing of the iPhone is deeply integrated into the company and they drive the manufacturing optimization that happens at Foxcon.
50:06
This is the whole debate about are they transferring too much IP?
50:08
uh but you know that you know that the the the iPhone team is thinking about what is available on the camera side.
50:14
How does that fit in and then how can they build software on top of that and that vertical integration has really led Apple to have this integrated approach that creates a fantastic product.
50:23
Feels like Nike was very much still in the mindset of like design so that it would be insane to not get a case.
50:30
Oh, what if we made it so the camera like the one of the primary use cases of the device protrudes across the back so it doesn't sit flat. That is brilliant.
50:40
So, by 2017, Flex's investors were bcking at rising costs at the company with some questioning why an outfoot that makes electronics was involved in shoe production.
50:51
Flex and Nike wound up the project.
50:53
I actually think it's smart.
50:55
It's like, hey, who's done really advanced scale consumer product manufacturing?
51:00
Okay, they've done it in electronics.
51:02
Probably a better generally a safe bet. Yep.
51:05
But uh wasn't safe enough.
51:09
And so Flex and Nike wound up the project by early 2019.
51:11
By then Under Armour had stopped mentioning to its investors.
51:15
Project Glory to make shoes in the United States.
51:17
That year, Adidas, which had also faced challenges producing complex shoes with robots, said it would close down production in Atlanta and Germany.
51:24
It shipped its speed factory technology to suppliers in Asia.
51:28
The three shoemakers stuck with their original offshore locations, Vietnam, China, and Indonesia.
51:32
Even after pandemic era factory shutdown showed the risk of having such a concentrated node of production, Adidas, Under Armour, and Nike, they are not talking to the journal about this.
51:45
Uh but uh now China, Vietnam, and Indonesia are in Trump's sights.
51:48
They got they got huge tariffs on them and uh there will be an army.
51:53
Uh Howard Letic uh said that the administration wants labor intensive industries to return to the US.
52:00
He actually said that there will be an army of millions and a millions of millions and millions of people screwing in little little screws to make iPhones.
52:10
So we're going to make them here.
52:10
That's a word for word quote.
52:11
Word for word quote in a recent interview with CBS. All right.
52:16
I threw my how to make iPhones book in the trash. Nope.
52:18
I need to try to get it out.
52:20
I need to go find a dump that they took it to. Yeah. You got to get it back.
52:25
The threat of new tariffs is pushing some to ask whether Nike or others will ultimately have to reconsider efforts to automate manufacturing and bring shoe production back to the US.
52:33
Uh people think it's could still be done.
52:35
Although it won't be easy.
52:37
You need some deep pockets and some patience because it's not going to happen fast. Um but yeah, we'll see.
52:40
You know what aren't patient? Who? Tariffs. They're hitting.
52:47
They're hitting right now. Right now.
52:48
But if you want to track the tariffs, head over to Poly Market.
52:51
I'm sure there's a bunch of great markets on what's happening in the tariff world.
52:54
And if you're looking to cut costs because of tariffs, head over to RAMP. Uh time is money. Save both.
53:00
They got easy to use corporate cards, bill payments, accounting, and a whole lot more all in one place.
53:02
Um let's move over to the next story.
53:05
Um man versus machine as China shows off humanoid robots in half marathon. It's pretty crazy.
53:15
Uh somebody came up to us in the gym this morning and was like, "You guys are talking about this, right?"
53:19
And we were like, "Ah, I actually hadn't heard this, but uh it is in the journal."
53:22
Uh, China, the Beijing half marathon featured a road race between humans and 21 different robot models, and it showed how far robots are from mimicking people.
53:34
Good example of why it's difficult to make shoes.
53:37
It's also difficult for robots to run in shoes. Yeah.
53:41
Um, they're both having trouble.
53:42
what we've been saying uh would like to see a humanoid robot drink 12 beers and play 18 and call it work. Exactly.
53:48
It's not happening this year. Not happening this year.
53:53
So the TN Kong Ultra, a humanoid robot runs across the finish line of a half marathon in Beijing.
53:58
So let's run through this metal asphalt.
54:04
Metal met asphalt in a in a half marathon that featured thousands of human runners and 21 Chinese humanoid robot models.
54:10
Saturday's road race involving human runners and a score of robots in Beijing has been bu as a showcase of China's cutting edge technology.
54:17
I gotta say this is extremely cool for the robotics industry and we should be doing this here.
54:20
Uh I mean you saw this with I posted remember I posted we need a we need a humanoid X games, right?
54:25
games, right? I want to see humanoids doing I want them to see do the 360 do the 1080 any any like athlete that Red Bull sponsors humanoid robot company should be saying hey we're gonna get our humanoid to go cliff jumping to go bungee jumping to uh you know do it you
54:42
know I want to see like you know we we really got to have the 1x the 1x founders on the Neo and a bunch of the other humanoid robotic founders on to talk about you know all our crazy ideas and force them to do it big wave surfing big wave surfing. That seems like the
54:55
That seems like the touring test for humanoid robotics because you can be waterproofed a GS.
55:01
Also, I mean, I imagine these robots are pretty heavy. You got to sink.
55:03
Yeah, that's the real test is a triathlon, right? Can you swim? Can you swim?
55:10
Humanoid robot that swims. That's crazy.
55:12
How's your battery going to hold up in good luck an hour in salt water? Yeah. Yeah.
55:16
Is a Is a honey stinger going to take you across the finish line? I don't think so.
55:20
You're gonna have to be recharging for an hour, robot.
55:21
Uh so it's the 13 mile race, first of its kind.
55:25
The interesting is the size difference.
55:27
Everybody just assumes uh humanoids will be, you know, six foot. Yep.
55:33
They had four foot humanoids. Four foot humanoids. Okay.
55:35
Little guys running around.
55:37
So China has want has said it wants the country to be a world leader in humanoid robots by 2027.
55:41
They are well on their way based on what I've seen.
55:46
Uh, Chinese authorities have lavished support such as subsidies, uh, talent bonuses and tax breaks on robotics companies.
55:52
In reality, the race telling me that for the first time, they know how to do some central planning over there. Yep.
56:00
You know, say what you will about capitalism back.
56:02
If you're looking to subsidize industries, uh, there's another model that seems to be working pretty well.
56:10
Um, in reality, the race showed how qu how both how quickly and smoothly some robots are able to run, but also how far away humanoids are from still being a from being able to mimic human activity.
56:24
Running is a ba very basic ability of human beings, says the chief technology officer of the Beijing humanoid robot innovation center, which developed one of the robots.
56:32
China is making a major push to produce more and more sophisticated robots in part to raise the automation level of its factories.
56:41
Oh, they actually put the uh they actually put the robot in sketchers. That's really cute. It's interesting, too.
56:48
So, these weren't autonomous, right?
56:51
There was people following behind the robots actually basically driving them with a joystick like it's an RC car. Yep.
56:58
Uh but still very incredible feat. Uh, okay.
57:01
So, the first one, the first robot off the mark was the Tien Kung Ultra. 5 foot nine. Hey, let's go.
57:09
Got to put you on the truth zone.
57:10
They're not four feet tall. No, six feet. Yeah. Yeah. Yeah.
57:12
5'9, 115 pounds, humanoid. Ideal build.
57:16
Featuring a pitch black head and sporting an orange tank top.
57:21
Three people accompanied it to help control the robot.
57:22
The race was the culmination of months of training.
57:24
They had to navigate the courses flat and hilly roads and maneuver around six left turns and eight right turns.
57:29
According to the organizers, developers had to train the robots to keep their stability and balance to avoid falling over.
57:34
The organizers initially planned to cut off the race at around three and a half hours, meaning that the minimum average speed was 3. 7 miles per hour.
57:41
Uh developers said the humanoid robots can typically operate for no more than two hours on a single charge of their batteries.
57:49
The faster they run, the shorter the distance they can cover.
57:53
Components and parts could easily break while running, so developers replaced plastic parts with metal and used extra strong but costly materials.
57:58
One company trained their humanoids by connecting the robot to a fitness machine like metal stand to prevent it from falling.
58:06
Uh Tien Kung Ultra was developed by the Beijing humanoid robotic innovation center uh a research institute also called exhumanoid and formed by robotics firm UB electronics and electric vehicle maker Xiaomi and the local Beijing government.
58:20
It could run an average of six miles an hour and could handle hills, stairs, grass and sand.
58:22
Um said a profile on impressive.
58:25
Yeah, this is this makes you think there we have a number of American humanoids.
58:29
Boston Dynamics, uh, 1X, Figure, Tesla Optimus, there's a bunch, right?
58:35
But we actually should have way way more.
58:37
If this is going to be very important technology, we should have 20 plus, you know, venturebacked companies sort of vying for this, right?
58:46
They don't all need to raise hundreds of millions or billions of dollars, but um, it's an area that we should be, you know, probably investing more in.
58:54
And so as they're running the race, uh, Tien Kung Ultra falls down because the battery failed.
59:00
The robots were allowed to swap batteries, and they changed the battery three times on this particular robot.
59:05
Uh, wound up finishing the race in uh, the human male champion completed the race in 1 hour, 2 minutes, and 36 seconds. It's pretty fast.
59:16
Um, followed by thousands of exhilarated human runners.
59:18
Uh, some were exhausted, resting nearby to catch their breath.
59:23
After 2 hours 40 minutes and 42 seconds, Tien Kong Ultra was the first robot to reach the finish line.
59:29
A large crowd of spectators, including government officials, was were eagerly awaiting the robots.
59:34
Uh, in the end, only Tienkung Ultra and Little Rascal N2 were able to meet the original cutoff time.
59:41
The organizers extended it to four hours and 10 minutes so that more robots could finish the race.
59:45
So, the humans still got it.
59:48
But, I mean, you look at the uh you look at the trend line here.
59:50
Uh, I think in a couple years they're going to be crushing this.
59:54
Uh, I I I I see no reason why.
59:57
It feels like a lot of it's probably battery technology.
59:59
Like maybe they keep swapping the battery and that's fine, but if they can really find some sort of battery breakthrough, uh, that's going to enable more power and and like running in a straight line and some slight curves here and there, that doesn't seem like as even as much of a challenge as like assembling a Nike shoe, right? It seems pretty simple.
1:00:16
Um, and so they should be able to win this.
1:00:21
Bell helicopter carrying a bunch of batteries with an extension cord. Oh, extension cord.
1:00:26
That humanoid is sprinting. Unstoppable for sure. 50 miles an hour. Unstoppable for sure.
1:00:33
Anyway, um, very impressive.
1:00:33
Should we move on to some big funding news? Yeah.
1:00:40
Uh, Grun's lands valuation around 500 million as supplement startups boom.
1:00:44
Uh you're familiar with this company. I was less familiar.
1:00:48
I saw I'm not super I'm not super familiar. Yeah.
1:00:50
Follow them a little bit just because I I I'd been aware that they had I had heard about their first year numbers and it was kind of unbelievable.
1:01:02
was kind of unbelievable. like it was a lot of companies have got to the their gross revenue I think in their first year is what best in is the run rate that companies will best-in-class companies can sometimes get to at the end of their first year
1:01:18
but they actually like did it in that in that year wasn't um um so clearly the founder is an absolute dog yeah was it Austin he was talking about just he says uh gruns was the most obvious billion dollar idea hiding in plain sight is a 600 $100 million revenue business. The
1:01:33
The founder of Gruns was like, "What if I took AG1 and put it in a form factor that is easier to consume and doesn't taste like [ __ ] easiest billion dollars anyone will ever make and I don't know that I I don't know enough about the company to tell you probably exactly why it's working so little nominative uh determinism.
1:01:57
Uh the founder's name is Chad." Okay.
1:01:59
So that probably uh has had something to do with the success.
1:02:04
But um yeah, the idea of taking is he the points guy or something?
1:02:08
People are saying he's like a points guy. I don't know.
1:02:11
Uh he says he's founder of Gruns Daily previous Stanford GSB.
1:02:13
He was a summer's partner summer part Summit Partners and he's on the board of Ruggable, Dr.
1:02:19
Squatch, Brooklyn, and Solo Stove and seven more.
1:02:21
So like really really great scaled up consumer products companies.
1:02:27
I've seen the I had a friend who was applying to work at Ruggable and the uh the economics of that business were fantastic.
1:02:32
Are you familiar with Ruggable?
1:02:34
Ruggable uh is one and then Dr.
1:02:37
Squatch is also very under the radar.
1:02:39
I think that's going to be uh a massive outcome.
1:02:41
Yeah, I mean we're seeing this with the Ridge Wallet guys.
1:02:45
Like there's there's increasingly a divide between like like good idea gets market traction, but the real killers come in and optimize the business and just actually get the full value out of the company or or out of the market.
1:02:58
So throwing Chad a follow, but let's read through some of the uh the information article on this.
1:03:03
So investors are on a health kick looking for new ways to tap into the fast growing brands and nutrition sector.
1:03:11
This includes pouring money into selling into startups selling supplements, particularly those that offer new ways for people to load up on vitamins, protein, and other ingredients.
1:03:17
Latest example is a 2-year old gummy supplement startup, Gruns, which raised fresh cash at at an up to $500 million valuation in recent weeks.
1:03:28
According to people briefed on the deal, their monthly sales were roughly 10 million earlier this year.
1:03:33
Uh they recently added kids products as well as a sugar-free version of its adult gummies and a fresh influx of cash nearing nearly 10 million according to securities filings. Wow.
1:03:43
Really low dilution if that's 10 million on 500 post. Not bad.
1:03:47
They're making uh they're making plenty of money themselves.
1:03:49
Um one thing uh the American consumer has always been undefeated.
1:03:54
Undefeated but is especially undefeated in the context of gummy.
1:03:58
Oh, everyone loves a gummy anyway.
1:04:01
Um it there I've seen so many different Yeah.
1:04:04
Uh Dan with create um uh I was initially I I talked to Dan um right when he was starting Create and I was a little bit skeptical around kind of the market that he was going for.
1:04:16
I'd taken creatine forever. Yep.
1:04:18
But um he was the first company that that I'm aware of to put creatine in gummy format. Yep.
1:04:26
And I talked to him like two months later.
1:04:29
it was like so obvious that the the business was was uh just working phenomenally well and he's been knocked off a bunch of times since then.
1:04:38
But um fortunately there's an interesting thing in the gummy space where there's always been issues with quality in the supplement space in gummies.
1:04:46
I I watched a YouTube video about this separately and then Dan posted about it.
1:04:51
How some uh you know mass monster bodybuilding guy tested like 20 of the top uh Amazon creatine gummies and they were a lot of them had zero creatine like Yeah. Like 0. 001% of the label claim. It was wild. Yeah.
1:05:06
Anyway, uh I mean Chad, if you're a real Chad and you're serious about growing this business, get on Numero. Go to Numeral HQ. Sales tax on autopilot.
1:05:11
You know, you're going to have to be paying sales tax on this stuff all over America.
1:05:16
You can spend less than five minutes on per month on sales tax compliance. That's right.
1:05:19
Anyway, I want to keep digging into this a little bit and then we'll move on to the next story.
1:05:23
Uh so they have other hot brands include create wellness.
1:05:25
That's our boy a line of creatine gummies which tapped into the muscle building supplements growing popularity on social media to raise 5 million in series A funding last fall.
1:05:34
The brand has launched in retailers and is on pace for around 25 million in annual revenue when it raised fresh cash.
1:05:40
Let's hear it for create gummies. Dan, good job.
1:05:42
Uh the rise in popularity of weight loss drugs is also providing marketing. Unilver too. Unilver is in the game. Oh. Oh. Was in that deal. Yeah. Oh, that's awesome.
1:05:51
Uh with brands claiming they enhance weight loss or help or help with side effects of the drugs.
1:05:57
Uh Gruns, for instance, has run digital ads.
1:05:59
I had to just look that up. I thought I was leaking.
1:06:01
Yeah, he did announce it last year.
1:06:04
I was like, "Wow, I just broke your your fund raise, Dan. Sorry about that."
1:06:08
No, no, it's in the information already.
1:06:10
Uh, and so, uh, Gruns has run digital ads targeting people taking drugs like Ozmpic who want to keep getting appropriate nutrients as their appetite fades. That makes sense.
1:06:19
Whenever you're on a cut and you're in a caloric deficit, whether it's caused by Ozmpic or just, you know, keeping yourself hungry because you're trying to get diced for summer, uh, you you want to maintain enough protein and enough creatine and enough hypertrophy training and hard training so that you don't lose muscle mass during that cut.
1:06:37
Micronutrients too is doing. Yeah.
1:06:37
And so they sell a monthly supply of gummy vitamins for $79 or 59 for monthly subscribers.
1:06:45
Uh they say the products pack more than 60 vitamins, minerals, and other ingredients into one serving.
1:06:50
And it markets his gummies as cheaper, more convenient version of popular greens powders, which are made of powdered vegetables and fruits mixed together into water or smoothies.
1:06:58
So did you ever hear about this company, Goalie? Yeah. Goalie.
1:07:00
For gut health, the apple cider vinegar gummies. Yeah. Are they really big?
1:07:06
They have they are large.
1:07:08
They do hundreds of millions of annual revenue.
1:07:10
I I think they've been somewhat unsuccessful in terms of like capturing kind of enterprise value around it.
1:07:17
Um that said, just to give you a sense of the scale, they have 364,000 reviews on Amazon.
1:07:22
So they sell 300,000 reviews. Yeah. So wow.
1:07:26
So run the numbers on what percentage of their customers you think have left an Amazon review and then you can get a sense of of their scale, but it's in the hundreds of millions of annual revenue.
1:07:36
Um, so again, uh, Americans are buying billions and billions and billions of dollars of, uh, vitamins.
1:07:44
You know what else has great reviews? Aid sleep.
1:07:46
Uh, I got an 81 last night.
1:07:48
I'm getting up on the routine.
1:07:51
quality and time slept is a little bit lower, but go and get a Pod 4 Ultra, five-year warranty, 30 night rice, 30 night risk-free trial, free returns, free shipping. Go to eightleep. com/tbpn.
1:08:03
Uh, last night, Autopilot made adjustments to boost my deep sleep by 42%. Fantastic. Thank you, Autopilot.
1:08:06
I actually went to bed later than I would have liked because Oh, you watched the Nathan for you.
1:08:14
It's not Nathan for you, but it's it's the rehearsal. Yeah, I love it. Was it good?
1:08:18
Uh, it honestly was giving me I almost was texting the the team because uh he ba they basically recreated the entire an entire airport. Yep.
1:08:26
Uh in in the episode uh just to for one of these our new studio u which hopefully we're signing the lease on today.
1:08:36
Uh we've been in um diligence for a while, but um uh we're going to have to just like recreate LAX if we you know if we want um you know to reach uh Nathan Fielder levels.
1:08:50
He he's been on a fantastic run.
1:08:53
I've enjoyed everything he's put out.
1:08:54
The old Nathan for you episodes are iconic.
1:08:56
Very rare to have business and comedy work so well together.
1:09:00
About uh aircraft safety. Amazing. So can't wait.
1:09:03
I'm I'm enjoying Last of Us, although uh it is terrifying.
1:09:07
It's essentially a horror film and it's not ideal for the deep sleep, I think.
1:09:12
So, I might try and push that to, you know, maybe Saturday afternoon.
1:09:16
It's so funny because because White Lotus ended. Yep.
1:09:18
HBO likes to keep the dread high.
1:09:20
My wife was like, "I'm very I'm I'm happy that we're not going to like be having these crazy cliffhers and like scary scenes right before bed."
1:09:30
And then Nathan Fielder is like sitting there like walking through like the history of like every terrible, you know, aircraft, you know, uh, accident and it's just like extremely stressful. So, um, go check it out. Go check it out.
1:09:45
I mean, uh, in in the world of Hollywood, the name of the game is getting an Emmy or a or an Academy Award.
1:09:52
It's Emmy season coming up.
1:09:52
And, uh, breaking news, we're going to be going for a daytime Emmy.
1:09:57
Uh, and we're going to get a billboard.
1:09:59
We're gonna get a billboard that a four-year consideration billboard.
1:10:02
If you've been if you've ever been driving around Los Angeles around Emmy season, all the shows have, you know, all their accolades up on the billboard.
1:10:09
We're going to go to adquick. com.
1:10:11
Out of home advertising made easy and measurable.
1:10:14
Say goodbye to the headaches of out of home advertising.
1:10:15
Uh anyway, uh let's go to some timeline.
1:10:18
We got our guests joining in 24 minutes. We got a great lineup.
1:10:22
We're going to start showing you the lineup during the show.
1:10:24
We're working on some new graphics packages, but we got David Tish from Fox Group, Mike Vernal from Conviction, Ed from Machina Labs, and then Will is coming on from Morgan Stanley.
1:10:35
Uh we'll be talking about uh what's going on in the venture market, what's going on at the early stage, mid-stage, what's going on in AI investing both at the foundation model layer and the application layer.
1:10:45
Uh I'm excited for the set of conversations, but let's run through some timeline posts.
1:10:50
In the meantime, we got a post from uh actually Lulu first.
1:10:53
You want to do Lulu first?
1:10:55
This is going to be a nice little surprise for you.
1:10:56
Um, Michael and Ben and Scott, if you can pull it up. Let's see.
1:11:04
Is it in the No, it's just in the main chat.
1:11:07
In the meantime, why don't I tell you about Wander?
1:11:10
They just launched Wander Malibu Vista. It looks fantastic.
1:11:13
You can see the ocean from that thing. Find your happy place.
1:11:16
Book a wander with inspiring views, hotel, great amenities, dreamy beds, top tier cleaning, and 247 concier service.
1:11:21
It's a vacation home, but better.
1:11:23
If you stay in that wander, let me know and go outside in the backyard and start yelling, I will be able to hear you. Wait, really? Is that close? Yeah, that's amazing.
1:11:31
You'd have to yell pretty loud, but I think I don't think our listeners will have a problem with that.
1:11:36
Yeah, with all the creatine and working out that's going on in our in our community. I think they'll be fine.
1:11:42
They'll be able to belt it out.
1:11:43
Uh, but man, what a beautiful little uh uh what is that?
1:11:47
Uh driveway, but it's circular. Looks really beautiful. I don't know. I'm excited. And here we go.
1:11:53
Uh, Lulu says, uh, if Kugan does the eyebrow while describing your business model, congratulations.
1:11:58
You're going to have a decorn.
1:12:00
So, she she identified, uh, uh, a specific, uh, I guess.
1:12:04
So, founders, this is something you can read into now, if you are um, if you are if if Kugan is talking about your business, just pay attention.
1:12:16
He's are underrated when you're trying to express.
1:12:18
He's subtly signaling something.
1:12:19
Um and uh in this case it's very positive. It's very fun. Yeah. What a journey.
1:12:25
Five years ago bored during COVID saw a bunch of people.
1:12:29
A lot of people already had podcasts.
1:12:31
A lot of people had Substacks.
1:12:33
I was like I want to do something different.
1:12:34
But I think it's good to get uh just be noisy on the internet because you're not really going to to happy hours anymore during COVID.
1:12:42
There's nothing going on.
1:12:43
Why don't I start yapping on YouTube?
1:12:45
Uh and uh the first year was really really slow.
1:12:49
I think it took almost a full year to get more than like a hundred views on a video, like 50 episodes in a row.
1:12:56
And it was just like it was cool.
1:12:58
I I mean, for me, I didn't understand any of the stats.
1:13:00
I didn't have any benchmarks because I was like kind of the only person in tech.
1:13:04
Eventually, Gary Tan was doing stuff, but he was so big.
1:13:05
I wasn't really like comping to him.
1:13:07
And so, I'd just be like, "Oh man, like I broke a hundred views on this video. Like, it's a banger. Like, this is great."
1:13:12
So many people would have.
1:13:13
And I was just like fired up about that.
1:13:15
But I enjoyed the process and and I was like I did a lot of motion graphics, spent a lot of time in After Effects.
1:13:20
Uh I remember I did a video on breaking down Varta and I built a 3D model of a satellite that was like wildly inaccurate based on like what they actually do.
1:13:29
But they hadn't announced what they were actually going to do yet.
1:13:32
So I was just like it looks like this.
1:13:34
And it was just wildly inaccurate.
1:13:35
But it looked really cool and that's what matters.
1:13:37
It illustrated the idea of of them launching a satellite and then uh bringing it back down.
1:13:42
And uh yeah, the first time I met Delian uh was in Miami at Hereticon.
1:13:44
He was like, "Oh yeah, you made like that video about uh about Va, we actually sent that to our real production team and said like make this but like for us accurate." Yeah.
1:13:56
I was like, "Oh, that's great.
1:13:58
I'm I'm I'm helping out." Um what what a fun.
1:14:01
And then took you five years to realize you just wanted to do a news to newsmax newsmaxing.
1:14:09
All right, we got a post here from Josh Pacini who I have uh did a call with a couple weeks ago.
1:14:14
Uh early TBPN listener quot tweeting Annie.
1:14:17
She says, "Name a better dopamine hit than running all weight."
1:14:22
And he says, "Getting a post read on TBPN."
1:14:24
Well, congratulations, Josh. Today's your day.
1:14:29
You got your post read on TBNN.
1:14:29
Josh also has a very cool um AI startup focused on automating drug development.
1:14:36
So, uh if you want to uh work in that space, go hit him up. You know, it's funny.
1:14:40
Uh, on one of the previous shows, we were talking about, uh, startup names, and we were saying that, you know, there's this new boom in the American manufacturing company of America or advanced manufacturing company of America, uh, Allen Control Systems, the San Francisco Compute Company, the New York Compute Company, the New York browser company, whatever.
1:15:02
And we said, all of those are done.
1:15:04
There's no more alpha there.
1:15:04
We got to go back to the Lys.
1:15:05
and Roy delivered on a Sunday afternoon.
1:15:09
He drops his launch video at 2 p. m. on a Sunday.
1:15:12
What a funny time to launch.
1:15:14
Uh on Easter Sunday, Clue Lee is out. It's the old school ly.
1:15:19
We're getting back in the ly game. I love it. Clue Lee is out.
1:15:21
Cheat on everything and uh has a very like very formally produced uh like really uh refined cinematic launch video with some great motion graphics in there.
1:15:33
We're gonna have Roy join the show tomorrow to break it all down.
1:15:37
Uh, but massive launch, 4.
1:15:39
3 million views, 16,000 likes as of the time of this show.
1:15:42
And Chad Buyers got in the deal, snuck a check in, and he breaks it down. The Cleul Manifesto.
1:15:49
We want to cheat on everything.
1:15:51
And interesting that he's using Wii, I guess he's already deeply involved with this company.
1:15:54
Normally think about as an investor, but this is clearly uh something he's really excited about.
1:15:59
Not sure the round he said was announced yet, but uh No, no, it it is. Yeah. Yeah. Five million.
1:16:04
Um Delian said that that he that they raised.
1:16:07
Maybe Delian leaked it, but then that's not on me.
1:16:10
Uh but Chad says uh we want to cheat on everything.
1:16:14
Yes, you heard that right.
1:16:15
Interviews, exams, sales calls, meetings, there's if there's a faster way to win, we'll take it.
1:16:19
We built Cluey so you never have to think alone again.
1:16:23
It sees your screen, hears your audio, feeds you answers in real time while others guess, you're already right.
1:16:28
And yes, the world will call it cheating, but so was the calculator. So was spellch check. So was Google.
1:16:32
Every time technology makes us smarter, the world panics, then it adapts, then it forgets, and suddenly it's normal.
1:16:40
Activate golden retriever mode.
1:16:40
This is golden retriever mode.
1:16:43
You don't need to know anything.
1:16:44
You just sit there and the clearly answers every question for you.
1:16:47
He says, but this is different.
1:16:47
Uh it isn't just AI just AI isn't just another tool.
1:16:53
It will redefine how our world works. Why memorize facts?
1:16:55
Write code, research anything when the model can do it in seconds. Founder knows.
1:17:05
The best communicator, the best analyst, the best problem solver is now the one who knows how to ask the right question.
1:17:11
The future won't reward effort.
1:17:11
It will reward leverage and being a golden retriever.
1:17:15
So start cheating because when everyone does it, no one is. Interesting. Amazing.
1:17:19
I'm pumped to have Roy on.
1:17:22
And uh very fun little fun fact.
1:17:26
Uh Ben made all of our sound effects with his own voice. That is so that is true.
1:17:31
If uh if you like this if you like the voice, shout out to Ben, our producer.
1:17:35
Could have been a voice actor.
1:17:37
Speaking of the other Ben, let's go to Ben Stiller. That's right.
1:17:38
Uh Sinners uh he's posting about Variety.
1:17:41
Sinners has a mass putting Variety in the Truth Zone.
1:17:45
Okay, we like to see that. A founder.
1:17:47
He's going founder mode because he's uh you know he's he's a producer.
1:17:50
He's the founder of many movies.
1:17:50
Um, Sinners has amassed $61 million in its global debut.
1:17:56
It's a great result for an original R-rated horror film.
1:17:58
Yet, the Warner Brothers release has got a $90 million price tag before global marketing expenses.
1:18:06
So, profitability remains a way a ways away.
1:18:08
And he says, "In what universe does a $60 million opening for an original studio movie warrant this headline?" It's a good question. I don't know.
1:18:17
It's not that crazy because I mean when when when big studios put hundreds of millions of dollars before be behind a uh behind a film like they're typically um hoping that they recoup all of that like on day one, right?
1:18:33
So, I mean, it doesn't seem like Variety is being like crazily negative here.
1:18:38
And they did kind of couch that in like, you know, they got a ways to go to get profitable, which is like maybe factually accurate if the price tag really is 90 million.
1:18:47
Um, but it seems seems cool.
1:18:49
I have I haven't seen the movie, but I don't know if I'll have a chance to go see an R-rated horror film.
1:18:54
That seems hard to rally the troops around.
1:18:56
But, um, have you watched the trailer or seen anything about sinners?
1:19:00
I don't watch uh horror films. No. Too scary.
1:19:02
might throw off your sleep score.
1:19:04
And we're in a knockout knockout.
1:19:06
I can honestly say I've never watched a full one.
1:19:08
I just don't I don't get uh What about the classics?
1:19:13
What about Scream or Cabin in the Woods?
1:19:16
I'm also not a movie person.
1:19:18
I've seen like three movies. Three movies.
1:19:22
Wolf of Wall Street, Wall Street, Wall Street 2.
1:19:25
Those are the three movies.
1:19:27
The only the only ones that matter.
1:19:29
Scarface, Good Fellas, Godfather.
1:19:29
No, you haven't even seen those, have you? I've I Yeah.
1:19:33
Not Not Have you seen Wall Street, too? No.
1:19:35
Have you seen Wolf of Wall Street? Yes. Okay. And you've seen more.
1:19:39
Have you not seen Wall Street? No, I haven't. Oh my god. American Psycho?
1:19:41
I have seen that, but I read the book first. Okay. Yeah. Yeah. That's weird. The book is weird.
1:19:47
The book is The book is weird.
1:19:49
Anyway, uh should we go to Jim Simons on the only rule he focused on in trading? Goat.
1:19:55
He says uh we'll we'll pull up the video, but he says something like uh never doubt the computer. It's pretty great.
1:20:02
Can we pull this video up?
1:20:05
Was probably never trusted trading decision that was made before ripping a heater.
1:20:10
Oh yeah, he's a big smoker.
1:20:10
Uh yeah, Jim Simons, in case you're not familiar, he's the uh worth $31 billion, 55th richest person in the world.
1:20:19
Uh he's the founder of Renaissance Technologies, a quant fund.
1:20:23
Famously uh there was a market crash.
1:20:26
Everyone asked them what did the what was going on at Renaissance?
1:20:30
what was your team doing while the market was crashing?
1:20:32
And they said, we were at a movie. Amazing. They're just Yeah.
1:20:36
They're just like, what are we going to do?
1:20:37
The machine will handle it.
1:20:38
Anyway, we're ready to play the clip.
1:20:40
The only rule is we never override the computer.
1:20:45
No one ever comes in any day and says, "The computer wants to do this. That's crazy. We shouldn't do it."
1:20:51
It you don't do it because you can't you can't simulate that.
1:20:53
You can't study the past and wonder whether the boss was going to come in and change change his mind about something.
1:21:00
So, you just stick with it.
1:21:02
And it's and it's uh and it's worked. I love it.
1:21:06
I mean, this is the story of like this is Lisa Doll move 37.
1:21:14
If if if the Alph Go team had said, "Hey, we think this is an error.
1:21:17
We're going to override what the computer's doing."
1:21:20
they would have thought that they were more likely to beat lease at all because it was a novel move that had not been considered by humans before and in fact move 37 was very important in that game.
1:21:30
It's interesting too to think of it in the context of if you uh interject and insert human decision-m Yeah.
1:21:38
you're actually taking away from the program's potential success in the future which could more than make up for a one-time mistake, right?
1:21:45
Like there could be if um you know the Renaissance team uh if the computer does something that loses money effectively um but then the next time it does that thing it makes it back 10x, right?
1:21:58
Like was it actually an error?
1:22:01
Gambler mindset in the computer.
1:22:03
You're only one computer aided trade away from I mean it literally works. He made $33 billion.
1:22:09
99% of computers right quit right before they hit 100x.
1:22:12
Just imagine it has like a side like internal thought and you're just like expand the reasoning on this one and it's just like I'm feeling I'm so due. I got a hot hand. I'm so do. Yeah, it's great.
1:22:26
Um I mean I wonder where else that philosophy has a potential impact like uh we saw this with the debate over social networking where there was there was a there was a question about you know the the algorithmic social feeds feed people confirmation bias.
1:22:43
They feed them content that they enjoy.
1:22:46
They feed them brain rot sometimes.
1:22:48
Maybe the steward of the network, Zuckerberg, who has essentially complete control over Meta, should step in and say, "You know what?
1:22:58
I'm going to dial back the brain rot or I'm going to dial up the, you know, turn down the news, turn up the entertainment, turn down the, you know, the turn up the educational content."
1:23:08
There's a bunch of things that in theory could could, you know, feel right, but maybe the end state is just step back and maximize for LTV. Who knows? I don't know.
1:23:19
I have a hot take on this which is like the the LTV is just like if you actually think about the true LTV of a of a user of Facebook, it can actually drive a much stronger and more valuable outcome as opposed to optimizing for ARPO, which is a one-year metric. Yeah.
1:23:38
So, if you're optimizing for a one-year metric, you're going to try and sell someone, you know, like gambling, mobile games, anything that just pulls money out of their pocket and potentially gets them distracted from creating value in the in the society.
1:23:52
But if instead you focus on what is the true lifetime value of this potential Facebook user for five decades, for their entire life, well then it might actually be valuable to instead of trying to get some kid to play Candy Crush and spend all their parents' money on Candy Crush tokens.
1:24:13
Instead, send them a bunch of educational content.
1:24:14
Get them really pilled on capitalism.
1:24:16
They go build something really valuable.
1:24:18
something really valuable. they make a bunch of money and then you can run advertisements on them for Rolls-Royce and Lamborghini million dollars on on meta ads maybe maybe and so and so the the there's an interesting theory where where like what
1:24:33
what are we saying when we when we when we say the kids are suffering from brain rot what we're really saying is like they will not contribute to society they will not become economically valuable but maybe they go and become slot farmers maybe and if they pull out of the slopes. It's possible. Of It's possible.
1:24:49
Of course, it's possible.
1:24:51
But I do think that there's something interesting where if you think about it from a ultra long time horizon, it can actually be economically efficient.
1:24:59
Our quarterly, you know, uh system will allow for that type of That's the beauty of Zuck.
1:25:05
He doesn't need to think in quarters, right? He doesn't need to.
1:25:08
And so I'm I'm I'm optimistic that that that someone like Zuck could take a really really long view.
1:25:14
Also, I mean, I think he has kids.
1:25:16
I think he has a very good strong moral framework.
1:25:17
So I think he might just kind of like you know optimize for that naturally.
1:25:20
But I do think that there's an economic model to optimizing for long long lifetime value lifetime value education and like human flourishing economic flourishing even in the face of declining arpoo in the short term.
1:25:36
I would love to think that Zuck is thinking of themselves as, you know, of Meta as operating digital railro railroads and it's like, you know what, we're gonna, you know, uh, we're just going to invest in, uh, delivering educational content to our children for the next 5 to 10 years.
1:25:51
We're going to get them really excited about, you know, learning and science and because we will make more money in the future. That's the key thing.
1:26:00
It can't just be purely altruistic.
1:26:01
It has to be, no, this is the right move for the shareholders.
1:26:06
But it's gutsy and it would never fly if you're optimizing the next quarter's return.
1:26:09
Anyway, I had a post on Saturday relevant to Facebook.
1:26:12
I said, "Someday you're going to look down and realize that the good old days are over. X will be like Facebook.
1:26:17
Your group chats won't be popping and your phone will just feel different.
1:26:21
Don't let this time slip away.
1:26:22
You need to increase your screen time now before it's too late."
1:26:24
And um funny enough, that thought came to me on a Saturday when my screen time is uh the lowest.
1:26:32
And uh but it it did um we I I've had this sense over time that that Twitter and X are just sort of Lindy um and we'll kind of always be here.
1:26:41
But um it's very possible that X goes the way of Facebook at some point and then we're just we're just the boomers on Facebook just all, you know, yelling at each other and arguing.
1:26:53
We'll become like a Tik Tok reaction stream. Yeah. Or whatever's next.
1:26:57
You always got to move on. We're going to Weimi. We're going to Weimi. It's on the come up. We talked about Weimi. It's the future. Yeah.
1:27:03
The FTC says that we is Facebook's, you know, biggest competitor. Biggest competitor. Or one of the top two. Snap and then Weimi. Snap and Weimi.
1:27:11
So, we'll see you on Weimi, folks.
1:27:14
Yeah, we really got to chase that down.
1:27:16
We got to We got to get that C.
1:27:17
I'm going to let you do this next one. Okay.
1:27:18
So, uh uh War Doll says, "You need to be dad maxing.
1:27:23
You need to be waking up at 5:45 on a Saturday to hit the gym and then coming back to wake up your wife and kids by blasting Creed while making breakfast.
1:27:32
You need to be dethatching your lawn.
1:27:34
I don't even know what that means.
1:27:35
You need to clean your car and then stare at it in the driveway for half an hour.
1:27:38
You need to be telling your kids that Jerusalem belongs to those who worship Christ.
1:27:42
You need to be treating Easter Sunday like Super Bowl Sunday.
1:27:45
You need to be absolutely mogging childless heathens with peak dad power moves as God intended. Very funny Easter post.
1:27:52
Obviously, there was a lot of like he is risen posts that went viral over the Easter Sunday, but I thought this one was was very funny.
1:27:58
Uh it is something that happens naturally.
1:28:02
You start waking up earlier and earlier in 5:45.
1:28:03
Uh that's 5 minutes later than I wake up actually if I look at my alarm.
1:28:08
5:40 uh on the Ashton Hall grind.
1:28:12
Anyway, Sam Demo, friend of the show.
1:28:15
He's been on uh he runs Impulse Stoves.
1:28:19
uh he had a good take about uh aliens that I wanted to share.
1:28:22
He says, "I remain convinced that the best evidence that the US government has not recovered crashed alien spacecraft is that one, that would be the coolest thing ever, and two, someone would have to share it in a group chat with the boys."
1:28:35
And he is, of course, sharing a screenshot of uh Secretary of Defense Pete Hegsth is said to have shared attack details in a second signal chat.
1:28:44
The defense secretary sent sensitive information about strikes in Yemen to an encrypted chat that included his wife and brother com people familiar with the matter set.
1:28:54
So send that headline to your wife and just say us us.
1:28:59
I was trying to do that earlier but uh kept getting paywalled by the New York Times.
1:29:04
The failing New York Times wouldn't let me screenshot the headline. Yeah. of an article.
1:29:08
Yeah, just take the screenshot from this. It's great.
1:29:10
We we got a nice clean screenshot here from Sam. Uh yeah, very funny.
1:29:13
I I actually think this is a great take.
1:29:16
think this is a great take. Uh, but we should have Jesse Michaels on the show to give us the latest and greatest in the world of UAPs, UFOs, and and uh alien aircraft because and spacecraft because he has gone way deeper and and regardless of what you think about aliens, I think that Jesse's done a
1:29:32
great job of finding interesting stories, interesting historical anecdotes, interesting anomalies, and uh even if you remain unconvinced after spending time with him, uh you will be entertained and you will feel like you are talking to an intellectual, not uh just a crazy conspiracy theorist. And so
1:29:47
And so um huge shout out to Jesse Michaels, one of the best to ever do it.
1:29:52
Uh we grew up on YouTube around the same time.
1:29:54
Uh anyway, let's move on uh to Atlas creatine cycle.
1:29:58
He says they should have a 420 for creatine monohydrate. So true. I love it.
1:30:04
Willow Brian says 106 because 10 grams a day and six looks like a G. Oh, okay. 10g. I get it. So true, King.
1:30:12
People are into creatine right now.
1:30:14
A lot of debate over whether creatine I have 15 15 grams not your max those are big scoops. Yeah. Yeah. Yeah. Big scoops. Uh yeah.
1:30:23
It's a little sandy but uh you got to get it down.
1:30:26
The creatine it seems like you know maybe it's a fast bargain but it's the best hair loss.
1:30:32
It's the best hair loss supplement in the world. Yeah.
1:30:33
A lot of people talk about hair loss supplements thinking that they're supplements that inc you know encourage hair growth.
1:30:38
But you know Joe Rogan's bald. Yeah. He looks strong. Maybe it's the creatine. Who knows? Maybe it's genetics.
1:30:43
Maybe it's a million other things, but yeah, you never know. You never know. It's worth it. Worth the risk.
1:30:48
Anyway, uh we we got four minutes.
1:30:51
Let's do a couple more posts.
1:30:53
Um this one by David Haber over at Andre was interesting.
1:30:56
Um he says, "In any fast growing organization, one of the best feelings as manager or CEO is having someone who works for you that I call safe hands.
1:31:07
It's this implicit sense that you can delegate a task, project, or line of business and trust that the person is going to execute with quality.
1:31:16
We've talked about this before with like the levels to employ to different employee skill sets, the pyramid that Sean Puri outlined.
1:31:21
Uh safe hands isn't just about competence.
1:31:26
It's about proactively communicating when issues arrive, acting as an owner when problem solving, and operating independently without the need for micromanagement.
1:31:35
Importantly, safe hands isn't tied to seniority.
1:31:37
It's just remarkable how inconsistent it is across an organization.
1:31:42
When you whenever I find someone who was safe hands at my prior company, especially junior folks, I pour responsibility on them.
1:31:48
Many of these individuals stretched far beyond their original roles and became leaders at the company with several now successful entrepreneurs.
1:31:56
It's one of the most common conversations I have with founders I work with, especially in the series B+ stage as they're beginning to bring on executive leadership for the first time.
1:32:06
Do you know, do you feel like you have safe hands across every functional unit of your business?
1:32:10
How much leverage do you feel you're getting from your team in those areas?
1:32:13
Every CEO instinct instinctively knows the answer to these questions and finding your safe hands is a key step towards scaling yourself and the company.
1:32:22
And Matt Grim says,"I call this uh folks you can turn your back on and they're insanely valuable.
1:32:30
In fairness, I didn't come up with that phrase.
1:32:31
One of those types who reports to me told me and I'm shamelessly stealing it." Got it. Jordy, what's your take?
1:32:36
No, I think oftent times uh you know, in this case, David and and Matt are founders.
1:32:41
Oftentimes, founders end up getting credit for the the things that these sort of safe hands roles end up doing. Oh, totally.
1:32:48
Um and so Matt is is But that's the goal.
1:32:50
You have to hire these people and then also train them and then take advantage of them and then you know back them when they go out to do their own thing. Absolutely.
1:32:58
Well uh before we have David uh we got to talk about Eric Eric Torber.
1:33:02
Speaking of Andre uh Eric Torrenberg has joined Andre Horitz as a general partner. Uh fantastic news.
1:33:10
He'll be on the show tomorrow to break it down for us.
1:33:12
Uh very excited to talk to him.
1:33:14
Uh been a good friend for years. Talked about media.
1:33:16
He's been in the podcasting space for for like multiple years now.
1:33:21
Bel Tarpentine uh done tons of interviews and uh one of the greatest networkers in human history.
1:33:26
I think probably he knows everyone and uh he will child of the internet.
1:33:31
He will throw you in a group chat with the world's most important people randomly and I'll be like why am I in this? But thank you. This is interesting. Yeah. Yeah. Great guy.
1:33:41
Anyways, very excited for him and uh and the firm and uh very excited to talk uh with him tomorrow.
1:33:48
Anyway, uh let's bring in our first guest, David Tish from Box Group. How you doing, David? Hi, guys.
1:33:53
How are you on having a new job today? I do. Yes. Thank you.
1:34:00
Uh yeah, I I'm I'm taking it from part-time, 50 hours a week to full-time, 100 hours a week, but uh it's only up from here. Are you hiring? We are. We are hiring.
1:34:09
always how do I how do I pursue this?
1:34:12
Yeah, the the VC to full-time news anchor pipeline, you know, once you've done it all in venture, you know, like you have, the natural step is when you when you start in this industry, I feel like that's the ambition, a daily show that you can pontificate across the whole industry.
1:34:29
So, well, we're hoping we're hoping to recruit you to be, you know, once a week, drop in, we will be your podcast.
1:34:37
Yes, this is this is actually like jokes aside, like the way that you like work for us is that uh when news breaks, you bring it to us.
1:34:44
When uh when when something happens that you have an opinion on, you come on the show and you talk to us.
1:34:50
Uh and you are elevating the the venture capitalist way to we don't have news.
1:34:55
We don't have things that happen.
1:34:57
I feel like you are really uh playing the audience.
1:35:00
What about like the most complex and naughty geopolitical conflicts?
1:35:03
I feel like you guys as a class are like the best in business.
1:35:07
Over qualified to talk about it. Oh, okay.
1:35:09
So, you you want to stay away from politics. All right.
1:35:11
So, let's start with let's start with tariffs. Yeah.
1:35:13
Let's start with tariffs. Iran, China, Russia.
1:35:16
Let's just go down the list. Good day in the market. Everything's great here. Everything's great. Everything's great.
1:35:22
Have you heard about the accordion effect to early stage?
1:35:25
Because that's our famous early stage line of like whatever happens in the latest stage market, it's like an accordion all the way to early stage.
1:35:34
So reverberates over the course of you know 3 to 5 years.
1:35:36
So we will not ever be impacted because the accordion basically never reaches early stage because something else will happen in the macro before then.
1:35:45
Uh which is a really nice uh don't ask us anymore.
1:35:50
There's no there's no liquidity so you can't panic sell even if you wanted to you.
1:35:54
15 years that's a that's a that's a third accordion in Ripple. Yeah liquidity.
1:35:59
I mean I mean that like there's no better example of that than uh we have SVB crisis. Uh interest rates go up.
1:36:06
It's the death of venture.
1:36:06
We're hearing about oh GPS are going to be giving back LP capital but then well well would you hear about this AI trend and all of a sudden it's time to rip checks. Saved us all. Saved us all. Who would have thought? Yeah. Wait. Yeah.
1:36:22
What uh wasn't there this was sort of before my time but there was a whole chatbot era right?
1:36:26
Wasn't there like pre AAI was like it was supposed to be the thing? It was sort of right.
1:36:33
It was just didn't have the the underlying tech trend. Is that right?
1:36:37
The I feel like the tech wasn't there, but also the branding of chatbot to uh chat GPT isn't that different.
1:36:46
Um but magically different from a outcome and uh sort of adoption curve.
1:36:53
If you look like the automated chat bots, I think is what they were called for that moment in time.
1:37:00
Uh they all fell on their face because they weren't interesting enough, good enough.
1:37:04
They were like pre-programmed answers and then uh open AI releases in essence just an advanced version of that in many ways from a consumer positioning uh and here we are.
1:37:17
And so you are like the idea of chatbot was right in how you're interacting with the tech.
1:37:22
the tech uh to your point just wasn't good enough. Yeah.
1:37:24
90% of chatbot founders quit right before they strike it back right before they saw the open AI.
1:37:31
I mean I I I don't know if I've actually we we should find some chatbot founder who stuck it out and hugging face. Hugging face is the one. Yeah, that's right.
1:37:40
Hugging hugging face is the winner.
1:37:42
They were in a boot camp for chat bots. No way.
1:37:44
And pivoted from a chatbot into hugging face. So that's your answer. That's amazing.
1:37:49
Yeah, we gota have it's it's Clement on the show. Yes. Right.
1:37:52
Yeah, we got to have him on. That'd be great.
1:37:54
Um, so yeah, I mean uh to get a little bit more serious like uh like how are you processing this year?
1:38:01
What are you actually excited about?
1:38:03
What is fatiguing you in venture these days?
1:38:06
We have our annual meeting tomorrow.
1:38:08
So I'm in the middle of practicing our our script on the market so I can just try it. Please audience. I love Yeah.
1:38:16
love Yeah. the uh you know the the pre-liberation day post liberation day market speech is uh is is different in some way but you really like I don't know you're you're investing at the earliest stage of a journey and trying
1:38:31
to attach two people starting a company with big long-term 10 15 year ambition to some moment in time macro or what today feels like an incredibly volatile macro where you actually can't even understand what stability looks like or a new framework. Like it just it's
1:38:51
Like it just it's impossible.
1:38:53
And so the daytoday job here is we meet people, we get really excited about what they're working on, we give them money and we wait 5 10 15 years to see how that all plays out.
1:39:05
So the the investment moment of investing in an early stage company feels totally decoupled from the macro of what's going on.
1:39:15
The macro does impact the companies you invested in years ago as they come to market for more money, as they figure out their, you know, revenue retention.
1:39:24
But the sort of day-to-day volatility, uh, I don't think has an impact on our day-to-day job of funding great people, starting big ambitious visions.
1:39:33
Do you think that there are going to be stories from this uh tariff cycle that are like what happened in COVID where everyone was com like Airbnb was particularly hurting as a startup.
1:39:47
I mean I they were public at the time but they were I still think of them as like a YC company, right?
1:39:52
Uh and it found it seemed like that the story of Airbnb was like COVID was a transformational moment for them and they emerged like a stronger company than ever.
1:40:01
Um do you expect that to happen to any startups uh maybe in your portfolio or just out there in the market that are getting beat up but might emerge stronger than ever?
1:40:09
So so like if you isolate tariffs that has a sure impact on a subset of companies very outside of software in many ways.
1:40:20
Um, I think there's software companies that involve logistics and shipping that definitely uh are impacted, but I don't know that the like come out of the tariff uh is a predictable framework like what does that actually mean?
1:40:35
Do we get to free trade in reality?
1:40:37
Do we get to some equilibrium?
1:40:39
Is this China only that we're talking about?
1:40:42
And I think within there lies a very hard to foresee again point of stability.
1:40:48
Whereas at least in CO if you said we return to a normal world that was a more predictable end point.
1:40:57
Um if you assumed you know early in co it was wait two weeks and then the world will start again and then it was some longer period.
1:41:03
But I think there was an appreciation for what coming out of it could look like.
1:41:08
Whereas I think right now you're in like the creation period. Harder to to guess.
1:41:12
Um yeah, you know the the easy guess is like American manufacturing windfall, but that feels like a big stretch like no one's building a factory overnight.
1:41:24
I feel like that happens in other parts of the world, not here.
1:41:28
world, not here. It does feel like it could be a catalyst for the American dynamism companies, the re-industrialized companies that have been placing that bet and kind of hoping for just a little bit of a boost that
1:41:42
you could see a lot of like, oh, that series A company went to series B like a little bit earlier because of this narrative and then that was enough to get them over an awkward hump, build some real infrastructure out and kind of like realize the vision. I think I've
1:41:53
I think I've been like cautiously optimistic from this.
1:41:57
If you're a super early stage investor, which which Box Group is, yeah, and then you're trying to time markets by being like, "Oh, there's tariffs.
1:42:05
Let me make a bunch of supply chain related or manufacturing investments, but then the company's not going to mature for five to potentially 10 years.
1:42:13
So, it could be a totally different environment."
1:42:14
And and one example that thought was relevant, we had the CEO of Astronis on last week, which I imagine you invested in like over at least over five years ago. 13 years ago. 13 years ago. Overnight success. Overnight success. Um. No.
1:42:27
And and that's like a good example of like you couldn't have you you thought that space was going to be important at the time and you like we knew actually we're we're early stage VCs.
1:42:37
We knew we didn't think you knew crystalall obviously. Um, no.
1:42:43
I mean, like if you think about AI, we're we're the earliest investor in a company called Clay.
1:42:47
And Clay created a vision, had a product, and it was in need of AI for it to work.
1:42:53
And and it wasn't created yet.
1:42:56
And so, if you watch this seven-year overnight success, it clicked when the underlying technology caught up with what the goto market product was.
1:43:07
And I think you see that a lot of the times that the founders who are early in a space and can see it through and wait for the world to come to them while they're sort of pushing the world to get there.
1:43:17
That's where a lot of that magic is.
1:43:19
And I think Arranis or Zipline again 14 years ago we funded Zipline is a is another example.
1:43:28
Um we are huge Zipline shills now that we talked to the founder.
1:43:30
I mean what an incredible story.
1:43:32
We left Yeah, we left that conversation.
1:43:34
We were basically like would invest buy at any price. Yeah.
1:43:39
Because I mean he he it's not just that he's been grinding it out for so long, but I feel like if you watch the launch video like he paints the it's the opposite of the Black Mirror vision.
1:43:47
It it's a very positive vision of the future and it's something that yeah, Job's not finished.
1:43:52
He's going to be doing that for 20 30 years easily.
1:43:53
Like there's so much to do and and now it's just a matter of manufacturing scale, economics, all the basic stuff.
1:44:01
But like he's got the drive.
1:44:02
He went to his office on a recent trip to San Francisco and I probably hadn't been in his office in in a decade and we walked in and it was just like this jaw-dropping Americanbuilt uh hardware software combo that you like you walked out of there and said this is the beginning of a another 15-year vision. Totally. Totally.
1:44:25
Uh, was there was there, you know, looking back at at these companies like Zipline, Astronis, was there ever a trend over the last whatever 15 years that you said that you were tempted to go all in on because I imagine there was a lot of investors of that era that said, "Oh, space is so cool. I'm going all in on it. Climate tech is so cool.
1:44:48
I'm going to do this, you know, instant delivery." Just NFTs. Just NFTs. Yeah, that was the one. one. Yeah.
1:44:57
No, I mean it we're generalists.
1:45:00
We we are founder driven.
1:45:00
We find people uh who have their own vision.
1:45:03
I think us as a as a firm, we should never be in charge of coming up with ideas.
1:45:10
We're not that's not our job.
1:45:11
Our job is to give people who have dreams uh capital and and support to help them.
1:45:16
Um, John, your company I think is the single uh most heavily debated internal decision uh we'd ever made at Box Group uh from a moral standpoint.
1:45:28
We had to like decide if it was uh good or bad.
1:45:30
Uh and so that was interesting.
1:45:34
So we almost pivoted into drugs.
1:45:38
Um that was an opportunity for us.
1:45:38
But uh no I like the world pivoted to AI, right?
1:45:46
So I think 2 years ago we were saying like 30 to 40% of the companies we see talk about AI or are oriented around AI.
1:45:53
Last year it was like 90 and this year it's 170%.
1:45:56
Um you you don't meet nonAI companies right now and if they're not AI first it's we're attacking an industry by bringing AI to it um which is slower and stale.
1:46:08
And I think that is a forced all-in versus a sort of opted all-in.
1:46:16
Does does do the different business models that seem to be emerging does that change the way you're underwriting these investments?
1:46:25
I'm thinking about highly capex intensive businesses where you're just expecting a ton of dilution or we've seen some companies that are kind of just doing like private equity style rollups but they're raising from venture capital.
1:46:36
And I imagine if you're not careful with PRAA or how you're sizing your bets like it it could it could turn out to be like ah we really didn't get the full bite at the apple.
1:46:44
But is that something that even matters to you or is it just like back the founder and we'll figure it out?
1:46:49
Uh private equity rollups in venture always ends well. Definitely. Always. Yeah. Yeah.
1:46:55
Definitely. Always. Yeah. Yeah. I mean we we we like to say private equities is not a mature business and so there's lots of opportunity for tech people to come in and disrupt no they don't take it serious famously not famously not cutthroat famously not cutthroat they're
1:47:10
happy to leave all that money on the table yeah exactly just dollars on the floor for VCs to come pick up right Jordy um no no I I I was more um how was has there ever you know it feels like over the last two years call it there's been more of these sort of rollups than ever before. Was there was there
1:47:27
Was there was there another did was there another kind of like general like marketwide crack or was it just sort of like individual areas um and opportunities?
1:47:39
I struggle to think that the way to build companies is to replicate another company that's succeeding.
1:47:46
In these moments, what you tend to have is one winner and a bunch of followers that don't win.
1:47:51
And so, can there be a single version of a roll-up?
1:47:56
If you look at Andreal, they're great at acquisitions.
1:47:58
That's a different phrasing of the word rollup.
1:48:00
Rollup is either we're going to buy a bunch of things that are the same size or we're going to buy like one big thing and then tuck in some other smaller things around it.
1:48:11
So, I think unpacking the nuance in these words is more important than assuming a general like success across a a wide variety of companies.
1:48:20
I think the like moment in time where you probably saw fast followers or fast movers win was in the ondemand business, right?
1:48:31
When Uber trained the world that you could touch something on your phone and something in the real world would happen, it unlocked like a a behavior across the world that was very different, right?
1:48:43
And that was this this phone tore world connection that I think opened up a ton of other businesses.
1:48:48
I don't think they all worked, but I do think that that was a horizontal opening versus a business model innovation uh that has been used in other industries or financial industries and then uh applied in venture. Yeah.
1:49:04
applied in venture. Yeah. the idea the the classic sort of like accounting firm roll up when I look at that it's like when it's been two weeks and my like CPA is not like on it or like is not you know like clearly like you know is busy
1:49:19
with other clients I'm like it's not like I want to switch immediately but I'm pretty quickly and so the idea that you're just going to buy like 20 accounting firms and then slap AI on it and not have like 90% churn it's just um interesting But have you tried AI? I mean, it's
1:49:35
I mean, it's pretty good. Yeah. Yeah.
1:49:37
It's pretty, you know, it's only going to get it wrong 5% of the time and that's going to cost you millions, but uh this is a topic that Jordy and I have been debated and it's kind of related to Uber opening the mindset of like this delivery boom.
1:49:54
Uh chat GPT rappers, it's been derided as a term.
1:49:58
Um we've now seen the VS Code fork wars.
1:50:02
Uh, but it looks like there's a chance that OpenAI buys Windsurf for three billion.
1:50:08
And I've been going back and forth with Jordy about this, like is this game on for M&A in the rapper market?
1:50:15
Because if OpenAI, who has, you know, great team, lots of money, can't like build it and they want to buy it, then do do we see Anthropic, do we see XAI buying stuff?
1:50:24
Do we see Amazon, all all of the Mag 7 buy stuff?
1:50:27
Then maybe even, you know, you get into some of the just fortune 500 companies that want to buy startups.
1:50:33
And so I like I'm in this weird scenario where I've I've been somewhat accepting of the idea that a a rapper might not be a power law outcome, but everyone involved in Windsurf, if this deal goes through, will do very well.
1:50:46
And so what is your take on on that market and and does this narrative shift?
1:50:50
shift? My take was broadly that like this is highly strategic for open AAI but that doesn't mean every rapper even if they have a lot of traction revenue is like suddenly strategic to a wide range of buyers right sure I think I
1:51:04
think in my in my 15 years doing this the thing I've underest or I overrated was the amount of M&A that would happen I think if you go back in the early growth of like this era if you look at you know anything from um you know
1:51:18
Facebook and and before that Google, Yahoo, Twitter, um on their way up they bought tens if not hundred plus companies and a lot of that was for stock and it was stock that appreciated postacquisition incredibly valuable for
1:51:36
both the company uh and the founders who sold and I think if you look at the next wave of startups a dramatic decrease in M&A because instead of buy you built and the aqua hire didn't prove out to the best use of equity. A lot of the the
1:51:49
A lot of the the decision of buy versus build became easier uh as you thought about where to innovate.
1:51:57
I think in sort of getting at your question, um I hope we're entering a period of M&A.
1:52:02
I think the easy way to rationalize it is the valuations of some of these companies that you mentioned are so big that they can afford to buy companies for substantial amounts that might actually align with the founder of the the acquired um expectations.
1:52:18
And I think that's where one of the other mismatch was happening is that the the company that wanted to buy another company was just not able to hit the price and so they didn't happen.
1:52:30
And so I think one way to rationalize it is whether it's the the big seven or uh the scaled AI companies have big enough valuations to go out and buy things for numbers that will hit a founders's expectation which I think would unlock um you know a lot of a lot of positive things in this ecosystem.
1:52:51
things in this ecosystem. one is liquidity, but two is like there there isn't really an answer to how do things end right now for a ton of companies and M&A is the piece that's been missing uh to me the most it gets uh less discussed and I think Fortune 500 companies need to modernize right and we saw that
1:53:10
pressure 10 years ago when you heard like all the non- tech companies are going to become tech companies then the public markets punish you on a quarter by quarter basis and it's like forget that we're not going to do any of that but I think today AI might be such a dramatic forcing function that smart big companies need to move quickly uh to get ahead of the curve. So is that is that buy versus you
1:53:31
So is that is that buy versus you you mentioned you know these super inquisitive companies maybe 15 years ago and then that changed.
1:53:39
Do you think that was a factor?
1:53:41
I mean you guys basically Plaid was built out of your guys's office.
1:53:46
There's a lot of uh there's been so much infrastructure built during the Plaid era.
1:53:51
Do you think that that was part of the calculus for some of these bigger, you know, companies that would have been acquirers, but they said, "Hey, like we could buy this tech and then maybe try to integrate it or we could get this team and and do it or hey, this infrastructure exists and we can just build it ourselves."
1:54:06
So many factors go like the biggest one to me is that the government didn't want big tech to buy more companies.
1:54:14
And so if they were going to buy companies, they had to be like the biggest ones, not the like they were going to fight the regulatory fights on big companies, not on small ones.
1:54:24
And that dried a lot of it up.
1:54:24
And then I think the valuation expectations was the other friction.
1:54:28
When a startup decided they were worth a half a billion dollars and the potential acquirers wanted to buy them for $und00 million, that was no longer interesting.
1:54:37
And so it just it decoupled uh rationality from playing out in the M&A world, right?
1:54:44
You have to have like a buyer and seller meet on price in order for the deal to even get going.
1:54:50
Um and I think that's where uh the biggest gap was. Plaid sold, right?
1:54:55
Plaid sold and the government stopped it.
1:54:57
And I think that that was uh you know to me a interesting moment in the government getting involved in something that was debatable.
1:55:06
And I think it also tells big companies like don't waste your time trying to get these things through and that slows down a whole wave of potential M&A. Interesting.
1:55:17
Um, I don't know if if the M&A market relates to what you're seeing on the LP side, but we've heard that there might be some fatigue from LPs on what's going on in venture, but at the same time, it feels like that David Gogggins meme where, you know, venture just keeps on chugging no matter what.
1:55:36
Uh, you're going to the AGM.
1:55:36
Uh, what what are you seeing broadly in terms of LP appetite for ever bigger venture funds?
1:55:45
We're a small adorable seed fund based in New York City.
1:55:48
So I we are not we are not responsible to answer that question.
1:55:52
I think our like your job if you take LP Capital is to give them back a lot of money one day.
1:55:58
And I think if you haven't um you you owe that to them and so at some point the patience should run out and I I appreciate that.
1:56:07
I think the challenge is the time for liquidity in early stage venture has gotten pushed significantly from where it was more predictable a decade or two decades ago.
1:56:22
You could say 7 to 10 years on an early stage fund and mean it. today.
1:56:24
stage fund and mean it. today. I think you say 7 to 10 years and you're like by that I mean like 12 to 15 y and that that's a substantial difference and I think you need to align with your investors on what that timeline is because it's in a rational timeline like the people we fund if you go back 15 years they were like pre-ele school for the most part right like if you look at the young founder that you're investing
1:56:50
in today a 15-year timeline is twothird of a life like it's an irrational number and so these like the go in motion of making an investment to the return the fund to LPS timeline is so decoupled from I think a psychological
1:57:08
standpoint that um your stakeholders are just dramatically uh unrelatable on that on that topic of like young founders um how has the early stage market evolved there's a it feels like there's a lot new a lot more products like YC see is bigger than ever. But then
1:57:24
But then there's also Zfellows, there's different fellowships.
1:57:28
There's people that are just giving away money to people to like go try a startup and then maybe I'll invest later.
1:57:34
Uh and obviously the teal fellowship is kind of like a scaled version of that, but there's a lot of other, you know, initiatives.
1:57:39
Uh how has that changed your strategy?
1:57:42
Has it changed your strategy at all?
1:57:43
Uh and kind of like what what does the early stage market look for you look like for you today?
1:57:49
I think there like there's always been these splashes of noise.
1:57:54
Um, and very very little like if you go back again, I'm old and so I've been doing this through these micro waves of uh change in the early stage market.
1:58:08
Um, very few products stay where they are.
1:58:13
They move in different directions. They either get later.
1:58:15
either get later. So if you think about new entrance as as funds into the early stage market most of the ambition is to become a bigger fund and in doing so you become later by nature of scaling uh your business and I think on a product
1:58:30
side uh you're only as good as the product you're in and so what has been amazing to me about YZ is they've just continued to compound quality of the satisfaction to their customer they are they are not free right you're giving YC equity. And yet, if you look at like the
1:58:46
And yet, if you look at like the happiness factor, the the MPS or whatever uh cheesy acronym you you call reviews, but like people go to YC because other people who went to IC loved it and they think it was entirely worth it.
1:58:59
And so I think the products that get invented need to live up to the cost.
1:59:03
And if the cost is free, like nothing is really free.
1:59:07
And so what comes with that freedom is that attaching to a brand.
1:59:11
And if you attach your company to a brand, is that a good thing to have done or is there some negative uh externality that comes with that?
1:59:21
And so I think um and it's probably not the end point for the the products that are offering free capital to be free forever.
1:59:27
Uh I would assume that's a hard thing to scale.
1:59:30
Um, and so I think you have to just like as a company aligning yourself with a brand.
1:59:38
The trusted brands have proven to be trusted for a reason and the the new ones need to earn that.
1:59:42
And I think um there are some that have been around long enough where you find good examples of uh quality coming from them and those are the ones that I would uh tend to trust more.
1:59:53
Uh you've been in the game for a long time.
1:59:55
Can you tell me the story of your first investment ever? Yeah.
2:00:01
Um our box group is named after our first investment.
2:00:04
Uh so uh it's like a combination.
2:00:08
So the first investment a company called Boxy.
2:00:10
Uh it was a Roku competitor.
2:00:13
Um uh we were in their seed round as a adorable check.
2:00:17
Um I didn't want to write my name as an angel investor.
2:00:22
So I created an LLC called Box Group to invest in Boxy which made me feel bigger.
2:00:27
Uh, the box was like a cool nightclub in the city at the time.
2:00:30
It felt like felt like a cool word.
2:00:32
Uh, I've looked with Aaron Levy. Uh, he registered box.
2:00:36
net like two months before I did Box Group. Oh, wow. Which I'm bitter about.
2:00:41
So, at some point he'll sue me and all over.
2:00:44
But he's been on the shows, too, so he can come on and debate for it. Yeah.
2:00:48
Well, it's collab we're all collaborative in this world anyway. No com no competitors.
2:00:52
Um, so, uh, you know, Boxy gets funded.
2:00:56
uh follow on by Fred Wilson.
2:00:58
I'm like, "Oh my god, I'm amazing at this." Yes. Yes.
2:01:02
I gotta read that guy's blog. This is so cool.
2:01:04
Uh and then General Catalyst came in and then Aer, the founder, went in front of Congress to like fight the cable companies.
2:01:11
Uh so it was like an amazing uh first investment to to like see the narrative of a startup. They sold to Samsung.
2:01:18
Um and they're they were built into like the Samsung smart TV technology.
2:01:22
Um, but the idea was right and the timing was challenging.
2:01:28
They had to fight every single battle on behalf of the people that came after them.
2:01:33
And I think that that was a a really good lesson.
2:01:37
Uh, is is do you see you have a ton of portfolio companies um too too many probably too many I'm sure to to manage.
2:01:45
Do you see uh and and Case text is one of them uh which is an interesting example because that was a company that no one nineyear overnight success. Yeah. Exactly. Exactly.
2:01:57
I'll hit the uh overnight success overnight success.
2:01:59
Um but uh uh are you seeing that across?
2:02:06
Can we get a soundboard for our like meet companies meetings like where we should Yeah, we should. Um handshake deal. Yeah. handshake deals. Raise your forecasts.
2:02:16
Yeah, there's an internal bingo that we run that every time they say buzzwords, we we're preempting.
2:02:21
But how do how do you think of AI in the context of the portfolio broadly?
2:02:27
context of the portfolio broadly? Are there are there a bunch of examples where it's kind of creating new momentum in the business or like tons of new opportunity or and you know I look at this like you know I have 50 odd uh personal investments just all of them will work all of them will work I'm sure
2:02:45
they've all been marked up so so I'm sure that's what the employer website told me which I was like yeah that could be 98% successful is it 98% hit rate um but but uh you know how often are you seeing it sort of create momentum versus momentum just being something that once you lose it, it's really hard to kind of get it back if you're not getting, you know, sort of lucky. I mean, Case Text
2:03:06
I mean, Case Text is like the most amazing story.
2:03:09
What a grinder into uh just being early in the right way.
2:03:15
And I think to Jake's credit, he was always like the smartest person in the room and figured out how AI could come into what they were building in a differentiated way and like split a team off, built a product, got to market first um in an industry that was ready for it.
2:03:33
And so I think it's this and he landed like he landed an outcome that he changed his life and changed the team's life and I think like a unique story.
2:03:42
um we're now like three years past that you're early to the game.
2:03:50
And so if you haven't figured out how to take your stale product or your product that's maxed out and begin to uh re-energize it, it feels like you're a little late.
2:04:01
Um there are those stories and I think Clay's the other side of it where it it AI happened and Clay uh benefited from sort of the the speed of it and now is running.
2:04:12
Um, but I think within the portfolio there are companies that benefit from pieces of it.
2:04:17
Um, and whether that's on the cost-saving side or the growth side.
2:04:22
Um, I think you can find examples of of everything, but it's within the verticals that I find uh the most interesting momentum just being recreated, right?
2:04:33
It's it's companies that are servicing customers and in some way if they can be the deliverer of AI to an industry that can't get it themselves.
2:04:42
That's a really big opportunity and I think Caseex sort of represents that. Yeah, that makes sense. It's awesome. Thank you so much.
2:04:48
This was a I came on here to promote something and let's go give us like the the Howard Stern of of tech and so when you go on Howard you have to promote or like the late night shows you have to promote.
2:05:00
So, we have we have a conference that we're putting on with Union Square and First Round and Lux in New York called Foundersenyc and it's foundersenyc. co.
2:05:09
It's free, which is different than most conferences.
2:05:11
Uh, it's for people who want to start a company.
2:05:15
We got the founders of Data Dog and [ __ ] talking to each other on stage.
2:05:21
to me a collection of all the the big great early companies that were built and scaled in New York coming on stage to just like sort of create the community effect of of building here.
2:05:30
Um and so if you're Yeah.
2:05:33
Do you want to take a do you want to take a victory lap on NYC just generally? I feel like No, no, no.
2:05:39
I'm I'm just saying when when you started Box Group, a lot of people like maybe maybe you No one had heard of we're like the the you know pioneers of New York City.
2:05:47
No one had heard of New York City before. There was that error.
2:05:51
That's true about tech, right?
2:05:51
The amount of the amount of the capital concentration in New York City right now is incredible. Never been more.
2:05:58
We're like the pioneers that discovered Miami in the co era. No, no.
2:06:02
Look, I I think New York's a home for people that want to win.
2:06:09
And I think whatever you do in New York, it's it's like an incredibly harsh environment.
2:06:15
And in order to stand out and win in New York, you have to fight friction. No one cares, right?
2:06:19
And that's what's actually fascinating to me about tech in New York is like if you went on a subway or to the restaurant nearby, nobody cares what you're doing.
2:06:30
Like no one cares about your business, no one cares about tech.
2:06:31
And I think that friction creates people to have to fight a bigger fight here.
2:06:36
Um because there's not momentum.
2:06:38
Like there's not somebody pulling you up and saying like we're all in this together.
2:06:42
And so I uh I think the community here uh is is just like people put their head down and grind and uh it's produced some real success and we're excited to put it all on stage. That's amazing.
2:06:54
Well, thank you guys very much for having me. I'll see you tomorrow. Uh we'll see you back. We'll see you back here. Talk to you soon. Bye. Cheers. Uh fantastic. Humble giant. Yeah.
2:07:06
Overnight success himself. Yeah.
2:07:09
Uh well, next up we got Mike Vernal coming in from Conviction.
2:07:11
very excited to talk to him.
2:07:13
We'll bring him in from the waiting room now. Mike, are you there? Can you hear me? How are you doing? I'm good, thanks. How are you two? Uh we're fantastic. Great to have you.
2:07:24
Uh it's Thanks for having me. It's a wonderful Monday.
2:07:26
I hope your Easter was well if you celebrate.
2:07:28
I hope your weekend was great and I hope you're off to a great start of the week. Thanks. It was great. Yeah. Fantastic.
2:07:34
Uh yeah, I mean I'd love to kick kick it off with just um a little bit of a little bit of background on your career and then um what you're excited about today and then we can kind of go from there. Does that sound good? Yeah, that sounds great. Um so let's see.
2:07:51
Uh my career I've been I've been an investor for the past decade.
2:07:56
Um kind of an accidental investor. I never aspired to this.
2:07:58
Um but I I sort of um I spent 15 years operating.
2:08:04
I was at Facebook for many years and um decided to give it a try and it's been great.
2:08:09
Uh so I've been an investor for the past decade.
2:08:14
Um I am like uh I'm very motivated just by like individual founders and folks that I find compelling and so like a pretty broad set of companies.
2:08:23
Um so like a mix of consumer and like SAS and hardware and infrastructure and some others.
2:08:29
Um, some of the larger companies are companies like Rippling and Notion and Verata and uh Clay uh and a couple others.
2:08:37
Yeah, we just heard about Clay from David Tish.
2:08:38
What what was the what uh Yeah, we we we just heard about Clay.
2:08:42
David Tish from box was on and he was saying about how Clay was this you know overnight success that took seven years but um as they usually do.
2:08:48
Uh I'd love to know the story of the first uh first investment that really made you catch the bug.
2:08:55
Obviously, you read Facebook for a long time, but then uh at at some point you were like, I got a knack for this and I'm going to take it more seriously.
2:09:04
Um, you know, I made a handful of angel investments before I like formally became an investor. Sure.
2:09:09
Um, but they were and I think in retrospect they're actually like they're they're pretty reasonable.
2:09:16
Um, like I only made a handful.
2:09:19
Um, I I probably made like seven or eight.
2:09:23
One of them was Notion, which is probably the best of the bunch, but like Wealthfront, it was like notion and Wealthfront and this company called um Human Interest, which is like a 401k provider, and a couple others. I know the CEO.
2:09:34
Sayan, I I I I'm like randomly friends with Jeff Schnebly who runs the company now.
2:09:39
It's a fascinating story because it was a YC company and then he came in as as like a founder mode CEO, but he had like like his background.
2:09:47
I mean, I I love him, but it's like not super He's He didn't start the company.
2:09:54
He has a MBA and I believe he has a PhD and he's almost like a manager mode type of guy, but he's been fantastically successful with the company and I'm just the biggest fan. Anyway, sorry. Exactly. Exactly.
2:10:04
Um, but I I think uh none of them really motivated me to do it.
2:10:07
It was more um I uh so I was lucky enough to join Facebook when it was relatively early like a few hundred people and I was there for eight or nine years and I uh I led a mix of sort of product and engineering at the company.
2:10:21
Um, and I was my wife and I had our first kid and it was the first time that I I took like six weeks off for paternity leave and it was a little bit of a uh it was like the first chance to catch my breath in eight or nine years and I was trying to figure out like what I wanted to do next.
2:10:39
Uh, and the obvious thing would have been to stay at Facebook and Facebook is was and I think still is just one of the most amazing companies out there.
2:10:47
Um, but I'd gotten to know a guy named Brian Shrier uh at Sequoa.
2:10:52
And Brian was Brian had actually backed two of my close friends at the series A, a guy named Matt McInness, who's now the COO at Rippling. That's right.
2:11:00
Uh, and Steve Gity uh at a company called Hearsay.
2:11:03
And Brian uh had kind of said, if at some point you decide to pop your head up and do something else, you should come come talk to us.
2:11:12
And so one thing led to another uh and I ended up at Sequoa.
2:11:13
And it was a little bit of an experiment to just see if uh investing was uh uh would be fun and interesting.
2:11:21
And as it turns out, it is both fun and interesting.
2:11:22
So, how did you um you know, people talk about quote unquote operate, you know, uh investors love to say if they've spent, you know, two years at a company, they love to say they've had operating experience.
2:11:34
You had some very real operating experience at Facebook. How did you feel? Um uh how did you feel?
2:11:43
How how durable were the learnings from the true operating experience versus finding sort of common truths about the way to do business and the way to build products and teams.
2:11:54
Um that you know because cuz I've found like for example like the influencer marketing that I maybe did in 2016 like no longer works, right?
2:12:05
And but yet there were things that I learned in that era that you know I still use today, right?
2:12:09
So how do how do you kind of delineate like you know sort of like a higher level of abstraction? Yeah.
2:12:14
Higher level sort of these sort of lindy ways of doing things that are durable and valuable to communicate to entrepreneurs that you back.
2:12:21
Yeah, it's a good question.
2:12:21
Um I will say one thing that I think I learned that still applies and one thing I had no clue about um uh which I didn't learn until like the investing side of things.
2:12:33
Uh, I think one thing one thing that's underrated, maybe it's not underrated, uh, about Mark at Facebook is he's incredibly patient.
2:12:40
Um, I mean, I guess you kind of see it.
2:12:43
I think he's like spent $80 billion so so far on like reality labs.
2:12:47
Um, but I think, you know, he's he is at the same time both like patient and impatient.
2:12:52
patient and impatient. Uh I mean everyone talks about like the move fast break things and I think there is like an like a very strong bias towards action and just like getting getting stuff done but he also like from a first
2:13:03
principles perspective kind of if something should work like if something logically makes sense he will just keep playing it out to try to to um until there's like new data that suggests that it that it is wrong which I think explains a lot of the reality labs investment. Um, and I think like I think
2:13:19
Um, and I think like I think one thing that's underappreciated about like a lot of great companies, maybe it's appreciated, I don't know, but like they take a long time.
2:13:30
Like uh I mean some companies are really fast and off to the races, but like you know Figma was like four or five years before it started to work.
2:13:39
Um, Notion was like four or five years before it started to work.
2:13:41
Clay was four or five years before it started to work.
2:13:44
I think Air Table like was founded in 2010 and didn't really sort of break out until maybe 2017 2018.
2:13:48
I mean all the sort of the PLG companies you know everyone's everyone says they like want to be a PLG company which to me is like yes I want to be a high growth company that is a I don't want to have to do sales.
2:14:00
I want to grow quickly without doing sales too. Yeah. Exactly. Exactly.
2:14:04
It's like I also want to be born rich and handsome but you know what are you going to do?
2:14:08
uh the it's it it takes I mean these companies just take a lot of bake time and I think I think if you've operated for a while and you've seen how um like if you viscerally understand patience and like if you're if you're doing random stuff or if you're like literally there's no hope then you should probably throw in the towel.
2:14:29
But in a lot of cases, if you're doing the right thing and it just hasn't hit yet, you just kind of uh sort of you got to play out another uh play out some more cards.
2:14:39
Uh and I do think I've observed a lot of financial investors having a like if it's not up into the right three months in it's like oh man we made a mistake and like what the hell are you doing?
2:14:51
Or I was talking to a founder recently uh who has like a financial investor on the board now and they were like your competitor is you know at 5 million in AR why aren't you at 5 million in AR yet?
2:15:03
Which is like I get it on one level but on another level like you just got to um I think uh I I think you got to make sure you're doing the right thing and sort of doing it in the right way.
2:15:15
Um so I think that was one lesson that definitely carried over from operating which I I think you kind of don't have until um I think is correlated with people who have done it and and sort of understand the the messiness of building something.
2:15:32
I will say conversely um at both I mean I was at Microsoft before Facebook and in both cases you're pretty insulated from like the market you're operating in.
2:15:43
Like I think uh once you get to a certain scale the you're you're kind of just thinking about how to build new things for your existing customers as opposed to how to really like start in brand new markets.
2:15:57
Um, and so I had no real intuition for how to like evaluate a business and whether it was a good business and a bad business and if it was how to think about um the like I there were obviously sales and marketing teams at both Microsoft and Facebook, but you don't really build an intuition for how to like attack a new business category, I think, unless you've worked at a startup.
2:16:19
So that that was all new to me. Uh, I have a question.
2:16:21
Uh, I've I've heard founders bring up the example of Figma as a company that took a long time to get a product out into the market and get, you know, traction.
2:16:32
It was what it was like four or five years until the first million of ARR.
2:16:36
Um, and I've heard founders use that as a reference point and be like, we're doing the Figma thing.
2:16:44
Like, it's going to take a while.
2:16:46
And I'm like, no, you're not.
2:16:46
And I'm like, no, you're not. they were doing something that was like fundamentally you know very technically difficult leveraging a new technology when there was no not of they were operating in a space that maybe wasn't hot for different reasons like you know sort of cloud design collaboration you know you you had I'm sure there was other players but they had sort of four years to be multiplayer wasn't a trend
2:17:08
multiplayer wasn't a thing they sort of created that so what what would you say to founders that that tell you you know oh yeah we're doing this but then you look at the market and you say, well, you have four or five other companies
2:17:19
that are have sort of more advanced feature sets, uh, maybe and are high, you know, it's highly competitive and um, I feel like that having the opportunity to do the Figma thing is like a luxury of being like very early. You got to earn that. You got to be You got to earn that.
2:17:32
You got to be Dylan to earn that. Yeah.
2:17:33
I mean, I I don't think Yeah.
2:17:36
If anyone came in and said it was going to take like four or five years to build Figma, I don't one, I don't think anyone would build Figma.
2:17:43
Two, I don't think anyone would fund Figma.
2:17:44
Figma. Um I I think there's an element of um like you you have to I think you have to be moving quickly in the short term like anyone you know sometimes founders will come in and say I mean what is it it's April and they will be like oh yeah we'll be ready to launch at the end of the year and unless
2:18:07
you're like taping out silicon or something that that scares the crap out of me because I don't know there's a lot of time between now and the end of the Um, and so I I think you want someone that is just like iterating at an insane clock speed and like learning at every iteration. And I think those iterations
2:18:23
And I think those iterations have to be pretty fast, but like sometimes you just don't know when it's going to start to catch.
2:18:30
And so I I know the Figma story less well than than some of these others, but it's, you know, you were just talking to David and you Notion notion's a good example, too, right?
2:18:41
didn't they were toiling away, you know, doing some, you know, in in relative obscurity until they found the thing that really was magical. Yeah.
2:18:48
One of my favorite set of videos, uh, you can find it online, uh, is there's like a notion product demo from, I think, late 2013.
2:18:57
Um, and it's this it's like this low code, no code environment.
2:19:02
There's all these blocks.
2:19:02
There's like a Stripe block so you can like drop payments in.
2:19:05
like it's a pretty uh it's a it's a pretty wild demo uh because you see like the through line from today to what it was back then.
2:19:13
Um and it's clear like all the ideas are basically the same.
2:19:17
It's like this canvas and it's got blocks and you can drop all this stuff in.
2:19:20
Um, but of course like the first iteration was like wildly complex and no one um no one really used it other than like a handful of like uh dieh hard folks that that that um that figured it out.
2:19:35
And so of course like the rethinking of it was okay make this an like an insanely great note-taking app and then go from there note-taking to Ricky and the lake.
2:19:42
And notion is an interesting example because I think uh there's like incredible stubbornness on the vision and I think like what Ivan has been trying to build has like not changed in 12 or 13 years.
2:19:53
But there was like a lot of flexibility on tactics and it was like okay this thing didn't work. Let's try it again. Let's try it again. Until something caught.
2:20:00
And I think that's kind of the like the the the art of it is like have a really clear vision of where you're like what you're trying to build over a five to 10 year window and then but just move insanely quickly and be willing to like constantly change tactics.
2:20:13
I've been super excited about the idea of a email client built ground up around AI.
2:20:20
It seems like an obvious startup idea.
2:20:24
We saw like the the mailbox era where Dropbox bought Mailbox and there was a new email client startup every few years.
2:20:31
Uh now we're in this dangerous territory where you know you have a company like Notion where founder mode CEO at the helm plenty of resources clearly not asleep at the wheel understands what AI is capable of.
2:20:42
Uh but as a as a product manager or just a user of email uh how do you think about AI in the context of email? What are you hoping for?
2:20:53
I have an idea in my head of what I want, but I think I might be wrong.
2:20:57
Uh, just because that's the nature of these things.
2:20:59
So, how would you break down that problem?
2:21:00
How do you think about AI email?
2:21:04
Um, yeah, I think actually this is one I've thought a decent amount about.
2:21:06
Um, actually in the beginning, so if I go back like two years ago, two and a half years ago, my like theory of the world of like where what what would be interesting to invest in was there's kind of like three levels of the stack.
2:21:21
There's kind of like the foundation models at the problem.
2:21:22
there's all this at the bottom there's like dev tools and infrastructure in the middle and then there's like applications at the top.
2:21:26
Um and I was like generally pretty pessimistic on the dev tools and infrastructure space because I think it's just moving really quickly.
2:21:33
Um at the application layer seemed like the obvious place to invest and you can kind of subdivide the application layer into like um I would say like head, torso and tail um or like horizontal and vertical, however you want to split it.
2:21:49
Um, and I was actually pretty skeptical at the time.
2:21:52
It like it was clear to me like I should not be writing my own emails in 2025 or to like a very small degree.
2:21:59
Like uh there is a massive amount of corpus.
2:22:01
I've had like a Gmail account for 21 years at this point.
2:22:03
Like it should just be able to write every single email for me.
2:22:06
Um, but it's 2025 and I'm still writing my own emails.
2:22:10
Like it it maybe gives me like two-word completions.
2:22:12
Uh, and I think I have two two conclusions from this.
2:22:16
One is like the large tech companies are just I think way too riskaverse to build anything good here.
2:22:22
Um like I I like this Google must be able to build this.
2:22:28
Like it it's just it seems like a very trivial problem.
2:22:30
Um I assume like the risk of me accidentally saying something really offensive in email is just too high for them to launch anything here.
2:22:39
Um, and so I like and I think that's kind of true at large.
2:22:44
Like there is also like there is not a great assistant experience yet.
2:22:48
Like Siri is still pretty bad.
2:22:50
You can get like the meta AI assistant is actually pretty good if you're wearing like the meta glasses, but you have to wear the glasses to use it.
2:22:59
Um, and so, uh, I actually think the horizontal like, uh, experiences are addressable by startups now because I I think the big tech companies are just going to be too conservative.
2:23:09
Um, and then in terms of like the experience that I want, I mean, I I don't actually want I don't really want I'm like pretty good at email.
2:23:19
I'm like I'm I'm a very diligent person.
2:23:21
I like go through all of my emails. I get a lot of them.
2:23:22
Um, reading them is like pretty fast.
2:23:26
I don't need it to be summarized.
2:23:27
If someone sends me a really long email, like I'm probably not going to read it anyway.
2:23:32
Like I don't need it summarized.
2:23:33
Like it is a negative signal if you send a very long email.
2:23:35
Um the the hard work is replying to them.
2:23:37
Um and if something could just give like read the thing and like write exactly what I was going to write, that would be an amazing experience, especially if that worked on mobile because like composing on mobile is a total pain in the ass.
2:23:48
Y um so I'm amazed someone hasn't built that yet.
2:23:51
It it doesn't seem like there seem like there are a lot of hard problems out there.
2:23:56
that does not seem like the hardest problem.
2:23:57
Um, and so I'm I surprised.
2:24:01
Yeah, I mean, we're seeing a lot of like uh bolt-on AI narratives in the public markets.
2:24:06
Uh, are you seeing uh the AI narratives uh take hold in the fundraising around like scaleups, growth stage companies that aren't AI native companies, but they make so much sense in the context of of AI.
2:24:29
Um I don't know that's a good question.
2:24:31
Um I like I think AI is like very core to notion for instance and so I think notion has like a has like a very deep advantage here.
2:24:41
Um and I think it's just like the next evolution of notion.
2:24:46
I think um I don't know I don't know if I have enough data points on this.
2:24:50
I mean the thing about AI in my view is like it is not to I never went to business school so I may get the definitions of all this wrong.
2:24:59
It's a lay person's understanding but like in the sustaining versus disruptive innovation I think AI is like pretty clearly a sustaining innovation in almost all cases. Sure.
2:25:06
Um I think it like might be disruptive to Google.
2:25:09
Um and so assuming like a scaleup is still competent and can still build new product.
2:25:18
I think they can just kind of uh adopt this stuff as opposed to being disrupted by it.
2:25:22
So yeah, this is the uh Salesforce for mobile was just Salesforce. Yeah.
2:25:26
And so yeah, you could imagine that Salesforce AI winds up working potentially or at least uh they'll they'll hold on to that for a while.
2:25:36
But uh it is it is a fun time because at least at the very least it's like a new battle for everyone to fight out like uh you know the Gmail team needs to pay attention because notion's coming for them. I love to see it.
2:25:46
Uh, do you recommend sbaticals? Uh, oh yeah.
2:25:52
Um, I most definitely do too.
2:25:55
I had never So, I had like worked continuously from when I was 14 years old.
2:26:00
Um, I like worked at a startup in high school and college and then I took off like a week between college and Microsoft and like three days between Microsoft and Facebook and a week between maybe it was 10 days between Facebook and Sequoa.
2:26:13
Um, I took off almost two years.
2:26:16
I I say took off because I I kept like 15 boards and so I was on like uh 30 hours of board board meetings and one-on- ones a week, but it was still amazing.
2:26:25
I got to like play baseball with my kids uh almost every day and now they're pretty good at baseball.
2:26:30
Um, I knew nothing about baseball.
2:26:32
Um, and so now I know something.
2:26:33
Um, and I was I I think I was terrified that I would just be like irrelevant after taking some time off, but um I don't know, maybe I am, but it was like it was it was very still got it. It was amazing.
2:26:46
And I I just like I would rather uh like getting to hang out with my kids is like the best thing in the world.
2:26:52
So, I feel like saying that you're on a satical allows you to release the FOMO, which like as an investor, there's probably no stronger force, right?
2:27:01
It's like the founder wants to talk and Oh, I should probably take this call.
2:27:04
Why uh why why Conviction and why why partner with Sarah?
2:27:10
I have a bunch of like, you know, guesses, but I uh I'm assuming you could have basically gone anywhere.
2:27:14
Um uh you know, any firm, you know, raised 500 million out the gates on your own, you know, whatever.
2:27:22
Um so, I'm curious to hear.
2:27:27
Uh well, I've known Sarah for a long time, almost a decade.
2:27:29
Um though I I recently discovered that she sent me like a LinkedIn message in like December 2016 which I did not see until like two weeks ago.
2:27:39
But I I I thankfully ran into her in other contexts.
2:27:42
other contexts. Uh and I think um my my theory on uh I I'd spent a bunch of time thinking about emerging managers and like what is the right strategy for starting a new firm and I think like I think one thing that is underappreciated or maybe I I certainly underappreciated
2:28:00
for a long time is the degree to which like venture firms you know there's this funny thing where like people who work in venture kind of get like a little bit ruffled when you call their thing a company instead of a firm because it's not a company, it's a firm. Um, but you know what, like firms
2:28:15
Um, but you know what, like firms and companies are like first cousins.
2:28:19
Uh, and they're they're all just organizations.
2:28:23
Um, and I think that uh you know, if you're doing a startup, um, generally you don't say I have an exception to this, but generally you don't say like I'm just going to go like build the next Google like that that seems like a kind of crazy thing to do.
2:28:36
Um, you like you start off with a market that seems kind of small.
2:28:41
you dominate that market, then you expand to like the next tier of market and then the next tier of market and suddenly you build something gigantic.
2:28:47
Uh, and I think the same thing is true in venture.
2:28:49
Like it's hard to just go out of the gate and be like, I'm going to be the next gigantic global multi-stage firm.
2:28:55
I think you got to like pick a market and go dominate that market and then sort of expand from there.
2:29:02
Uh, and if you look at over the past like 10, 15, 20 years, the folks that I think have done this well, it's like Matt at Paradigm with crypto.
2:29:09
It's Mickey at Rivet with fintech.
2:29:11
Um I and so I you know AI was clearly going to be uh a thing the most important thing over like the next 10 or 20 years.
2:29:21
And so I think building a firm focused on AI and intelligent software and becoming the sort of preferred partner for founders category and then expanding from there seemed like the optimal strategy.
2:29:37
Uh, and I think that's very much the founding thesis of Conviction.
2:29:42
And, uh, I think like Sarah and Pranov and the team here, we're off to an amazing start.
2:29:47
Um, so I've known them for a while.
2:29:49
I think the thesis makes a ton of sense. Um, and it's really fun.
2:29:53
I think like one thing that is underappreciated in venture, back to this like firms versus company thing, is like venture is fundamentally predicated on the idea that there are things that like a 2-year-old startup can do that like are hard for a 50-year-old company.
2:30:10
Um, but then when you like but venture is like not always that introspective about itself.
2:30:14
Um, so I I think there is a belief that there are things that like a 2, three, fouryear-old startup that is in founder mode can do that maybe a 30 40 50 60 70 year old venture firm cannot.
2:30:26
So So, so when's the IPO? I know's going out.
2:30:29
General Catalyst, we're hearing rumors about conviction. I'd like to see them.
2:30:34
How have you uh what's your broad uh I'd love to give you a few minutes to talk about enterprise adoption of AI.
2:30:42
We talked last week about how Johnson and Johnson had basically piloted, you know, 30 plus different AI tools.
2:30:50
Wasn't it like it might have been, it might have been like 300.
2:30:53
I might have it off by like an order of magnitude.
2:30:55
It was like they basically tried hundreds of tools across different teams.
2:30:59
It was called like the thousand and then they've just been, you know, basically culling, you know, a bunch of stuff that was cool but didn't quite work.
2:31:05
Um what are what are you seeing as somebody who is I'm sure at times a big buyer on the other side at Facebook.
2:31:15
Um well I mean certainly there are there are things that are working like obviously cursor um cursor is working in a gigantic way.
2:31:22
I think Harvey is working in a gigantic way.
2:31:23
Decagon and Sierra seem like they're they're doing really really well.
2:31:27
I I I caught some of the conversation you were having with David right before this and you guys were talking about rappers a little bit and I like I I don't love the term rappers and to me there's like back to like the bottom, middle, top thing.
2:31:45
you can kind of like if you're going to build an interesting company, I think you either have to attack it from uh you have the handful of researchers that can literally push the state-of-the-art forward, which is like maybe eight organizations in the world or you know and understand and like love and can serve a type of customer better than anyone else in the world and just like build down from the customer.
2:32:10
So, you're either like building up from the technology or building down from the customer.
2:32:14
Um, and so I I mean I think that the companies that are working I think just like deep just have really good taste and like deeply understand what their buyers like what the c the customers need and are figuring out how to like adapt the technology to that.
2:32:31
And then I think there's a bunch of folks that are like not are kind of like hypothesizing what people might need but don't don't like viscerally understand it either because they haven't worked in it or they're just like they they haven't figured it out yet.
2:32:44
And I think that's the stuff that is probably turnurning.
2:32:47
But I mean I think it's kind of standard at this point in the cycle like some things some things are working some things are not.
2:32:52
I think more things will work.
2:32:53
I think the defining characteristic will be like uh a deep focus on the customer like the the buyer and like building a great product for them.
2:33:03
Uh do you expect big tech broadly to get more and more inquisitive?
2:33:07
to get more and more inquisitive? Uh we were talking with David too around you know just this era maybe 101 15 years ago when you had companies buying you know a single big tech company even a Twitter would buy many many many
2:33:22
companies and it was uh probably healthy for the ecosystem broadly and then you know now you know on the show last week we spent a bunch of time talking about you know the FTC's you know what they're doing to Meta around something that happened you know many many you know over a decade to go. Um, but I'm
2:33:38
Um, but I'm curious, uh, you know, what what's your outlook on on M&A? I mean, I hope so.
2:33:47
I mean, I think all the I think it all got shut down just due to government action and I think like misplaced government action.
2:33:53
Um, I mean, I I I think there hasn't really been anything interesting on the consumer front in over a decade.
2:34:03
over a decade. like probably I mean if you ignore Tik Tok and Tik Tok has like a million asterisks next to it um there I mean uh it's basically chat GPT I guess uh and chat GPT is obviously super interesting um but there haven't um you
2:34:19
know if you look at the 2008 2009 2010 era there was Instagram there was Snapchat there was there was Uber there was there was Door Dash there was Instacart um and part of that is just like the mobile phone and obviously like all the change that that ushered in. But
2:34:34
all the change that that ushered in. But I I think part of the problem is uh especially starting around maybe it's probably 2016 2017 there was just this crackdown on the ability to do acquisitions by big tech and I mean especially for consumer consumer has the this property that like
2:34:58
very very very few things are going to work but if it works it works really big but if it doesn't work it's often like really talented entrepreneurial founders that can then either like um take what they've built to a larger company and adapt it for that or work on something else at that larger company. And if you
2:35:13
And if you shut down the outs for those consumer companies, like eventually people just stop funding them.
2:35:19
So like every year some venture capitalist is like this is the year that consumer is back.
2:35:23
And I just I don't think consumer will fully be back until either there's like some massive platform shift or um like those companies can get acquired.
2:35:35
Speaking of speaking of consumer, uh I'd be interested to get your thoughts on consumer agents.
2:35:41
I was talking with a founder last week that had an idea for a consumer agent in the insurance space and it was a great idea.
2:35:47
seemed like it'd be a very very valuable product and I knew right I basically explained my point of view is that uh even if OpenAI is not explicitly saying that they're they're going to do this specific use case the sort of natural evolution is that at some point you will be able to use OpenAI to do a series of tasks like he was describing.
2:36:10
How do you think of the con do you see opportunities in consumer agents broadly or given the foundation model sort of emphasis on the application layer and building consumer products that it's sort of a dangerous path to be building on?
2:36:27
Yeah, I thought for a while there seems like there's some intrinsic tension between between being a great research lab and being a great product company.
2:36:37
Like the with the exception of OpenAI like most of the other folks haven't shipped that much product.
2:36:41
Enthropic is starting to ship some product.
2:36:43
And then OpenAI has had some uh complexity over the past couple years.
2:36:47
Um I I think I don't I think there may just be some like intrinsic tension between being great at research and great ated product.
2:36:57
And I think that is good news for founders that are building like horizontal products.
2:37:02
And so I do think I mean this is kind of like the email thing.
2:37:06
There's no there's no great like AI assistant yet for people and it seems like it is a thing that is possible to build.
2:37:17
Um, I think the more general purpose you are, like if you're building something literally for multiple billions of people, I think the higher the likelihood that you're like in a really square dead conflict with the foundation model providers.
2:37:34
But I think I mean this goes back to the like building down from the customer.
2:37:38
If there's some set of customers you can go build for that is even hundreds of millions of people.
2:37:45
millions of people. Um I think I don't know like a a company um a lab like OpenAI has um like they have so many important things to work on like there has to be and I have incredible respect for for OpenAI but it's like if you have
2:38:06
some strict stack ranking of the teams within OpenAI like only 10 problems can have like the top 10 teams on it and so if you're any if you're doing anything after like the top 10 problems that they have, then you're like by definition don't have one of the top 10 teams working on it. And that seems like a an
2:38:20
And that seems like a an opportunity for a startup. Great opportunity. Yeah.
2:38:23
Well, a great conversation.
2:38:25
I have a lot more questions. Yeah. Yeah.
2:38:26
We'd love to have you back soon because I mean, you can just talk about everything that's going on. It's amazing.
2:38:31
Uh we really appreciate you taking the time. So, thanks.
2:38:34
Thanks so much for having me. Okay. Have a good one. We'll talk to you soon. Bye. Bye.
2:38:37
Uh next up, we got Edward Mayor from uh I'm gonna mispronounce it.
2:38:42
It's I think it's Machina Labs.
2:38:44
I got it wrong the last time I tried to pronounce a word like that.
2:38:48
Uh but we'll hear it from him.
2:38:48
Uh he's got a bunch of a bunch of interesting topics and he uh is going to talk to us about a bunch of stuff.
2:38:57
Uh let's bring Ed in and then we will have him introduce the company so I don't mess anything up and uh we'll go from there. Ed, how you doing?
2:39:04
Hey guys, good to see you. What's going on?
2:39:07
Uh would you mind just uh kicking it off with a little background on the company?
2:39:10
Uh maybe how you pronounce it.
2:39:12
It's Machina Labs, right? Yeah, you're right. Machina Labs. Fantastic.
2:39:16
Uh, and yeah, just give us a brief overview for the listeners. Yeah, for sure.
2:39:23
Um, you know, I think Mocking Labs is a response to something that maybe actually Peter Teal said once, right?
2:39:29
Um, I I I don't want to butcher it, but was along the lines of like, you know, we wanted flying cars, we ended up getting, you know, 140 characters, right?
2:39:36
right? um which kind of comes down to the cornerstone of like okay software develops really fast hardware doesn't I think you know some people interpret that as a um you know as something that says oh like we don't have as much glass still in the hardware side but reality
2:39:50
is there's a lot of technological challenges in terms of um building building hardware right so what we're trying to do at mock labs is just making that much easier right you know if you're a software developer you know two people can develop a program you know rent server from AWS Amazon deploy it. If you want to build a hardware, you
2:40:07
If you want to build a hardware, you pretty much have to go build a factory, right?
2:40:10
Um, and that's why development timelines are like seven years, 9 years.
2:40:15
Um, so what we're doing at Machina is we're using robotics and artificial intelligence to build a basically what we call a robotic craftsman.
2:40:20
And it's a robotic system that can do different types of manufacturing operations.
2:40:24
Um, and it's powered by AI.
2:40:26
basically can figure out how to pick up different tools, do different types of operation to make different types of parts, physical parts, um, without anybody having to program it or having to handhold it, um, or build tools for it.
2:40:39
Uh, and can you talk about, uh, the specific kind of like first instantiation of the thesis?
2:40:43
I saw Justin Lopez over at uh, Bass Powering and it looked like he used a picture of your warehouse. Is that you guys? Is that right? That is us. Oh, yeah.
2:40:53
And so he breaks he breaks the new uh the new manufacturing space down into three categories.
2:41:00
Manufacturing SAS companies, make parts for other people companies, and change the way the part is made companies.
2:41:08
And uh which which one do you guys fit in?
2:41:11
I think you're in the third, right?
2:41:12
Change the way the part is made.
2:41:14
Uh developed a way to make stampings complex geometry form sheet metal parts without dyes.
2:41:18
And so, uh, can you break us down like what is actually going on when this massive robotic arm is, uh, like pushing into this metal?
2:41:26
Explain why this is important, how it works. Yeah.
2:41:29
So, we wanted to build a system.
2:41:32
We call it robo craftsman, right?
2:41:33
The idea is like where do we get started?
2:41:35
Um Um, and we started as sheet forming.
2:41:38
Sheet forming is largest metal processing sector today.
2:41:42
I think it's like $280 billion industry.
2:41:43
You know, most of the metal parts you see dayto-day are sheet metal parts.
2:41:47
parts. like you know you're sitting in your car you like you know you're in a sea of sheet metal every other car body is sheet metal form sheet metal parts or aircrafts are basically sheet metal cans um but today it takes a very long time to get your first batch of parts in sheet metal world right you have to go
2:42:01
make dyes put them in giant stamping presses like fourstory tall buildings um and then stamp your your you know your parts out um so our first kind of application of being sheet forming um we have two robots that form um you know start from flat sheet amount metal between two robots. They have these
2:42:17
They have these giant fingers that are super strong, but they can basically push and pull on the metal that form it into a very complex shape without the need for any any dyes or tooling.
2:42:27
Um, basically, you know, you get your from idea to to the first part in a matter of hours, right?
2:42:32
Um, it's similar to how a potter forms a clay bowl with their fingers.
2:42:36
They're coming into the sheet, they form it um and and and shape it into different shapes.
2:42:40
So, yes, it's a new um paradigm, a new way of doing manufacturing.
2:42:45
Um, but it doesn't stop there.
2:42:48
We're already doing like trimming.
2:42:49
We're already doing uh slotting, hole making.
2:42:51
So, the robot can literally pick up another tool, figure out what it need to do with it to do the next operation and does that.
2:42:57
So, today does forming, it does a lot of subtractive work like trimming, hole making.
2:43:03
It also does a lot of QC.
2:43:03
Are you do you get worried?
2:43:05
There's been a ton of there's been an explosion of new manufacturing startups over the last couple years.
2:43:11
couple years. um you know uh long since you guys started the company many of which promising to automate with robots the creation of various products um we covered earlier today on the show just how hard it is to like manu for Nike to manufacture a shoe right which sounds somewhat trivial like you know humans have been make doing them for years you should be able to uh maybe we should be
2:43:33
able to do it with machines but then in practice it's like there's so many different factors down to temperature um out of all you When you look at manufacturing broadly, are I'm I'm assuming you have some type of framework for evaluating whether something like can be automated to the degree that people would like to see out of manufacturing or or areas that you know basically shouldn't be touched, right? Like something like like shoes which
2:43:59
Like something like like shoes which have infinite sizes and a bunch of different factors. Yeah.
2:44:03
Yeah, I think it's like a combination of like three things like the market size, opportunity and how how technology ready it is to to to be to be kind of disrupted, right?
2:44:14
So, um uh but also I think there's a lot of conversation around automation that that is in the previous paradigm.
2:44:20
I think for now for the first time we have this concept of LLM this concept of we can actually reason very complex sequence of operations as long as we can train the robots on that sequence of operations.
2:44:35
So it comes down to what data do we have available to train the robots.
2:44:38
We already figured out okay you know neural networks LLMs uh these uh transformers if you give it enough data it can actually learn a very complicated complicated task.
2:44:48
The real key was okay what where do we generate enough data where do we have enough data to train it.
2:44:55
it. Um so it needs and and unfortunately for a lot of manufacturing tasks the data is not out there right you cannot train a very complicated model on it for you know for chat GPT the internet had trove of free data that you could use to create a very complex um kind of chatbot for us coming up with right sequence to
2:45:14
make car doors um or shoes doesn't exist yeah so the key is can you actually provide a solution that can scale with limited amount of data with human intervention limited amount data so you can deploy it in the field and get enough dra enough enough traction so you can have now enough data that comes from your machines to train your model. Can
2:45:33
Can you talk and that's why they went after she form. Yeah.
2:45:36
Can you talk a little bit about the N uh the Nvidia announcement today? That's very exciting.
2:45:41
Uh seems like it might be a little bit of like a side quest or is this like in the critical path to you know mass uh mass production? Yeah.
2:45:49
No, Nvidia's today announcement was was a kind of little fun thing we did.
2:45:54
Nvidia is an investor in us. So cool.
2:45:57
Fundamentally they are very interested in what we are trying to do.
2:45:59
You know being able to capture data from physical phenomenon and build models that can manipulate the physical world. Cool.
2:46:04
Um but I think we work with their artist and resident.
2:46:08
This actually open AI is artist and resident that we work with Nvidia collaboratively.
2:46:11
Fun to really turn a just the artist speaking to a system into a piece of art. Right.
2:46:16
So, Alex, the artist that would work with us, basically spoke what he wanted to build, what sculpture he wanted to build, and the full stack of generating the code, running the robots, all were done autonomously.
2:46:29
So, from speech, from intent all the way to the physical part without anybody ever touching anything. I love that.
2:46:35
I I was I was when the when the Studio Giblly moment happened, I was taking some photos of like my kids playsets, studio giblifying them, and then immediately I wanted to print them out because I wanted to have some sort of like physical instantiation, just showing the phone was like not satisfactory.
2:46:50
So the idea of like speaking words and then getting like a sculpture out like that's that sound really awesome and futuristic. Yeah. Yeah. Go.
2:46:58
How uh a lot of an investment in humanoid robots lately.
2:47:05
many of those promising to revolutionize manufacturing, replace human labor, automate the production of lots of things.
2:47:11
As somebody who's been doing it with robots since 2019, uh how how do you think of that form factor in the context of um manufacturing?
2:47:23
Yeah, I'm actually super bull super bullish on on on human rights, right?
2:47:25
I think the question is to to your point, is it going to be is the first application going to be in uh in manufacturing? I don't think so. Right.
2:47:35
Um I think the biggest problem if you want to like think of a startup as like what do you need to derisk first the first thing you gave an example of like Nike figuring out how to do the shoe manufacturing.
2:47:45
The first thing is actually intelligence.
2:47:46
We have kinematic frameworks for a long time.
2:47:49
You know we could do what humans does in terms of kinematic freedom with industrial robots.
2:47:55
We can actually do it more precisely with higher force with industrial robots which was what we need in manufacturing setting.
2:48:00
So the missing piece really was intelligence.
2:48:02
So we're kind of a little bit intelligence first, right?
2:48:05
We don't need to solve the joints.
2:48:07
We don't need to solve people walking around like robots walking around and you know having the human form factor.
2:48:12
If you solve the intelligence um you have enough kinematic frameworks which is industrial robots to do what you need to do.
2:48:18
But that being said, I think you know humanoids is huge opportunity maybe not in manufacturing but but downstream in homes and all other places that that that that we can consume it.
2:48:27
What is uh what is the you know I imagine you guys are a beneficiary from some of the tariff stuff in some way but on the actual machine side are you are you know you and the industry seeing challenges of like you know h how much of like the actual robots that you guys are leveraging are are sourced uh outside of the country and yeah Chris at Hrian was saying like yeah my capex goes up by 10% but the demand demand has way way higher.
2:48:59
So what's been your experience? I mean that's true.
2:49:02
I mean like what we can what we do here.
2:49:03
So so the the challenge is for a lot of work that we do and especially in the defense aerospace defense there is no industrial base. So we have to do it.
2:49:12
So demand is always there.
2:49:14
Now 70% of our bomb is off the shelf and we intentionally try to do that so that we can actually finance it easily right use you know multi pusher offtheshelf equipment so that we can easily finance um the full harbor stack.
2:49:28
That being said, you know, a lot of the hardware that we use is either produced in America or it's in very like allied countries.
2:49:34
For example, robots were produced in in in um Japan.
2:49:36
Um so the tariffs are a little bit more digestible there.
2:49:42
But that being said, you know, we are one of the very few manufacturing companies that are like doing almost software like merchants.
2:49:47
So there is enough room for us to to be able to absorb that cost and still, you know, have a very high value, high margin. Interesting.
2:49:55
Interesting. is the idea on that note is the idea you know we've seen we've seen some investors kind of uh you know strictly software investors kind of poke fun at at VCs that are investing in manufacturing you know expecting uh
2:50:10
software-l like margins and and there's sort of this sense that uh well manufacturing hasn't for most things hasn't historically had software-like margins you're seeing it today um how how do you think about the margin profile for advanced manufacturing over
2:50:27
time even you know even in things that are like non-chips right yeah yeah so um so so it's it's it's interesting arbitrage I think there's a lot of people thinking about manufacturing solutions in different ways I think you can think of it as
2:50:43
automating what traditionally has been done and I think that's usually end up being very low margin right but if you're creating something new that you know has some kind of arbitrage on either labor or has some kind arbitrage on equipment, right? In our case, for
2:50:55
In our case, for us, you don't have to make dyes, right?
2:50:59
And a die, a single die for um you know, a car door can be up to a million dollars, right?
2:51:05
So, for us, we're faster, but we get rid of that asset.
2:51:08
Um so, it allows us to have the you know, at the same costerity, have higher margin, but I think yeah, down the road, um the margins could erode.
2:51:16
And that's why we're thinking about it as a platform, right?
2:51:20
It's a robotic system that can do forming today.
2:51:22
HMR is going to do your next operation.
2:51:24
It's going to do bending.
2:51:25
It's going to do hemming.
2:51:27
It's going to do forging. Right?
2:51:27
So, you're constantly expanding the capabilities of the system which allows you to sustain a very uh larger volume business um as some of the margins of the older processes kind of erode. Right. Um yeah, last question.
2:51:42
Um uh you you mentioned that you might have a Golden Dome take.
2:51:48
We did a little deep dive.
2:51:50
There wasn't a lot to dig into. I'm curious.
2:51:52
Uh, it doesn't seem like it would interface with your business too much, but maybe you just have studied the industry.
2:51:58
Uh, so what do you think's going on with the Golden Dome?
2:52:00
He's making an actual dome. Dome. Yeah.
2:52:02
Make a physical out of gold to go over the United States. I would love that.
2:52:06
If you like the blue, get, you know, enjoy the blue above us while you can.
2:52:10
It's going to just be all gold soon. Yeah. Yeah. Yeah. That's right.
2:52:14
I mean there was like people were talking about how it's like compete with a golden dome at Nordame which is like actual thing um in terms of the name naming conflict. No.
2:52:23
So um why do we actually fit into this paradigm?
2:52:26
We we work a lot with missile manufacturers right um and manufacturing components for them the body of the missile for them.
2:52:31
Um hypersonics is one of the main enablers of how we can actually have defense against some of these uh missile attacks.
2:52:39
So for hypersonics, you need to start processing very complex materials.
2:52:44
Materials that were traditionally very hard to to process.
2:52:46
Think of high temperature alloys like nickel. Yep. Uh titanium.
2:52:51
Um and and because of that robotic process that locally manipulates the material, we have way more control to process the material without getting to failure.
2:53:02
For example, we can form titanium sheets without tearing it in at room temperature, which is traditionally not possible, right?
2:53:06
So I think there's a lot of interesting um opportunities there for us and we're exploring with our you know with our primes that we work with.
2:53:16
Uh but yeah I mean the whole concept of Golden Dome is going to be interesting.
2:53:20
You know what I've heard last is it's going to be 7% of the DoD budget in the next two three years right once it becomes pro programmatic. Yeah.
2:53:27
So that's that's what I what I've been great obviously next year we're looking at like I think 20 billion on missile defense but over long term I think the plan is roughly 7% of defense budget.
2:53:37
So, um, it it's going to be a huge opportunity and obviously I think something that's probably very necessary for America, but uh, yeah, we can dive into details, but I don't know how much time we have. No, no, no.
2:53:45
I mean, uh, I I think we'll have to have you back on when there's more details that emerge and we actually get, uh, you know, a a deeper dive into how the uh, how the program record evolves, what the subcontractors might be looking at.
2:54:00
Right now, all the all the names parties are no comments, so there's not too much to dig into for good reason.
2:54:04
Um, but it'll be super interesting to see how this uh pans out. Uh, Jordy, anything? No, this was great.
2:54:10
Yeah, we got to have you back on soon, but uh congratulations on all the success and uh I'd love to come see the the machines in action in person.
2:54:19
Yeah, we're like we're half an hour away from downtown, so downtown LA. So, both of you. Yeah, that'd be awesome. Let's do it.
2:54:24
I want to I want to do the show with with a robot, you know.
2:54:26
I want a robot to make us a bigger gong. Biggest gong possible. Got to make it happen.
2:54:30
We got to make it happen.
2:54:33
That's something you actually do very easily. I Yeah, I can imagine.
2:54:36
It's a perfect It's a match made in heaven.
2:54:38
Our people will talk to your people.
2:54:40
Our people will talk to your people. Uh this is fantastic. Thanks for coming on. Thanks so much, Ed. We'll talk to you soon. Bye.
2:54:45
Uh next up, we got William Brown.
2:54:48
The idea of like a a like a 20 foot gong is very appealing to me.
2:54:55
Very appealing that you could have watched be made from the raw material. Yes. Be be perfect. Be perfect.
2:55:02
And I mean, you could actually etch different things into it.
2:55:04
There's so many cool things you could do with that technology. It's very, very awesome.
2:55:08
Anyway, welcome to the show, Will. Uh, how's it going?
2:55:10
Good to have you on here.
2:55:12
Uh, been a big fan of your of your posts for a long time.
2:55:15
One of the top Many people have said one of the top post poster of the year, I think.
2:55:21
Potential poster of the year for sure.
2:55:23
He's been burning up the timeline.
2:55:23
Uh, Jordy, where where should we start?
2:55:25
I mean, I have a bunch of I have a bunch of stuff that we can go through.
2:55:29
I mean, we can just dive Do you want to give a brief intro on yourself and and who you are before we Maybe we should start there.
2:55:37
Started talking about the timeline. Sure. Yeah, sounds good. I'm Will Brown.
2:55:40
I'm a researcher at Morgan Stanley.
2:55:42
By the way, nothing I say here is Morgan Stanley opinions.
2:55:44
This is all me kind of just like sharing my thoughts. Um, I am a researcher.
2:55:47
I work a lot on stuff related to LMS there.
2:55:49
Uh, my background was in reinforcement learning theory.
2:55:54
I did my PhD at Columbia, so I've been in New York City for a while.
2:55:58
Big fan of New York City.
2:56:00
uh great things happening here all the time.
2:56:02
Um and yeah, I also just like like talking about the stuff on the internet and like participating in the open source community.
2:56:08
There's lots of like cool either projects or code bases or papers or models always coming out and like I a lot of my job is like needing to know all that stuff and being a liazison essentially like from the internet to the company where like people want to understand like okay what's the latest model for XYZ that I should be using?
2:56:27
the right open source toolkit, especially because like at a big regulated company, we need to understand the landscape, especially for things that are like downloadable off the shelf without needing to like onboard a vendor.
2:56:37
Like some things you a vendor for, but it's also like a much heavier lift.
2:56:40
And so like we need to like map the landscape of like what's the right tool for the job for everything LM related.
2:56:46
And so that's a large part of what I do.
2:56:49
And it was also like my excuse for being on Twitter all day. That's great. It's a great excuse.
2:56:54
Where do you want to start?
2:56:54
Where do you want to start? uh where uh aside from X like where where are you getting signal without giving away without giving away all the alpha because I imagine like you know by the time Darkh has like done a podcast on something like you know it's been that the information you know disseminated it's not necessarily alpha
2:57:14
there's a surprising amount of alpha still on X just from places where you don't like like so there's places uh I have not found much alpha on LinkedIn um the group chats, the anons, um the uh open source uh like GitHub discussions, lots of really good stuff is buried in like a GitHub issue or like a feature request where someone's like, "Hey, this thing would be cool." Yeah. Yeah.
2:57:40
And these ideas are just all over the internet, but you got to like know where to look for them. Yep.
2:57:42
Uh I want to talk about I want to talk about humor.
2:57:45
You actually posted about this like a month ago or two months ago.
2:57:49
Uh, you said it's interesting that 1.
2:57:51
5 billion parameters is all you need to crush math competitions, but you need like 15 trillion to make the model be funny.
2:57:58
Maybe humor is the right measure of true intelligence.
2:58:01
And for a long time, my eval has been tell me a funny joke and every time there's a new model, it's tearing up X and I ask it to tell a funny joke and it's always the worst joke I've ever heard.
2:58:11
It's kind of an anti- joke.
2:58:13
Um but but but h is that just a lack of of like uh of labeled data essentially uh or do you think there's something more uh more like innately human to the idea of humor?
2:58:24
I mean, I think humor is like really hard and it's also very hard to it is very hard to label, but I think it's also really hard to like, okay, this is like a a big debate topic is like would it like I think people kind of assume, oh, just like train it on funny data. Yeah.
2:58:43
But like the things that make things funny are really subtle and pretty buried.
2:58:47
Um, I a friend of mine, co-orker actually did this experiment where he like he tried to like have one model be a judge of like is this funny or not? Yeah.
2:58:56
And what they do is they like lean into things that feel funny and that the results were actually kind of funny, but what it ended up just doing was swearing more.
2:59:03
So like if you swear a lot, models think that's funny.
2:59:06
Um, and so there's all these like kind of cheat like there's a lot of cheat codes. Sure.
2:59:11
and RL and training these models is like the path of least resistance is to find the cheat code.
2:59:17
Um, but like real humor like the people that to be really funny, you have to be really smart.
2:59:22
Like think of like your favorite comedian like uh Norm or Lar David or like all these people are like really smart.
2:59:27
You can tell from the way that they compose jokes like there's like an attention like attention is in like transformer attention of like ideas that need to combine in a very sparse like precise way to make a good joke is like you can't just like shove things together.
2:59:46
You kind of got to thread the needle to make a joke land.
2:59:49
Um it's not a very coarse mechanism.
2:59:49
And so I think like uh GPT 4.
2:59:54
5 is like ginormous model trillions of parameters most likely.
2:59:57
Um and that like there's more room in the model to have these like little sparse connections materialize as you go through layers of the transformer.
3:00:08
Um and I just haven't seen anything like that come from a smaller model.
3:00:13
How scale are you based on the results from 4. 5?
3:00:19
Are you off to the races?
3:00:19
Scale is all you need or lesson pill or what?
3:00:22
I'm very RL scaling pled like uh I'm not big transformer pill really like the training wall real um I think we don't like is if there's another 100 trillion tokens of data sitting out there ready to train on go for it. Um, I don't know.
3:00:40
Like one I think is like people should do some math and math about like how long until big GPUs are readily available and it's it's going to be a while before people can really run like even DeepSec R1 with like easy resources.
3:00:56
Like you can you can kind of do it on one node, but like the quality bump over things that are much smaller is just like like the we're hitting diminishing returns on capital investment is a lot of it.
3:01:10
Like they're taking out of the API because like the they can sell other things with the same GPUs and make more money is part of it.
3:01:18
Like it's a lot of compute to keep a model up like that. It's slow. Yep.
3:01:21
and the things that it's better on are like not that economically valuable.
3:01:27
And so the sweet spot appears to be in this like between like 30 billion active parameters and or parameters total and like couple hundred maybe.
3:01:37
But like I don't know that like like when Meta releases this behemoth model, I don't think anyone's really going to run behemoth. Sure.
3:01:45
For like their day-to-day stuff. Yeah.
3:01:47
It's just probably not going to be worth it.
3:01:49
So I mean given that uh you know Tyler Cowan's calling 03 G AGI like the the the economic results of these like big but not crazy big models seems to be pretty good run it on a node.
3:02:01
pretty good run it on a node. Uh should we be talking more about hey it's good enough let's bake it into an ASIC let's just bring down the inference cost to basically zero like we did with Bitcoin hashing and whatnot like is that the conversation we should be having or I
3:02:17
mean I think that's been that's kind of been like Groanova are like playing that game um and I I think like those businesses make a lot of sense like they can kind of say okay this version of a transformer we're going to be in this ballpark for a while we can bake that into our plans a little bit. Um, I think
3:02:32
Um, I think what's the next way 2025 like people have been saying you're the agents, but like what I think that means is agentic RL like we're realizing RL works like the reason 03 is good is because it's trained to use tools.
3:02:46
The way you train a model to use the right tool for the job is reinforcement learning.
3:02:49
Uh, and they've said as much like deep research reinforcement learning.
3:02:52
Um, people have been throwing like random tools at models for two years and it only now works even though the models are the same.
3:03:00
like GPT4 was a bigger model than a lot of the models people are using now.
3:03:04
It had just as much training data, but it was not or at least doesn't seem to have been RL in this specific way.
3:03:12
And so that's kind of like my bet is like people are going to really want to train models to be agents.
3:03:17
And I think you can get that to work well with a pretty small model.
3:03:21
Does that mean like a flourishing of RL uh big transformers for different tasks or are we still searching for like the god model that can do everything all at once?
3:03:33
I mean, so okay, I think there's a a couple leaps we need to have models that can do everything all at once for like a super long amount of time.
3:03:40
Um like my I tweeted something about this like my what I'm happy to call 03 10-minute AGI. Mhm.
3:03:48
And I think like framing AGI in terms of like length of time it takes a human to do a task is like more reasonable than like a global framing.
3:03:57
Like sure there's a bar of like drop and replace for a human that we are like definitely not at yet like for general jobs but most things a human can do in 10 minutes you can like get 03 to do that pretty well.
3:04:08
Um and so that's a very even like a like make a website in 10 minutes. Okay.
3:04:16
You have a good human designer is like probably can whip something pretty quickly. Yeah.
3:04:20
But but not like an entire design system that works together across all the different Yeah. makes a ton of sense.
3:04:26
Um so uh so so uh seems like we're transformer maxing, we're RL maxing.
3:04:31
Uh is there a new paradigm that you're excited about? Uh program synthesis?
3:04:35
Are we bringing back symbol manipulation at some point?
3:04:39
Like what what are we going to pull from the tool chest to make this thing go to the next level?
3:04:42
I mean I think it's just tool calls.
3:04:45
Like like I think when people say program synthesis like we're already there like 03 is program synthesis but the programs are like JSON and Python.
3:04:53
Y like you can do a lot with that.
3:04:56
I would have I did some tests on like RKGI problems where you give the screenshot to 03 and it can like basically figure it out in 10 minutes of like zooming in and looking at the thing and then writing some code to see if it like reproduces the thing and like it doesn't nail it but I would they haven't released the benchmark but I would guess it'll do reasonably well. Yep.
3:05:12
Um what as much what are you seeing around AI adoption in finance broadly?
3:05:18
adoption in finance broadly? I feel like we've been promised like, you know, somebody can just press a button and generate the deck and like generate, you know, a 20page investment memo and and uh are are are these types of uh you know, deep
3:05:35
research style tools being used extremely heavily already or is there just even a greater now that they're being used, you know, hey, let's go uh spend our time talking to, you know, three times as many experts so that we actually get proprietary kind of um insight into the into the business. Great question and I think there's a
3:05:56
Great question and I think there's a couple different answers I have.
3:05:58
couple different answers I have. one is that like on one hand we at least have been pretty I think fast at certain things at adoption like the day GPT4 launched Morgan Stanley had integrations because we had been working on it and these were like we had press releases
3:06:12
for these like we are ready to go and so there are certain things where like initiatives can happen where a effort comes together to make a thing ready to use but the like there's a long tale of smaller tasks that you do that you there's not a drop in replacement for if
3:06:30
there's a vendor it would take a long time to onboard them and it's not going to really solve the problem right away and you could build something from scratch if you put a few people on for a few months but like engineer your hands are like few and far between um and so I think like deep research is an example
3:06:47
of a product that I would say like works really well um for the thing it's built for it's doesn't quite like do other stuff super well like if you wanted to give you a table of like 50 very precise things it's going to make some mistakes there. Um because the report format has
3:07:00
Um because the report format has a lot more room for like slop um without it seeming like seeming like noticeably bad.
3:07:08
Uh the PowerPoint Microsoft PowerPoint copilot is not great.
3:07:11
It's not a thing that I have heard many people say saves them time. Um uh what are you Yeah.
3:07:23
Do you have do you have any takes on enterprise AI adoption broadly?
3:07:26
We were talking, I think late last week, about how Johnson and Johnson had tried hundreds of different tools.
3:07:32
Like they went through the effort to just like try a bunch of stuff because there was a top down mandate and now they're just like cutting probably 90% of it.
3:07:40
And so all the startups that were like, "Yeah, we have a pilot with J&J. It's going great." Like they won't churn.
3:07:47
Well, 90% of them just churned maybe.
3:07:49
I mean like most of these pilots are very much intended to be churned.
3:07:53
like they're not they're very much in a they're not being rolled out broadly.
3:07:57
They are coming through in a kind of walled off environment for people who are like going to be the beta testers.
3:08:04
And so like we have a crew of people who like and I'm involved with this as well like there as new stuff comes like online we test it out we give it feedback.
3:08:14
Um, most of these do not convert um because like we're very willing to like try stuff um in terms of taking the call in terms of like looking at a demo but actually making a thing be part of like the companywide workflow is like a pretty heavy lift.
3:08:32
Um think of like onboarding a cloud provider.
3:08:34
Like a lot of the reason like maybe this is like one reason cloud providers have the margins that they do is because like moving clouds is really hard.
3:08:42
um moving all your stuff from like AWS to Azure is a pain.
3:08:47
Um I think especially in regulated industries like a lot of software onboarding faces the same sort of hurdles um where you can't just like bring it in and use it.
3:08:58
You got to like go through a whole process and some companies are have kind of planned for that and some have not.
3:09:05
So like one example is I think like a reason that Windsurf has been successful as a cursor competitor is they lean they've leaned way harder on enterprise than cursor has like they have really designed for enterprise integration whereas cursor really has not.
3:09:22
Um do you uh was the windsurf news surprising to you at all or was that just obvious?
3:09:27
OpenAI cares about coding and they should have a sort of dedicated enterprise, you know, coding. I mean, it makes sense. Yeah.
3:09:36
I think like to me it felt like the a sign of there being some more friction with Microsoft because like originally it was like, oh, they already have that as co-pilot. Sure.
3:09:45
Um, and so this to me seems like them move taking a step away from Microsoft.
3:09:56
Um, but I mean this is just speculation. Yeah.
3:09:59
Um, it does make sense that it's cursor and that is Wind Surf and not Well, I mean they did try to buy wind cursor supposedly.
3:10:06
Um, like I'm I'm a cursor user. I think it's great.
3:10:12
I haven't seen anything that really sold me on like go move to Windsurf. I'm sure it's fine.
3:10:15
Um, it's just like the bar to like switch for me is like I need a thing to have a a feature that's like killer that the other one doesn't have and I have not like seen that yet, but I'm sure they have plenty of good stuff.
3:10:28
What uh now that we're like a couple months out from the Deep Seek moment?
3:10:32
Uh, what is your takeaway?
3:10:34
Is it something around kind of the the optimization of of of the inferencing these models or uh h how have you processed uh that news now that we're a few months out?
3:10:46
Yeah, I mean they're like incredible engineering like I think a lot of their open source releases have really like like since R1 they've released a lot of code and details on their inference stack and training stack.
3:10:59
Um, even even Sam Alman was posting like, hey, if you work for a high frequency trading firm, like come work at OpenAI.
3:11:05
And I read that as like, oh, they want to optimize their models now, right?
3:11:09
I think the some of the big players have realized they didn't have to that much.
3:11:13
Like, of course, uh, Deep Seek is serving all of China on 2,000 GPUs.
3:11:19
It's kind of silly that Anthropic can't like has to have the warning about, oh, we have high limits.
3:11:24
Please try again later for like their paid users.
3:11:28
like that's whereas like the deepseat chat is literally free anywhere.
3:11:31
Um and so I think some other people probably are working on upping their uh inference efficiency game but it's also like hard um because it's very model specific.
3:11:41
Every model has its own quirks and you got to optimize around that.
3:11:44
Um it's hard to like have things be reliable and fall tolerant.
3:11:48
Um, so also just so you think you can't just port back like FP8, wasn't that one of the uh things or like the the the mixture of experts blocking all those different things?
3:11:58
Like it seemed like there was some stuff that where it was like week two we were getting, you know, analyses and I was like this will probably be open sourced and ported to Llama and all the others like pretty quickly.
3:12:07
But has it played out differently?
3:12:09
I mean to me the surprising thing was not any individual one thing.
3:12:14
It's that each of these is maybe like a 30% 50% gain, but they have like 10 15 of them that all stacked.
3:12:19
And so getting all of these to stack nicely is what's hard. That That's interesting. That's a great take.
3:12:25
Yeah, that's interesting.
3:12:26
What are you expecting out of uh Alibaba and and Quinn three?
3:12:29
Oh, I I mean I'm really excited.
3:12:32
Like they make I think still the best model suites for like doing research.
3:12:36
So like I do almost all of my experiments on Glen models just because they have a bunch of them.
3:12:43
There's like so many versions like for every different model size there's like seven different versions.
3:12:47
There's like a code one, a math one, a multimodal one, an audio one, a base instruct RL like they really are optimizing for user friendly like open source model ecosystem um in a way that uh Llama was doing last year.
3:13:02
This current year we'll see if they can get their act together.
3:13:08
Um, so are they are are they also like RLP pill at this point and and moving? I'm sure they are.
3:13:14
Yeah, like they have they have a reasoning model.
3:13:15
The reasoning model is not as impressive like but it's it is a pretty small one.
3:13:20
It's like a 32 billion parameter model that like does well in math competitions.
3:13:25
Um, I don't know that anyone has really bitten the RL bullet in the way that Deepseek did early and then OpenAI has been doing lately. Sure.
3:13:37
where they're really kind of betting like OpenAI seems to be essentially betting on scaling up RL as the path.
3:13:44
Um, and that it's not just longer reasoning, but it's reasoning integrated with system interaction.
3:13:49
And I sure like that's kind of the drum I've been trying to beat for a while.
3:13:53
Like a lot of the open source work I've been doing is around uh multi-turn tool calling RL.
3:13:59
Um, I think that we need better ecosystems for that.
3:14:02
like very like there's starting to be more tools you can go use, but for a while there was just nothing that supported this.
3:14:10
Everyone was very RLHF pill. Yeah. For too long.
3:14:12
What what do you think is going on with Llama 4?
3:14:14
Did they just miss the memo about RL or uh is there something else at work?
3:14:19
I mean, doing this stuff at a huge company where a lot is on the line is hard and it's way easier to do nothing than something. Sure.
3:14:30
Um, and Meta also like doesn't have to print money off of their models.
3:14:37
Like they don't uh like I think one analogy that I've been giving to people is like why did Amazon not win NLP?
3:14:42
They had like 10,000 people in the Alexa team. Sure. In like 2018. Yeah.
3:14:49
And they have just now released like okay language models.
3:14:55
Um and they're really betting on Anthropic to drive their revenue there.
3:15:00
Um how are you thinking about uh the trade war in the context of the data center supply chain and is that even an do you pay much attention to that or is it kind of you're busy enough on on model? I do follow it. Yeah.
3:15:15
I mean um I tend I listen to Dylan Patel talk a lot of stuff. Um he's great. um semi analysis.
3:15:25
It's hard for me to like give a real take.
3:15:28
I don't think I have I think I think it's like important to keep an eye on.
3:15:31
There's a lot of pieces of the supply chain um that will get I mean it depends on what happens like Yeah, it's going to get messy for sure probably.
3:15:39
Um but I I don't have like a a hard stance.
3:15:43
You should become a VC and then you can just give it anyway. Exactly.
3:15:46
Yeah, you should you should flip over.
3:15:48
No, I just think it's funny like uh in a year we'll be like, "Bro, we were worried about the wrong Transformers." Yes. Yes. Yes. For sure. With the energy thing. For sure.
3:15:55
Um I I have one last question.
3:15:57
Um uh h how's the culture at Morgan Stanley?
3:16:01
I mean, it seems like talking to you, you sound like a Silicon Valley founder, uh but you're at like a company that's like over 100 years old.
3:16:08
Uh is are are you part of a new guard or has Morgan Stanley always been this way and you're just the first to kind of post about it?
3:16:17
Uh what's it like working there?
3:16:17
I'm definitely like the first to like tweet a lot about it, but like the team I'm on has been around for maybe like it's like a machine learning research team where we try to we try to be like a version a finance version of like an MSR or a Bell Labs where we like sure off keeping up with the research.
3:16:35
We write papers, we publish, we go to conferences, we tend to we kind of hang around as like expert consultants and advise on a lot of efforts throughout the company.
3:16:43
Um, and so the company has definitely been like betting on machine learning for a while.
3:16:49
Um, I think we are like probably ahead a lot of a lot of like the quant firms in terms of the how early we were on realizing deep learning was important.
3:16:59
Um, and I think we've done a pretty good job at like keeping up with and following the like L and craze like we were we partnered with OpenAI before CHBT. Um, wow.
3:17:08
Wild and Yeah, that's amazing. This was great.
3:17:14
Let's make it a regular thing. Yeah. Yeah, this is fantastic.
3:17:17
Yeah, I I definitely want to have you back when there's a new model release or something or new paper publish.
3:17:20
Uh I'm excited to follow.
3:17:23
If I can do a quick plug, um in June, uh I will be at the A Engineer uh World's Fair in SF giving a talk as a follow-up to my previous one. Very cool.
3:17:33
I think we're going to see you there. Oh, sweet. Awesome. See you there. Yeah.
3:17:36
Um and then also I am doing a course of some kind soon.
3:17:38
Official announcement pending, but you really want to like how to get rich quickly. Uh agents and RL stuff. Okay.
3:17:46
Um and so if you DM me your email, I'll put you on the list for more info. That sounds awesome. Awesome. Good luck with it. Uh I'm excited. Cool.
3:17:55
Yeah, this has been great. Yeah, thanks so much. Talk to you. Thanks for coming on. See you. We'll talk to you later. Bye.
3:18:00
Uh should we go through some timeline and then get out of here?
3:18:04
Yeah, you're a timeline addict.
3:18:05
Um there's I mean there's so many posts and we don't have enough time.
3:18:10
There's so much in the Monday is always a stacked day because you have a whole weekend of posts to catch up on.
3:18:17
Uh Adam Morin says, "You're coding at the bar.
3:18:19
I'm drunk at the office." Respect to that. I love that.
3:18:21
Um Aiden says, "The most pe the people I'm intellectually respect the most have quite lopsided output input ratio.
3:18:28
They write, build, create more than they read, study, or absorb.
3:18:31
Geniuses are not sponges. They're volcanoes." I like that framing. That's interesting.
3:18:38
Uh, I'm I'm kind of going backwards.
3:18:38
I like this one from Telmudic.
3:18:40
The first time I heard of the GoMad, gallon of milk a day diet, I laughed my I laughed.
3:18:48
There's no way I'm having my milked intake for a diet.
3:18:50
Goad, what are you cutting?
3:18:54
The idea that he's doing two gallons of milk a day. Hilarious to me. Timeline and turmoil.
3:19:00
Timeline and turmoil to uh this weekend.
3:19:02
Uh Paul Graham taking shots of Palunteer saying you shouldn't work at Palunteer.
3:19:07
Uh Gary Tan chimes in and says, "Is now an awkward time to mention I helped come up with the very first Save the Save the Shire t-shirt at Palunteer?" Uh very very funny.
3:19:18
Uh you know, we love we love Palunteer.
3:19:21
We love PG's writing and he is a foundational member of uh the the tech elite.
3:19:27
Um but he gets he gets spicy with the political takes.
3:19:30
He has strong political opinions and he brings them to the timeline and uh well I like that uh Barry and PG can have a little bit of fun. Yeah, exactly. Love to see it.
3:19:40
Uh WebDev Mason always with some great takes.
3:19:44
There's uh there's some people uh the Blue Origin story has really grown since we first covered it.
3:19:48
Now people are very upset that Katy Perry went said it was an affront to real astronauts.
3:19:55
They didn't go all the way to space.
3:19:57
They only went to the Carmen line.
3:19:57
I saw people saying uh that, you know, how could you go to space when there's problems on Earth? Yes.
3:20:04
Then you could also kind of extend that out to how could you do anything problems on Earth? Exly.
3:20:10
I think that's so funny because like obviously it's like why are you wasting money on space when there's people that are hungry?
3:20:14
But we literally talked to a farmer who was like Starlink is increasing farming yields.
3:20:18
It's like no actually like going to space and spending on like the crazy thing actually help people eat which is great.
3:20:27
Uh, but webdev Mason says, "Uh, this is so sad. Dead culture stuff.
3:20:30
Eject me into the timeline where she yanks the microphones in and shouts," I flew over our home world and saw us, the pale blue dot, the blessed sprouting seed of the Virgo supercluster, and I must report to every living soul that it is dope.
3:20:43
Uh, so she wants she because the story is that Katy Perry now regrets going and says that there's been too much backlash.
3:20:52
She says that she shouldn't have gone, but Mason says, "No, own it.
3:20:56
go and celebrate it because it is fantastic.
3:20:59
Uh question about watches.
3:20:59
Uh we already talked about bezel, but is it appropriate to wear a Rolex GMT as an analyst? I love this answer. It says it's it depends.
3:21:09
Imagine your first allnighter.
3:21:11
There you are burning the midnight oil when you decide to take off your Rolex GMT and place it on your desk to allow your wrist better control over the almighty Excel shortcuts you are about to employ.
3:21:21
You take a look around and what do you see?
3:21:23
Every other analyst has placed their PC Philippe Calatraa travel time in front of them for the exact same reason.
3:21:30
Now if you sir are capable of bearing the overwhelming feeling of shame that will inevitably conquer you then by all means consider it absolutely appropriate. Hilarious. What a great post. That's hilarious.
3:21:46
Anyway, um we should cover the VCs, uh the what's going on in higher ed at some point, but there was a good post by Connor.
3:21:54
He says, "Vs are attacking higher education.
3:21:56
Trump squeezes higher ed funding.
3:21:58
Universities sell PE holdings to fund their operations. VC funding dries up."
3:22:03
I don't think this is actually going to happen.
3:22:06
Uh what the the news is that Yale is like selling off a portion of their venture capital portfolio, but I doubt that that really has affected funding landscape.
3:22:15
And I don't think I don't think the university endowments are a major major source.
3:22:19
I think like pension funds are even bigger now and and uh and like uh sovereign wealth funds are even bigger. Yeah.
3:22:25
And also very likely again uh I doubt they're reacting to shortterm pressure.
3:22:32
It's probably, you know, uh, Yale selling $6 billion of secondaries probably is part of a larger strategy to generate liquidity on long duration investments, right?
3:22:47
So, here are some posts that I want to follow up on.
3:22:49
We'll highlight them today, but we'll either have these folks on the show or do deep dives on these topics.
3:22:54
Uh, person of swag, Adam, uh, says, "Vibe sheeting, is this anything?"
3:22:59
Uh, he's built cursor for Microsoft Excel.
3:23:01
uh going into the lion's den competing directly with Microsoft Copilot.
3:23:06
But we just heard it that some of the Microsoft Copilot products are falling behind a little bit and so I thought it was fun that he was uh building a like a plugin just into uh Excel and I could imagine this becoming a great company.
3:23:19
So excited to talk to him about that and dive in deeper.
3:23:21
Uh also uh there's a new planet Delian mentioned this K218B.
3:23:25
I did a whole uh deep research report on K218B.
3:23:28
Uh very very fascinating. Uh interesting. I had science.
3:23:35
Uh I had Chachi tell me an entire speculative science fiction story about how we might get to K218b.
3:23:42
Spoiler alert, it's going to be like 500 years to get there even at like.
3:23:47
1 C or something like that, which is the speed of light because it's so far away.
3:23:51
Uh but a multigenerational math to understand how many episodes we could do in 500. Quite a lot.
3:23:58
Interestingly, I had chat GBT.
3:23:58
I was like, "Uh, tell me a sci-fi story."
3:23:59
And like, "Why don't you just make me the character in it?" And it was very weird.
3:24:05
Uh, and it was like, "Now you're in cryosleep for a hundred years."
3:24:07
Uh, it was like, "Now you're an old man, but because of uh life extension technologies, you're 150 years old and like uh like all like your kids are now older than you because of this weird time thing."
3:24:20
It was very fun, very weird.
3:24:22
Not appropriate for the show. Anyway, get on.
3:24:24
Uh, last one is uh uh Notable Cap.
3:24:28
This is a leak from Arthur Rock.
3:24:30
Arthur Rock uh series B uh to uh browser base, right?
3:24:35
Paul Klein who's been on the show.
3:24:36
So uh I assume one last one last one uh from Near. I like this one. Very end.
3:24:41
Reminder of how far AGI goalposts have moved.
3:24:43
It's from an old book or something.
3:24:46
It says an AGI could beat you a chess, tell you a story, bake you a cake, describe a sheep, and name three things larger than a lobster.
3:24:55
It's also solidly the stuff of science fiction and most experts agree that AGI is many decades away from becoming reality if it will become reality at all.
3:25:04
So wow the last three models I use could describe a sheep.
3:25:10
Couldn't bake the cake though.
3:25:10
We need the humanoids for that.
3:25:11
So but it can tell you how to bake a cake and it really can do all of those things and more. So yeah, AGI is here.
3:25:16
It's just get on with your life. Unevenly distributed. Yeah, that makes sense.
3:25:22
I was at the beach over the weekend and I was thinking to myself, I was looking around.
3:25:27
I was like, none of these people are AGI pills.
3:25:31
And then everybody went back to enjoying the beach.
3:25:33
Yeah, that's the nature of it.
3:25:35
Anyways, thank you for watching today.
3:25:36
Uh we will see you tomorrow.
3:25:38
We got a great show for you.
3:25:40
Bunch of news breaking and we'll talk to you then. Bye. Looking forward to it. Cheers.