SPEAKER_02: i'm so tired i've slept six hours in three and a half days six hours so wait wait seven hours to the world series of poker with hellmuth so you spent no no first of all i flew public took southwest what yeah cost me you were on a southwest flight yeah because i got a ticket for

SPEAKER_05: 49 bucks it's like so incredible did you sit in seat a1 also known as jcals reserve seat

SPEAKER_02: it has my name on it well it's in the front row yeah you it's great like southwest has these

SPEAKER_09: numbers and what they do is first class on southwest no no no premium plus he was in the front row no

SPEAKER_10: there's like these signs that that like have a number that attaches to your ticket and then you stand in that line and then you go on in this orderly way and so i was i had like a5 or something so i was like the fifth person on and then i sat in the front and i put my bags up top yeah you carried your bags yeah an hour later i was in vegas it was so easy southwest is phenomenal and then i flew back anyways and after after losing as much money as i did it felt good to fly for 49 i got to be honest

SPEAKER_11: with you well austerity measures have their benefits you can sleep better at night with austerity

SPEAKER_12: measures and did you go to the all you can eat buffet and take it to go as well no i never do that you

SPEAKER_25: never know why did you fly southwest why did i fly southwest well the planes in europe was not that's

SPEAKER_00: number one and then two i just wanted some flexibility to get in and out depending on when i busted these tournaments so let me tell you about these tournaments the 100k is literally like a

SPEAKER_10: murderer's row of like every great poker pro so it was like 96 of us or something i got to be honest

SPEAKER_00: with you it was so much fun so much fun why how so you know we play 40 minute levels and you have to really get the chips moving which means that there's only so much like you know game theory optimal poker you can play and at some point you just got to gamble it up and you got to take your shots and you have to be willing to bluff and you know you have to bet for thin equity it's so much fun and

SPEAKER_10: then you see some of these guys that just lose their minds anyways it was it was really really a lot

SPEAKER_00: of fun i finished in the middle of the pack like 42nd or something so that was brutal but then the thing that i'm the most proud of is and i went from that and i hopped in the 3k six six handed right

SPEAKER_10: afterwards and there was like 1250 people in that and i got to 81st and whoa the problem was i just couldn't get anything going i couldn't get any real car like you need some card distribution like you

SPEAKER_00: have to at some point yes make some hands you can't just you know i could bluff my way to 81st but i needed hands to really to really really have a chance to chip up and run it then i played in the raz which is seven card stud low there's 126 people now when i finished in the middle of the pack there and then this 10k bounty there's about 600 people and i finished like 150 those are all respectable i mean i know play tournaments i have no idea how to play tournament to be honest i know how to play poker i know how to play cash games but tournaments are very different so i feel like i was very ill prepared but i had so much fun and you know the crazy thing is when you're playing for 12 hours a day you are focused for 12 hours the thing that i forgot is that you end up i ended up losing like two and a half pounds oh wow if your brain is going going going eight hours a day biggest thing and i was i was mentally devastated at the end of each night you're just lying in bed eyes awake yep and then you can't sleep and so then you know you go and you play crabs and buck

SPEAKER_39: break the casino for a couple hundred times a night there and then go back to bed that's great

SPEAKER_44: sounds like a great week the good news is uh is phil was live blogging for you on your behalf from uh

SPEAKER_00: from over your shoulder oh by the way the other thing was i brought my phone but i don't have anything on my phone like i don't have twitter i don't have any of that stuff so i was just totally in the zone it felt so great to not be connected to social flow experience it's a flow experience when

SPEAKER_48: you can turn everything off and just be in the moment congratulations on the on the you know coming in

SPEAKER_49: the 10 of the field in the 3000 or 8 of the field that's incredible dude if i really practice at

SPEAKER_53: these tournaments i think i would bank a couple but i just don't have the time the paradox there is

SPEAKER_48: that you did best in the largest field if you can do this for three or four years and you go into those razz fields when it's a hundred or those 25k 100ks when it's 100 200 people in the 100k there's

SPEAKER_00: 90 95 96 people like you really do have about a one percent shot of winning the thing which is pretty

SPEAKER_60: incredible yeah razz too razz is always 100. i love tournament poker that was my that was my game

SPEAKER_62: back in the day it's because it's because you're afraid to lose money so it's perfect for you put up

SPEAKER_25: a small amount of money and you can play for many hours you're probably in your mind you're probably dividing the number of hours by your buy-in so that you can think about what your hourly cost is

SPEAKER_65: i actually like tournament poker because when you lose you lose like you can't just yeah rebuy and

SPEAKER_68: basically and then there's a broader macro strategy you start to play the chips against the other

SPEAKER_63: players because you know how everyone else is playing based on how many chips they have in

SPEAKER_10: front of them it's a great point this theory of icm which is like independent chip model is exactly what you just said the craziest thing freeberg happens right at the money bubble these guys start i've never seen this before they ultra tank ultra tank oh yeah they go in the tank for two

SPEAKER_02: minutes and then they fold so obnoxious when i used to play tournament poker i'd sit i'd fold ace king

SPEAKER_68: suited you'd fold pocket queens you'd fold everything if you're on the bubble of getting into the money or you're on the bubble of getting to the final table it just doesn't matter because if there's three guys with you that have smaller chip stacks let them get blinded off and now you're in the money or now you're on the final table and then you'll play poker it's like go get up go have some dinner come back in an hour don't even look at your cards if you can cruise your way there it's a

SPEAKER_63: totally different kind of game totally different there's a guy that was tanking he's a really

SPEAKER_00: lovely guy a british guy i offered him the min cash just to stop tank i was like please stop

SPEAKER_63: tanking for four minutes and folding around i'll just need it as action i had a guy do this in a

SPEAKER_78: tournament and i timed it with the rest of the table i took out my phone i timed it and i said whatever number of seconds you do i'm doubling and i just did that it was so obnoxious that the

SPEAKER_02: entire table like the guy finally stopped well in the in the 100k we actually had a better setup which is that we had these ipads on the table and everybody gets a 30 second shot clock so you can't

SPEAKER_10: really around but then in these big field events they're not going to have ipads on 100 tables obviously yes and so instead people just tank forever and then you have to call the floor and you have to call time on these folks the funniest thing though was like the last tournament i bought in for was a 10k secret bounty which means that on day two when you knock somebody out you actually everybody stops the action you go up on stage and you pull like out of it's like treasure hunting yeah and some of these bounties can be like a million bucks and obviously there's a lot of people there that kind of recognize me and when they saw me in the 10k secret bounty there was like

SPEAKER_43: a movement it was so funny they were like if you win the secret bounty and you win the million we're

SPEAKER_84: going to revolt and burn the room and i thought you know it's true it'd be patently unfair if i was

SPEAKER_89: the one that won that million dollar secret bounty how many selfies did you take on southwest that's the question everybody wants to know how many people recognized you on southwest and what was

SPEAKER_11: the shame that you felt it was great i'm gonna fly i'm gonna fly southwest to vegas all the time

SPEAKER_10: now it's great it literally was honestly to get from my house to the gate where you stand to get into

SPEAKER_43: the plane it was like plus or maybe 15 minutes more than just driving on the tarmac like it was a

SPEAKER_92: nothing burger jaco it's amazing what a down market could do to people you know what this is what is

SPEAKER_93: it brad this is the victory of the age of fitness chema absolutely chema dictators southwest you were

SPEAKER_96: in the first the first row you were in the front row southwest all right listen there's a big docket

SPEAKER_12: today we've got a lot on the plate david sachs is busy in a star chamber right now selecting the next president of the united states he's picking a vp and the cabinet for whoever's going to to win so he couldn't make it today so we brought the fifth bestie brad gerstner back and great

SPEAKER_48: timing because the fed decided to pause rate hikes in just yesterday i guess and this would have been the 11th consecutive rate hike is that about right sounds right and they anticipated two more 25 basis hikes before the end of the year so here's just the um fed funds effective rate so you can all can see it here over time and there was also some big news this week a couple of 2023 ipo candidates started to be floated obviously we've been hearing about reddit and stripe for a while but arm confidentially filed for an ipo back in april and they're seeking to raise between 8 and 10 billion later this year that would be uh a pretty big deal for softbank which owns the firm and that could i

SPEAKER_101: guess save softbank which has had a just a brutal run with the two different

SPEAKER_60: opportunity why would they say save softbank the arm deal is being done is what's the valuation it's

SPEAKER_59: being done at softbank acquired arm for 32 billion in 2016. they don't know the valuation right now they

SPEAKER_65: haven't put that out there yet i mean they were originally looking for anywhere from 30 to 70 but i mean

SPEAKER_106: brad where do you think that comes out i have no idea i think that you know they're shopping to get these anchor tenants into arm if the ipo market was super hot and this was really easy to get done this would have been done you know so i think you know they're tiptoeing out because they need to get it into the public market but you know i i think that anybody who does an ipo today okay is going to demand a you know a rate of return into that offering that is a significant margin of safety relative to all deals that were done in 20 and 21 and what that means is on the roadshow investment banks go and they try to drum up demand they want to have anchors in the ipo book they call firms like ours they call strategic investors they want to size up what that is and you know they give you their expectation of fair value or where they think the ipo will trade to of course you meet with management you ascertain

SPEAKER_96: and develop your own expectations of fair value what i'm saying is maybe in the peak you know in 20 and 21 people were getting thinner and thinner in terms of the margin of safety they demand i my expectations whether it's arm whether it's databricks or other companies that are starting to queue up and these are fantastic companies right make no mistake about it nvidia tried to buy arm and you know

SPEAKER_106: databricks is doing over a billion in revenue growing extraordinarily fast led by an incredible team

SPEAKER_96: but in order to get into the public markets they have to you know step in with a bigger discount than

SPEAKER_48: they would have you know pre this correction and uh just to put some numbers on it arm had agreed to be acquired by nvidia for 40 billion before the talks fell through that was due to regulatory scrutiny

SPEAKER_101: bloomberg says arm is aiming for evaluation of 70 billion obviously there's no confirmation about that

SPEAKER_111: that's just the news hopefully intelligently speculating about it what exactly would you be

SPEAKER_00: buying for 70 billion dollars yeah i thought they missed it yeah arm had a window where they had to make some really important product decisions to be able to actually have reference designs that actually made sense beyond just simple mobile processors and cpus and they don't have any of that so they have a good cash cow business it's probably worth 25 billion they paid 30 some odd 32 yeah they could have gotten it out for 40 something but it's a mid 20s billion dollar company no more and if you're and if you buy it at the 20s you're buying a cash flow yielding asset and the good luck to all

SPEAKER_118: the players i guess is what i would say clearly it's a function of soft bank seeking liquidity right i mean

SPEAKER_60: they uh they've had this position for how many years now four years no more i think this was six maybe

SPEAKER_121: six years something like that yeah when did they buy it was it 2016 i think it was longer yeah 2016

SPEAKER_60: they bought it so this has been and in the fiscal year ending march the vision fund alone took a 32 billion dollar write down obviously soft bank has a big chunk of the vision fund but this is not in

SPEAKER_128: the vision and it's not in the vision fund exactly it's on balance sheet yeah but it's clearly a liquidity driver for them i don't think that this is you know first of all jason i don't think the

SPEAKER_105: point about soft bank needs saving is necessarily true it's not a company that needs saving but

SPEAKER_130: they've certainly taken some hits well they could use a win this is what my point yeah yeah they could

SPEAKER_131: use a win but the last two uh quarterly calls you know masayoshi-san was pretty repentant and

SPEAKER_60: humble right in fact he used that exact language we talked about it here so there's a big liquidity need at some point for the business so they can get this thing out and get liquid on it it would certainly support the balance sheet more than anything you know the the market is on a tear this

SPEAKER_48: year brad is there now that everybody's buying back into tech stocks i guess two questions why is everybody suddenly buying into tech stocks when people think there's going to be a recession and consumers have run out of money and yeah what what's the thinking here of why so much money has

SPEAKER_101: gone to tech and this incredible rally it's been great to watch but uh people seem to be perplexed by the rally and then the fed even though they didn't do a rate hike seem to put cold water on and say hey listen don't get ahead of yourselves we're gonna probably do rate hikes later in the year we all

SPEAKER_135: have short memories you know tech had a devastating year in 2022 so this is a bit of a reversion to the

SPEAKER_106: mean right if you remember the chart we you know we showed a lot on this pod last year at the at the end of starting at the end of 2021 was that we were trading at you know multiples that were you know 50 to 70 percent above their 10-year average um and at the trough we were trading about 30 to 40 percent

SPEAKER_96: below the 10-year average right so the market corrected it overshoots on the upside because rates went to zero and by the end of 2022 when you had larry summers and others who were going out and saying we don't know what the upper bound of inflation and rates is that's really scary to the markets and so you had this overshoot to the downside we're still trading below the 10-year average for internet and software companies based upon our numbers and so when you look at this as we said at the end of last year the framework changed early in this year all of 22 was about two things is inflation going higher and our rates going higher and those things are debilitating for growth assets and growth multiples by the beginning of this year we started to have confidence that inflation had peaked which meant that rates were largely in kind of you know spitting distance of their final destination and the framework shift to economy are we going to have a hard landing now remember at the start of the year mike wilson from morgan stanley who made give him credit made the call in 2022 he said 22 was going to be awful he was right but he stayed in that bearish position at the beginning of 2021 said the economy is going to hit the skids in q1 and that we're going to revisit 3 000 or 3 200 on the s p instead we're sitting at 4 300. so he was tragically wrong at the beginning of this year because again he was fighting i think the last battle that he that was won as opposed to looking ahead and saying this is no longer about rates and inflation so the nasdaq has moved 30 percent to start this year you know the fed said yesterday um we're hitting pause um you know the quote uh from chairman powell that has everybody a little bit perplexed is on the one hand he says we're data dependent on the other hand he said we're likely to have more rate hikes right and that rates are going to stay high for a couple years well if we're data you can't have it both ways if we're data dependent we have a hard landing like stan druckenmiller thinks we're going to have in q4 this year then rates are going to get caught but the market seems to like this and let's talk about where we are the 10 years at about 3 7 the s p is basically flat in terms of where it was and where it was before and where it was after the rate hike announcement the economy appears to be reasonably good 80 percent of earnings in q1 b the guides were okay stan druckenmiller said i shifted you know he pulled forward when he thought we were going to have a hard landing based upon the bank crisis which now feels like it's a distant memory for many and he's now calling for potentially a hard landing in q4 so the fed just came out and said no we're not going to have a hard landing in q4 we think we're going to have one percent growth they think that inflation to end the year is going to be at 3.2 percent versus the 3.7 percent goldman consensus what i would say is you know we've moved but we're still below the 10-year average but you have to start paying attention now to individual stocks that have likely gotten ahead of themselves or where they've taken what's an example of that a lot of the return facebook apple no you know i think a lot of the ai related stocks are not at or within 10 percent of our end of year price target and when you when you see that take something like nvidia if i'm playing at home or i'm a you know a professional investor i may say oh i'm going to go sell some long-dated calls to buy a little protection to the downside i don't think there's any problems with nvidia i think they're going to continue to perform i think they're beat their numbers for the balance of the year but you know remember at the start of the year people thought they were going to miss their numbers for data center people thought data center revenues this year were going to be negative and they just revised their guide from 7 billion in q2 to to 11 billion in q2 so they've absolutely blistered their numbers and the stock's gone from 125 you know to over 400 so when you have those parabolic moves just like we did with rates last year i think it's wise to say it's a fantastic company we want to be involved it's going to continue to be one of our larger positions but if i can lock in 20 yield by selling calls six months out i think that's a reasonable risk reward so you know i think that's where we are and now now the question really becomes and it's unknown and unknowable right are we going to have a soft landing a medium landing a hard landing in q4 and rolling into next year and i would say as an investor i'm very data dependent

SPEAKER_143: you know we're talking to companies we're looking data point should we all be looking at at this point

SPEAKER_121: chamath you think jobs unemployment just these stocks and their multiples and how are you looking

SPEAKER_144: at it as a capital alligator tomorrow still on the sidelines what have i said like a broken record rates are going to be higher than you want and they're going to be around for longer than you like

SPEAKER_114: and now powell is basically telling you the same thing so we're almost at the end of i think the bottoming though i don't agree with drunkenmiller i think he's wrong on which part that doesn't be a hard landing in q4 yeah yeah and and the reason there's not going to be a hard landing is you just saw china today basically say we're going to start to rip in trillions of dollars they're going to

SPEAKER_115: stimulate the economy you can't have a hard landing when china's printing trillions of dollars

SPEAKER_00: not possible so i think that what powell was forecasting is that if china starts to basically

SPEAKER_114: turn on the money printer and go through a huge spate of quantitative easing it's going to just inflate everything because they're just such a critical artery to the world

SPEAKER_58: economy and so you just have to get prepared for rates just being sticky and inflation being sticky and i think that that's probably the most reasonable base case for the rest of the decade rest of the decade seven more years yeah yeah and so in that environment the problem is that now you have the seven or eight most valuable tech stocks priced to perfection yet again if you look at their enterprise

SPEAKER_00: value over their net income these things are trading at astronomical yields that are less than half

SPEAKER_10: of the two-year note and now approaching sometimes less than half of the 10-year note government bonds that makes no sense and if you subtract out those seven or eight biggest companies in the s p 500 the s p has not been a great asset so the i think the equal weighted index brad you probably know the exact number the equal weight index is just so what does all this mean i think it means that people are psychologically exhausted with having lost money they are psychologically wanting to will the market up they want to psychologically believe whatever will allow them to influence rates getting cut but i think the most reasonable bear cases rates aren't getting cut and whatever hope you had of

SPEAKER_58: rates getting cut just went out the window today because there is no world in which you cut rates when china's going to print a trillion dollars or more freeberg a lot of people seem to think

SPEAKER_48: that this big run-up in the markets is ai related everybody's got an ai angle for their company some

SPEAKER_101: of them super valid like microsoft and google other ones feels kind of speculative we talked about buzzfeed talked about having ai journalists in their stock which looks like it's going to be delisted got a quick

SPEAKER_48: pop do you think the ai ai actually coming to market driving efficiency driving revenue for companies

SPEAKER_98: is going to happen in the short to midterm or is it more like midterm to long term in other words is the ai rally overheated and fake i don't know if i would assume that the pricing of equities is

SPEAKER_60: entirely driven by ai i think there's a lot of factors including a lot of folks maybe being on the

SPEAKER_118: sidelines for too long and needing to come and buy in at the same time but i do think that there's a

SPEAKER_60: radical realization underway that a lot of businesses that are dependent on services that might be replaced by ai don't necessarily have the longevity today those are businesses that have been severely hurt by the point of view of what ai could bring to market and bring to industry so there's a lot of

SPEAKER_68: stuff happening on the short side on the long side i think it's a distribution curve on like how quickly different industries will be affected in a positive way by ai certainly the most obvious is demand for chips because all this infrastructure is being built out so that's the first and so you know with a high discount rate which is the environment we're in today you can more quickly bet on those opportunities and so you see a disproportionate bubbling in valuation and multiples on businesses like that there are other businesses that are further down like services businesses how long until lawyers are able to leverage ai how long until banks investment banks are going to be leveraging ai to create new products and improve their margins there's a lot of speculation a lot of ideas but the discount rate is quite high right now and that's probably a few years out to the point that the fuzziness wipes out the potential

SPEAKER_162: value what you mean by the discount rate in this context so let's say that i think that investment banks are

SPEAKER_68: going to be severely changed in seven to ten years that's the time horizon i think that their business

SPEAKER_118: is going to improve because of ai as an example or you know factory machinery automation right

SPEAKER_68: manufacturing businesses probably 10 to 15 years so i can't really give them that much credit today and so i i apply a higher discount rate there but there are other industries that are certainly being affected much more quickly and you know i'm sure brad and his his team of of finance wizards probably have a board you know where they have this list of all the industries are going to be affected in what time horizon i mean that's the way you could kind of think about this and over different time horizons you could kind of accrue some net improvement to earnings and say okay you apply a discount rate to

SPEAKER_118: that here's how much the stock should trade up and and you know that's one way to kind of think about

SPEAKER_60: this i don't think it's everything everywhere all at once great drop uh great reference brad great by the way i i gotta tell you i just saw that movie i have to reference it because i saw it over the

SPEAKER_68: weekend have you seen that movie of course of course yeah what we should take what an unbelievable film

SPEAKER_118: i had not seen that film i had kind of postponing it incredible anyway uh what is it what movie it's uh

SPEAKER_24: everything everywhere all at once it won best picture at the oscars last year i don't trust the best

SPEAKER_25: picture awards at the oscars for a few years i've just been duped by that award like you get some

SPEAKER_63: really crap ass movies that are like uh watch this movie and then report back let's i won't even say

SPEAKER_169: anymore politically correct or whatever and then yeah there's a little virtue signaling they were

SPEAKER_171: doing a little catch up because of oscar so white the movement to try to balance things out in fact

SPEAKER_144: people well that doesn't exactly reinforce trust that actually destroys complete trust and reputational

SPEAKER_101: credibility yes they're uh virtues and actually that was something people said that everything everywhere all at once was possibly uh because it's an asian cast i don't even know what is it what is

SPEAKER_98: it it's a sci-fi adventure surrealist kind of romp michelle yo is in it uh she's very famous from

SPEAKER_78: crouching tiger hidden dragon correct there you go exactly kwan is in it who played short round in indiana jones and the temple of dones yep and i don't do my impersonation but i would get cancelled for doing

SPEAKER_179: the impersonation of him in that that guy must be like 70 years old yeah yeah no he's like he's like

SPEAKER_177: 50 or something but it's pretty great because he also won best actor best supporting actor and there

SPEAKER_78: are all these pictures of him with harrison ford and steven spielberg the guys cast him the guys that made it

SPEAKER_68: are like music video directors you got to go see this film it's really good it's it's on a visual basis you have to go to a theater you have to go to a theater just watch it on apple tv in one of

SPEAKER_187: your theaters tonight go to your main theaters go to the theater on your plane and watch it tonight

SPEAKER_179: i actually will do that i'll do that this is the best moment i ever had i go to the bathroom you know it's a long flight back from italy so i'm taking my time and i see that as sub as sub soap that i can't afford it's like 100 yeah that one so i see that and i'm like wait a second so i look

SPEAKER_78: in the cabinet underneath i'm like if there's two up here there's got to be six down below so i check in the galley of course there's six down there i put two i boost two on my back which i'm still not through in my uh bathroom here that's about 300 bucks then i go and i check out he's got a tv set up so i'm like oh i wonder what movies are on here and it's netflix he's got so much bandwidth on that freaking plane that he can stream on the plane you don't have to download your movies before you take

SPEAKER_89: off on shama's plane all right let's get back to work here have you seen it by the way have you seen

SPEAKER_203: that movie yeah it's incredible movie incredible he's got to watch it it's not the best film of the

SPEAKER_204: that was tar but i don't want to get into it with chamath oh my god i tried to watch tar

SPEAKER_30: that movie sucks balls it was so boring and self-absorbed and controlled it is a little self-absorbed i'll admit but that's what i like about it for anybody that has sleeping issues and otherwise needs to use a sleeping aid i would just put that movie on because you know once you see cape

SPEAKER_10: blanchett blathering on and walking back and forth from her kids school and then some crappy piano

SPEAKER_212: play you'll fall asleep and then what puts you to bed faster chamath ambient tar or science corner

SPEAKER_05: science corner which one i love i actually love science corner i don't take i don't take sleep

SPEAKER_10: aids i've taken ambien three times in my life yeah but that was recreationally during the day

SPEAKER_02: that's no no it was prescribed to me because i wanted to try it i also tried lunesta once they were they're both like be careful with that stuff man you'll start tweeting weird be careful i tried those sleeping aids four times in my life and you wake up i wake up with a it has

SPEAKER_10: a really bad tail effect so like you wake up yeah i wake up super groggy i only get five or six hours of sleep with it so anyways i uh i do take melatonin and that's really good that works that's natural it does the trick does the trick for me and then i try to just everybody on melatonin all of a sudden

SPEAKER_00: like everybody i go takes melatonin are you just a natural way to ease into sleep yeah for me what i do is like i use it as part of a routine where like typically around 10 like all the you know devices

SPEAKER_39: things get silenced and do not disturb mode and i'm in a really restful place by 10 o'clock and then usually by 10 15 to 10 30 i'm out like a light man i just got this infrared sauna you guys have

SPEAKER_98: infrared saunas this is like the new hotness and uh man that is incredible uh you get a heart rate goes way up and you sleep like a baby okay let's get back to ai brad are people getting too much

SPEAKER_226: credit in their stock prices and is the contrast therapy after the infrared sauna i don't know what

SPEAKER_05: contrast therapy is what is that meaning when you do the hot then you jump into like an ice bath no i'm getting an ice bath of course it is next for my little outdoor sauna area and i'm going to start doing

SPEAKER_98: that i did that with my friend antonio a friend of the pot antonio gracias he's got one of those cold plunges we jump in the cold plunge together he's got this thing at 45 freaking degrees he says don't try to do too much you'll have a heart attack so i go in i get 45 seconds in i start breathing like

SPEAKER_230: really heavily and i'm like i gotta get the hell out of here he does six minutes the man is an animal

SPEAKER_30: i do four minutes three times a week what temperature what temperature 42 to 44 the research paper

SPEAKER_39: that everybody references now i think her name is suzanne's soderberg or something like that she's

SPEAKER_00: the swedish researcher huberman did it on his podcast that rogan did it anyways the the data is like an hour like a little less than an hour of heat and 11 minutes of cold per week per week was shown to be physiologically optimal so i do it three times a week 20 minutes of heat followed by four minutes of of cold plunge three times a week do you have like a cold plunge where you set the dial

SPEAKER_57: on it or is it no no i just cold water in no i do old school i uh i use the mr steam in my bathroom

SPEAKER_00: because i after i'm done working out the last thing i want to do is schlep outside find my slippers put on a robe i'm not going to do any of that so i walk upstairs i do the mr steam at like 110 degrees and then while i'm working out somebody just makes an ice bath for me and put a bunch of ice in the tub fill it with water and just jump in it so that what do they do do they carve like a bunch of

SPEAKER_89: baby seals in ice and then they just when you get no no push all the ice cubes were made to look like

SPEAKER_240: my face so i have these shapes it's like a bust of yourself so i'm like looking at myself yeah no

SPEAKER_54: infrared gets up to like 130 140 degrees and your heart rate will go up to 140 150 degrees in this it's a

SPEAKER_44: whole different experience than the mr i was recently on a ski trip yeah and it's a this extraordinary

SPEAKER_106: sauna it's wood fired okay you told me about this so that so the lead guide on this trip we go down there he gets the thing heated up before we show up right and i show up with this great athlete down there and we go in the hot box because you know this is the thing to do yeah and i look at the

SPEAKER_96: thermostat i think it's broken it shows like it's hard pegged at like 225 degrees fahrenheit that's a lot you know you walk in you're i mean you slow cook meat at 225 my lips are burning you know and there's this person sitting in there he's like now this is the way you do it right tough guys you know do it at this temperature we're like all right so you know we sit in there for like 10 10 minutes and then we go and we jump in you know in a frozen lake and about 10 minutes later we're taking off we had to leave this particular place and you know people are waving goodbye and they're looking you know they have to look over the sauna because they're waving goodbye and they notice that the

SPEAKER_248: sauna is on fire ah yeah that's that's basically what happens when you go to 220

SPEAKER_251: people are taking this a bit to the excess it's a little crazy all right so back to back to the docket

SPEAKER_12: here ai yeah are people getting too much credit in public markets kind of running with this before it's actually ready for prime time just to add to freeberg's point here uh there are a lot of people talking about doing things in startups i see people actually doing things i would say four out of five startups uh that we've invested in and the majority of them that we're being pitched on right now are

SPEAKER_98: not only talking about ai which they were all talking about it three years ago they're actually implementing it now because the tool sets are available i just had a call thank you freeberg for putting me in touch with the google people i just did a call this past week with google and they gave me some previews of what they're working on i can't say anything about it but it's it's mind-blowing i think google's not as far behind closed ai as you think what's your read on google now we had a little

SPEAKER_118: kerfuffle when i when i talked about sundar's reaction to you know ai and the concerns that others might have on folks talking about the company i mean what's your point of view on google today i don't

SPEAKER_106: know if you talked about this yeah yeah no well i mean i think you know i had this i had this dinner this incredible dinner actually thursday at the pub and you know it's it's how i know silicon valley's back because it was just like serendipitous we threw it together in the afternoon you had this like four extraordinary you know ceos there and you know you had mustafa one of the co-founders deep mind there and rich barton from expedience oh just just an incredible group and the 20 trillion dollar question is how does the entire open architecture of the web get re-architected in the age of ai another way of asking the question because we're all kind of playing small ball right we're all

SPEAKER_96: talking about oh you know our tpus and gpus you know and what cloud is going to get accelerated that's kind of the small ball the big ball here the 20 trillion dollar question that has defined the entire internet for our entire careers has been web search and i will tell you that the conversation around that table there's increasing confidence that whatever comes next right the top of the funnel is up for grabs and all of those dollars over a period again a five years seven years is going to get redistributed and make no mistake google is well positioned to make its claim for that but its claim is going to come at the expense of search and so google may create the best ai in the world but it's going to have to fight off the cannibalization of core search and the question is in this new thing and so rich barton you know an incredible um product founder right think about expedia zillow yes and zillow so he's founder of expedia and founder of zillow two vertical search engines that dominated right their particular verticals and the topic of the conversation was that the new ui right the new way you interact with your customers right is not optimizing a web page it's not optimizing an application he calls it the race to intimacy the race to intimacy that you

SPEAKER_263: have to build a conversational ui i've run that race a few times and no and it's a related race i think that was a sprint not a marathon sometimes it's been a marathon and the penultimate question

SPEAKER_00: who ends up on top yeah are you always on top jama this is i i run really good middle distances and then if i need to just make it you know run a boston marathon i can if i want but i don't oh okay

SPEAKER_271: yeah sure you can yeah but i think you know i took a poll does he think sorry does he think zillow is

SPEAKER_30: dead now i mean he's no longer sure what i'm guessing i mean he's the ceo he's the ceo is the

SPEAKER_00: chairman sorry sorry not zillow sorry not zillow uh expedia pardon me what does he think about

SPEAKER_96: expedient travel well you know you and i had this conversation girly chimed in the cash chimed in on twitter well you're not on twitter but they chimed in you know about the conversation from last week's pod every vertical search engine and google itself that architecture is suspect right the question is just over what timeline is it suspect but the world is moving past 10 blue links information

SPEAKER_274: retrieval that looked like a card catalog is not where we're going you no longer need an index you

SPEAKER_276: need someone to do the retrieval from the index for you that's the the new service and extract the

SPEAKER_68: knowledge and that's what i mean it's personalized to you but i think one of the questions that's still outstanding from an architectural point of view in terms of your interaction with all the data in the world is are you as a user going to have a number of independent relationships with your travel search engine your travel agent or are you going to have a relationship with your doctor medical office and you're going to have a separate relationship with your you know financial advisory type service or are you going to end up seeing similar to the challenge we have in search versus vertical search today an aggregator of services because there's value in cross-populating the data and the knowledge about you across those services so yeah sure you know it's going to be the no it's going to be the

SPEAKER_240: former because like obviously there's value to the company because every company wants to profiteer

SPEAKER_00: from you and monetize you in nine different ways and maximize the equity for their employees and their founders and their board and their shareholders but that's not what people want right so like for example when you go to a doctor would your doctor be better off if you know she also had all of your data from i don't know pick your other service providers of course maybe you could make a claim that if they had all of your personal information and they understood your risk factors they could do a

SPEAKER_10: much better job but the reason you don't do that is you want some segregated relationships a it gives you some level of privacy but mostly it gives you control in a world of ai where you can be more conversational the idea that humans will want to get reduced to talking to one thing for everything i think is wrong i think instead it's actually going to be hyper fragmented because the best in class use cases are going to be the things that people want and then b it gives the user control the one thing i would tell people is do not give up all of your data to one thing who's professing to be at all they are going to lie to you and trick you and you're going to have a social media problem

SPEAKER_68: writ large all over again there's another way to think about this which is the reverse of the client server model where today you as you are the client you are the node on the network and you're communicating with the center of the network which is the server which is the service provider and all the data sits there and you're getting data in and out of their product to suit your particular objectives but the future may be that more data is aggregating and accruing about you you end up becoming a server you then have the option just this speaks to kind of the architecture on the internet imagine if every individual had their own ip address and that individual's ip address uh had behind it behind your ability to control lots of information and lots of output about all your interactions across the different network of service providers that you use that's a very smart framework i think and that

SPEAKER_39: object and ultimately what would you agree you want to be able to rent that data to services

SPEAKER_71: i don't even think it's about rent as yeah as much as it is about you should have the reason the

SPEAKER_30: reason rent well that then you have to rent it you can't give it to them because that's be stupid

SPEAKER_68: that's right your goal your usage like let's say the doctor says hey can i can i get access to your health data or sorry your apple watch activity data for the last year you then have the option to provide that data to them through some permission type service that you then make available so there will be interconnectivity amongst the service providers but entirely gated and controlled by the

SPEAKER_10: users look you're seeing this up front now with reddit when in this world the people that create the content are actually in control of the data and if you try to like monetize an api without paying the people that created it or giving them control they'll revolt and you're seeing it happen now reddit lost what 90 some odd x percent of all of their content because the mods were like you to reddit well think of that writ large where everybody is asked hey i'm this great service hey i'm this great service i can solve x y and z problem for you just give me all your data and then if you're not asking well what do i get and they're like well you're going to get a whiz bang service and that's all they say you're being tricked it's already happening i mean there's an open source

SPEAKER_54: lawsuit against uh microsoft's github co-pilot and you have the getty lawsuit and you know it there's

SPEAKER_98: also a corollary for this google tried to kill xpedia already they tried to kill yelp and it didn't work folks use both services and you know if you if you are trying to boil the ocean like chat gpt might i think it's not going to work because there's going to be so much innovation so much community brand and all that kind of stuff and these llm seem to be trending towards good enough or some kind of parity that i think your rights from off people are just not going to give their data over wholesale i've been using the zillow plugin to get exact and the zillow plugin sucks on chat it's getting slightly

SPEAKER_12: better why is the wrong model it's going to live this will exist on zillow site zillow will not let

SPEAKER_78: chat gpt take their business you're going to go to zillow site and you're going to say hey show me homes that have a sauna and an ice plunge and that are four bedrooms and that are you know fixer uppers or are not fixer uppers and have ensuite bathrooms and whatever and it's just going to give you that list

SPEAKER_287: why would they give that business to closed ai can i take a crack just at summarizing what we all just said because your your initial question was friedberg was like what does this mean for google

SPEAKER_96: well we all just said that what we're certain of is there's not going to be 10 blue links and a model of search that they have monetized in a way that has been the most prolific business model in the history of capitalism right so that's a problem for them you know from a business model perspective so they're going to have to reinvent their business that's going to be hard to do it's like rebuilding the plane on

SPEAKER_106: the fly but then if you just go to the two models that we outlined here on the one hand chamas said you know you're going to work with all of these agents that's exactly what mark zuckerberg said on lex he's like you're not going to have one super agent you're going to have you know hundreds of agents and he he thinks of whatsapp and all of his platforms is the new open web where all of these agents will live and you'll interact with them you'll interact with your zillow agent there you'll interact with your booking agent your expedia agent there etc on the other hand chat gpt

SPEAKER_96: hi at inflection reid hoffman and and mustafa and many others think that the benefits of general intelligence this will be the first time that we solve vertical problems horizontally right and their metaphor the compare there is we all don't have a hundred personal assistants in in our office you have one that gets to know you really well they get to know what you like to wear how you like to travel the things you like to eat the things you like in your house how many kids you have so there's real leverage in that in that horizontal knowledge that can then be applied to these different verticals i think the answer lies in between your personal assistant will do a lot of the general stuff okay whether it's pi chat gpt whatever googs comes up with etc but i think they will subcontract the work whether they chain it out via you know behind the the scene you know mechanism to other agents so for example if you're interested in traveling to this undisclosed location that i'm at in in in europe right now you know my my my assistant doesn't know anything about this place so she may interact with a specialist for this particular place to get you that magical experience so you know i i don't think you have to pick one or the other but architecturally 20 trillion dollars of value on the web is built around web pages advertising and e-commerce right and sending traffic to those other places and what we're all saying unequivocally is it's moving to knowledge extraction and intelligent agents and i think

SPEAKER_98: that's tectonic but i still think there'll be clicks i you know i have been using bard a whole

SPEAKER_54: bunch and they have made massive rapid progress just to give you but one example here i'll share my screen it is pretty extraordinary how quickly they're figuring this out and i did this just in the search

SPEAKER_101: while we're talking about the five best greek restaurants uh in the bay area obviously came up with

SPEAKER_12: kokari and you know other ones and then i asked it like hey what are the top items on each of these menus they started putting images in and linking to yelp and then i said hey tell me the most expensive wines and it actually got that from kokari they have a chateau margot 2009 for 1500 chamath i don't

SPEAKER_98: think it's good enough for you but what year is it it looks like 2009 yeah they have screaming eagle

SPEAKER_12: 2013 i don't know if that's a no-go harlan 2014 no yeah so it's a no-go yeah but anyway oh harlan's good harlan's good okay okay good so we got to save anyway this i did the same exact search like four weeks ago it didn't have images and it couldn't get me the items on the menus and these could all be links and these could all be paid links so in these results there's nothing to stop google from saying hey yelp if you want we'll put your information here or not up to you you choose do robots.txt except we're going to call it ai.txt you tell us what you want to be included and if you want to be included we want you know it's it's a marketplace for clicks we'll include three of your images with clicks and we'll call it sponsored they are going to be able to insert ads into here that are better than the current ads and that will perform at a higher level and they're going to know us this could actually be

SPEAKER_05: the reverse of what everybody's thinking this could lead to higher cpms and higher cost per click because of intent by the way i mean the the 2014 is actually i think uh underrated and it's very

SPEAKER_114: highly rated but i think it is an excellent one all right that's an excellent um jason as to your

SPEAKER_106: point i think all of that i mean that's impressive it's true you can also do it on chat gpt i'll just say the number of people i interact with on a global basis who talk about chat gpt versus bard is like 10 to 1 today now google's got massive change real fast google's got massive distribution power let me tell you what i what i also think but to your point brad about this but i just want to i

SPEAKER_78: want to respond to your bard versus gpt4 that's people are not paying attention and when google puts this on the home page yeah and starts sending five percent of users to it i mean obviously it's expensive to do it people are going to have their eyes jump out of their head i think google's going to beat chat gpt4 i'm saying it right here right now i think they're going to beat them because i think that they're better at indexing all this information and understanding it than anybody on

SPEAKER_12: the planet and they have the largest ad network if they get this done in the next six months i think it's going to increase the cost per click because they're going to know so much about each user

SPEAKER_106: continue brad i know i'm in the minority here well i no no i i mean listen i think a lot of google's revenue comes from de facto navigation of the web they turn they turn the url into a navigation box there are a bunch of ads there that you know people click on they don't even know they're clicking on ads and you know they generate a lot of revenue off of that but let's be clear google is firing at this i just don't think it's going to be as monopolistic as they are in search i think there are going to be other competitors who are going to be well financed you're going to have access to the data and today you have well north of 100 million people who are paying to you know a huge percentage of those to use chat gpt and i think this you know it's the first competition but you know friedberg knows

SPEAKER_96: he and i had the back and forth they reported earnings stock was flat i said on cnbc that we had sold our shares we bought back some of the shares around google io because we do think they're going to be a player we think they're going to be a beneficiary but i i would say as i sit here today the distribution of potential for them is less than it was before chat gpt you know the the distribution of upside they have some competitors now who are going to be vying for this next new thing and i wonder whether or not you know as they try to navigate you know you're saying they're going to take some of this traffic from search and feed it into this other thing this other thing better monetize as well as search or by definition that means revenue goes down it feels like google is so

SPEAKER_101: close to figuring this out i mean i've been doing some barred searches how much google do you own how

SPEAKER_317: much google not enough not enough i've just been playing with the j on my jtrading.com he's thinking of adding this i'm like a buy-in or two but uh let's pivot over to this reddit thing because

SPEAKER_48: it's super important that people understand this i'm gonna buy 17 more shares i'm gonna buy like a

SPEAKER_98: harland 2014 equivalent in shares no my j trading is up right now jtrading.com shout out uh to my trades i did it as for fun i'm at like 20 returns versus six percent well jake you went kind of quiet for a

SPEAKER_328: while so when you're down you don't talk about it no i had all the stocks i liked and i didn't want

SPEAKER_294: to change them do you track them and exit or what do you do with i'm going to track the exits but

SPEAKER_330: i haven't exited any if you go to jtrading.com you can play along but how can you only be up 20

SPEAKER_96: when facebook has doubled oh sorry 22.43 yeah but facebook doubled off the bottom the whole goal of trading jacow is to maximize your highest conviction maximize your highest conviction what'd you do put a nickel do you put five percent into facebook i don't know i just i just bought

SPEAKER_334: companies as they went by in the group chat when you talked about it on group chat you trade companies

SPEAKER_337: you got to put a lot of chips on the table i mean okay that's i think what's the optimal number

SPEAKER_89: as a day trader brad what's the optimal number of names i should have you're not a day trader you don't

SPEAKER_339: trade as a long trader what's the optimal number of names i should be you have too many um i think

SPEAKER_96: with your level of insight um you know on the things that you really believe in there was such asymmetry in that position at the time that you bought it it was your best idea you should have put at least 30 of of whatever you were going to allocate to this into that position that's what i'll

SPEAKER_341: do yeah i mean it's done okay i mean listen i'm up what did i buy i bought uh 500 shares and uh my basis is 47 i'm at 140. so it was a good trade yeah now imagine if you could you know 10 or 20 exit

SPEAKER_89: yeah that's what i should have done yeah all right wait hold on how much have you put into all those stocks 1.5 million i i basically took i had some money in an index fund laying around and i was like

SPEAKER_12: let me do this five million i did j trading as a blatant attempt to get a sponsor for this week in

SPEAKER_317: startups so i was like maybe robin hood or e-trade will sponsor this and i'll make it like a segment

SPEAKER_98: so i was like let me do like a segment where i trade and see if i can secure the bag and it you know the markets are so crushed that nobody's spending advertising from robin hood or e-trade or interactive brokers or bloomberg but i also did because i wanted to become better at public market trading to your point to last week's conversation conversation is under understanding when to sell and when to go long when you get distributions as a vc that was really the reason i tried to do it was just to try to have skin in the game you know just like learning to play plo or you playing the razz tournament i just wanted to it's it's very it's very easy once

SPEAKER_39: you get shares distributed to you immediately sell them and then and then hold on yeah and then

SPEAKER_10: re-underwrite from scratch from there okay once the money is sitting in your account even if it takes a day a week a month it's not like the stock's gonna rip and triple on you over the next 30 days

SPEAKER_226: re-underwrite it from scratch and then see if you have that much conviction when you see the money in

SPEAKER_357: your bank account i like it i would just tell you guys the distributions you know there were a lot of

SPEAKER_106: firms that rode you know a lot of these big names right all the way down um you know that they got public but they didn't distribute much if any and it's really don't say the names but what do they rhyme

SPEAKER_362: with i'll i'll i'll start i'll start uh but what's moving up go ahead you go you go next you pick one

SPEAKER_364: not doing it um bliger bliger andresen borrow it's melounders land look at freeberg just dropped off

SPEAKER_367: he's like i don't want to be associated with this conversation all right listen we all got our asses

SPEAKER_334: kicked except for me uber's up 67 percent this year oh come on come on let me what what i was doing andresen blurowitz

SPEAKER_96: it's just right all of it what tends to happen is things hit the bottom they bounce a little bit and people are so relieved that they bounced a little bit did they immediately start distributing because lps are hammering them lps are like why did you not send us this you know snowflake when it was at four hundred dollars a share why did you not send us this doordash why did you you know go through any name you know that you can think of but then people compound the mistake i think that as soon as they get a little relief boom it's out the door and then the thing doubles on them and i just think you know part of the benefit of me having to get up every day 5 a.m deal with public markets and think about long or short it's what we're talking about in the fall of 2021 around the table before the poker games which is the public market is tilty it doesn't feel good and that you know we stopped making venture investments in october of 21 principally because of how we felt about public markets at that point in time and so i think a discipline in fact i was meeting with a managing uh partner of a very big venture firm in silicon valley this week they just started about a year ago or six months ago doing public market investing and they said it was incredibly valuable to how they

SPEAKER_376: think about distributions and what they're doing in the venture business that's what it's been for me

SPEAKER_12: i have now started to learn how these things are priced and i just think it's great as a learning

SPEAKER_48: thing for me i have to make two decisions one to distribute to my lps which i just give them the shares the second i have them and let them make the decision but that's also personally for me do i

SPEAKER_330: want to hold it or do i not i like your suggestions uh nice punch up chamath of looking at the cash

SPEAKER_240: and then deciding if you want to reality did you guys see this crazy thing from calpers where they were like yeah you know they were they were down like four i mean horrible returns on the venture

SPEAKER_00: side and so their decision was to double down well they didn't have they took like a decade off

SPEAKER_02: from venture the returns of what they did invest in venture were atrocious to be fair i don't know

SPEAKER_96: how the board has changed to calipers but calipers brought in a new cio an entire new team and so

SPEAKER_106: they can't be held liable for you know that track record i did kind of see something on twitter about i was shocked it would be very difficult to manufacture that battle track record if what i saw was accurate however i would say the team that is there now the portfolio that they're putting together and doubling down where i think venture is kind of at the bottom or bottom third and oh this is

SPEAKER_96: the bottom yeah i think bottom third i don't know bottom half you know listen in venture you got to find three things you got to find you know you don't want to invest all your money at the top so you got to get the bottom you know bottom third none of us can call the bottom but you know last year at the end of 22 we were certainly in the bottom third evaluations then you want to be early in a major platform disruption i think we all agree we're early in a major platform disruption is probably the third of our careers and then you want to back one of the best firms best brands in silicon valley like if you do like and people know who those are if you do those three things you know simultaneously like you got a good shot at producing really incredible vintage just to put some numbers on

SPEAKER_98: the calipers thing so you can respond to each month america's public largest public pension calipers that's california's manages some 444 billion in capital on behalf of california's 1.5 million

SPEAKER_48: state school and public agency employees is leaning into venture after years of bringing down its vc exposure to one percent to a one percent target the institution investor is now looking to increase its allocation more than sixfold obviously six percent from 800 million to five billion the financial

SPEAKER_00: times reported so thoughts on that shamanth well i mean this is the moment there's a lot of venture firms that are hard up for money and so they may have a shot they can get allocations now in order to get into these funds the problem that they have to realize is they may be buying a bunch of toxic assets or a bunch of toxic partners in the sense that a lot of these people have been getting run over for the last three or four years we don't know again we've said before most venture investors are probably unprepared for this shock because they've never lived through one there's been a lot a lot of junior muckety mucks that were hired because they were xyz middle level exec at rando company

SPEAKER_10: means nothing so i don't know i think it's smart that they have i mean they have to be thoughtful

SPEAKER_00: first of all how is it that the calif like literally it's like i know it's in trillions of

SPEAKER_10: dollars being made in your backyard and some genius was like well the thing that's creating all the wealth in the world which is in our backyard where we probably have the best pitch in order to get into

SPEAKER_247: these funds and you just call doug leone and say hey oh my god give me the list of 10 and put money on those 10 and call it a day oh my god that's unbelievable that's unbelievably derelict that's

SPEAKER_00: unbelievably sloppy and emotional sorry that's it sorry let me just finish that's actually that's the best

SPEAKER_10: word to use when you look back on a decision like that there's no numerical justification that one could have made in the 2000 teens to have made that decision except that that was an emotional decision and when you're running a half a trillion dollar fund there is no room for emotion you should not be allowed there should be some so i think that that shows a very hollowed out and at a minimum imbecilic risk management infrastructure at calipers that needs to get fixed so whether the allocation goes up or down if i was a teacher and my money was being managed by them i'd say man these people i would want to understand how they're making decisions because i don't i would want to see the investment memo that got them to decide as an investment committee that one percent in the most important asset class that's in your backyard that's making all the money in the world made any sense i just want to read that memo and see could i agree with that because then i would want to then hold on then i would want to read the investment memo that says we're going to change course and get back to five or six percent because if that process isn't fixed these decisions are just going to be equally bad you're just going to compound bad on top of bad because again they clearly did a terrible job and then on top of that they had horrible partner selection because the people that they did invest in because the data is public have performed horribly so what changes now all of a sudden

SPEAKER_56: right because if the best firms still don't want you in increasing it from 800 million to 5 billion just means you're going to lose 4.2 billion more than you would have otherwise at 800 million it says

SPEAKER_171: here they made uh two bets they had bet on uh looks like light speed in the last couple years

SPEAKER_287: light speed and tpg so we'll see how those go those can be tough i do have some data on this because we've spent a lot of time talking with them and as you know jason we were just talking you know like

SPEAKER_106: talking to folks like my bottle who are yeah california is one of the largest sovereign wealth funds in the world right this is not just a state this is bigger than most countries right this is one of the biggest

SPEAKER_96: economies in the world i would say that what my interactions with the new team have been very impressive and i will tell you they're not just looking at funds and building a new portfolio i think they are thinking about doubling down at the right time and they're thinking about thematic bets against super cycles like ai to say let's allocate this much money who are the three or four deep strategic partnerships that we can have to drive you know return on that so i i think it's i think they're on their way but i i agree with you they're the sovereign wealth fund of the most prolific state in the world and they should have outsized returns not trailing returns i just want to point

SPEAKER_198: out uh one thing here is the power of writing chamoff you decided to write your annual letter a couple

SPEAKER_98: years ago brad you also uh will write a letter famously the facebook getting fit one or the meta getting fit one the power of writings and i just did this for uh the launch fund four and when you write a deal memo compared to doing a deck there the questions you get are so qualitatively different and the people you attract are so different it is extraordinary i am advising startups across the board to write really tight deal memos because i write these we make these deal memos internally when we make an investment but man is it great for clarity of thought and for the person on the other side to just stop and read a thousand words or two thousand words as opposed to go through some stupid performative deck it's just so such a better process maybe you could you both could speak to that or friedberg i don't

SPEAKER_397: know if you're you've been writing deal memos but maybe for people who are listening or capital

SPEAKER_81: allocators and founders i've been writing for years why and what is the what is it what does it do for

SPEAKER_240: you yeah well it allows you to actually find people who will critique things in a thoughtful

SPEAKER_00: intelligent way it's hard to critique decks because you use broken english you use fancy graphics all of a sudden somebody that's very good at like graphical layout can dupe somebody else and so you don't get to good outcomes because this weird group think effect sets in when you look at decks

SPEAKER_10: so i'm not a fan of decks i i use them but they need to be a companion to some sort of long form narrative document and i just think it's more useful you get people who can really think about what they agree about what they don't agree about it shows the intellect of the person writing it quite honestly i just think it's a it's like a basic skill that people should have it's a useful skill to teach people as well decks are very dangerous i think if you're going to make decisions i would encourage you the bigger the decision the deck is insufficient it can be a companion piece but it needs to be attached to long form documents but the long form documents don't need to be long either two

SPEAKER_39: three four five pages but without it i think you're going to allow some really bad decisions to creep

SPEAKER_401: into some good ones freeberg any writing from you for deals okay uh he ran he doesn't want to talk

SPEAKER_48: about writing all right as everybody knows uh reddit is on strike right now i should say the mods who run reddit are on strike 95 of reddits went dark they basically turned off new posts or they just went private basically nobody could join nobody could see the content i believe is what that means between monday and wednesday and the reason they're doing this is because reddit decided it would start charging for its api so who gets impacted by using the api it turns out apps we saw this at twitter as well when

SPEAKER_98: twitter started changing its pricing for its api and this means that some of the really high-end reddit apps would have to pay 20 million dollars a year for access to reddit's data now they originally had said they were changing this pricing because they were going to train ai models and they wanted anybody using that data i.e uh google bard or chat gpt4 and closed ai they wanted them to pay for it and steve huffman

SPEAKER_101: was uh explained all this in the new york times profile in april you can go search for that reddit wants to get paid for helping to teach big ai systems the largest app is called apollo uh just so you

SPEAKER_98: know if you're if you're not into those reddit really came out with their app really late this was a function of the 2005 to 2015 time frame back then web 2.0 startups didn't have a lot of capital so they let other people build on their apis and build apps some of those apps caught steam and are actually better than the apps developed at least for a while it was better than the twitter app and

SPEAKER_60: the reddit app so thoughts on this freeberg there's a history here um is mimicked at both facebook and twitter both of whom had open apis that provided access to third-party app developers to build tools on top of the platform by accessing either user data or content off of the network and then making that available via some different product function some different ui than the native tools allowed if you'll remember facebook started to kill off its api and in the process killed a number of these third-party app developers the most prominent of which was zynga i think this was around the 2012 time frame you guys know if i'm right on that i think it's right sounds right and you know we saw the same thing happen in twitter where if you guys remember in the early days a lot of users accessed tweets from people that they followed through third-party apps and third-party apps all competed for the user and ultimately twitter's management team realized that having direct access to the user being able to control the ui the ux and not just become a data network but to actually become a a service for users made a similar sort of change so you know reddit's motivation around ai training is is an interesting one but it does speak to this idea that these social network companies

SPEAKER_409: social in the sense that the users themselves are creating the value they're creating the content

SPEAKER_60: at both facebook at twitter and at reddit ultimately the company loses the value if that data that content gets extracted and they can't monetize it or capture the value somehow and it's a lot different

SPEAKER_68: you know every company ultimately wants to become a platform company meaning that they can offer multiple products or services to users that sit on top of some you know network that they've created and in the process create a network effect because more products more apps creates more users more users you know begats more more uh more apps and so on that works well in some contexts like an apple app store context but in the context where there is a network unto itself like facebook twitter and reddit

SPEAKER_409: meaning that there is already a user network that is generating value in the form of the content that's

SPEAKER_60: being produced and consumed you don't necessarily gain anything by then building an app network on top of

SPEAKER_68: it and i think that's been one of the kind of key learnings that's repeated itself over and over with twitter and facebook the thing about reddit it's always been a community service if you guys remember like in 2014 2015 ellen powell stepped down after there was a reddit revolt she was the ceo at the time and she fired an employee at reddit that ran the q a site for reddit's mods and their users

SPEAKER_60: and the network the community was super pissed off when this happened and they all revolted and they were going to shut down the service and ultimately ellen you know got removed from her role as ceo when this happened so you know because so much of the value of reddit isn't in the management team it's not in the work that the software engineers do that run the company it's not the the vcs or the shareholders the value of reddit is inherent in the community it's inherent in the individuals that build the content on that platform and that community has convened many times in the past at reddit to change the rules to say this is what we want this community to become and this is how we want this management team to operate and so it's a really uniquely positioned business says a lot

SPEAKER_68: about how social networks in this kind of modern era are operating it really speaks to how much of the value ultimately accrues to the shareholders in a business like this when the users themselves can step in and unionize and say hey you know what we're not going to allow this much value to be pulled out of the network in this way we want this to change so i think it'll it'll have a lasting effect in terms of investing in social network or social media type businesses where the users are generating so much of the value and have the ability to kind of communicate with one another

SPEAKER_98: and control where value ultimately falls you were at facebook i think when zuckerberg realized enabling a bunch of folks to use the api wasn't as good as controlling the user experience having a uniform user experience and uh it got deprecated and i remember zynga and a bunch of other people had games and were sucking users off the platform there was a linkedin competitor at one point that was growing at an incredibly violent pace and i guess you all made a decision we don't want you sucking our user base off the platform maybe you could expand on what the decisions were at facebook at

SPEAKER_234: the time that was one of the things i ran i think one of my teams was responsible for facebook platform

SPEAKER_00: yeah it was just a very clinical decision around enterprise value look the the thing with reddit is that it's a hot mess and in order to try to create enterprise value they decided to really leverage the mods so that there was some level of control and that control was necessary so that they could

SPEAKER_10: basically sell ads that's how this thing moved in lockstep because the minute that there was corporate venture investors and other investors and a need to generate enterprise value and is it worth two billion or five billion or ten billion whatever the number was that they thought they were worth they had to make money right and huffman was very straightforward about that and wanting to go public and the whole nine yards but because it was such a hot mess the mods became this integral part of the ecosystem so that they could actually drive reasonable revenue

SPEAKER_00: but then what happened was it also allowed them to basically take over and i think that it was

SPEAKER_10: a pretty significant miscalculation because i think that what they needed to do was really redefine the economics of how revenue generation splits would work before they could do all of this stuff and try to monetize the api so i think like they got the order of operations wrong but i also think it's very fixable and i think that they have some very smart people around that table so as long as they're again willing to be clinical and unemotional like we were they'll get to the right answer which is give them a healthy rev share that's the future freeberg you know what version of what freeberg said is true the content creators need to get paid you know why you know you see content creators now on youtube making millions tens of millions hundreds of millions in a few unique cases billions and then we are all content creators yet most of us on these old legacy

SPEAKER_00: platforms make nothing so that exchange that exchange has to change brad any thoughts with us derail the

SPEAKER_418: ipo not on that but i mean i think you know you bring up this thing about meta you know did anybody pay

SPEAKER_106: attention to you know such launching project 92 right and project 92 is going to take on twitter it's a text-based social network that's going to pay creators and they're courting apparently oprah winfrey the dalai lama and all of their creators that are already on their current sites are saying we will use this thing to interact and we will compensate you and then on lex friedman's podcast

SPEAKER_96: he mentioned something about um you know having been inspired a little bit by what was going on with blue sky so i'm super intrigued you know he talked about it at this all hands meeting i think chris cox talked about it so it looks like meta may be revisiting some of these things that they shelved a while back you know it doesn't have any direct uh implications on the reddit front but i think there's a suggestion here that it may be more about putting the control back in the hands of the user from a data perspective and a monetization perspective that would be a pretty gangster move

SPEAKER_380: you know and an interesting way to leverage the platform and i don't really hear anybody talking

SPEAKER_143: about it there's a really easy solution here for reddit those mods most of them do it for fame glory

SPEAKER_48: and affiliation uh community but if some number of them wanted to monetize their activity there why not allow people to subscribe to subreddits and pay a membership fee and uh split it with the mods that that seems like it would be a high scale move adding patreon you know subscription services to let them make a little bit of money and then with these it's only three apps that are being affected

SPEAKER_101: they should just either buy those apps and bring those teams internal or i think it's not or they could split revenue with those apps they could tell those apps that that's the key issue it's like

SPEAKER_00: it's like if you're going to try to monetize apis on user generated content i think what's happening

SPEAKER_10: here is the internet is saying okay that's enough because we're going to leak our activity someplace else where we can directly monetize it so that's the whole point i think in this current version of the internet the value is going to go and again further further further erode away from these centralized apps and more towards the individual people or in this case these hubbard spokes these mods and not to the centralized organization that that has the housing the reddit the facebook of the world so that's just the trend that's happening the instagrams of the world and there's nothing wrong with that youtube as well it's just where the pendulum is swinging so i think

SPEAKER_00: that reddit just has to cut them a deal pretty easy to do speaking of ai funding there was a breaking

SPEAKER_101: news story just in the last 48 hours a startup named mistral ai has raised 105 million seed round

SPEAKER_423: they're calling it should be at about a 240 million valuation according to reports one of the co-founders is a deep mine researcher i guess people were saying this is insane because they haven't written any code yet and they've been working on the company for a couple of weeks why these rounds are so big is i

SPEAKER_39: guess that you have to buy all these h100s and they're expensive and these rigs are very expensive i

SPEAKER_00: mean you guys saw this that nat friedman basically published that he spent 75 million dollars on a bunch of hardware and if you were one of his portfolio companies you could use it so getting these h100s and a100s to train on seems to be a non-trivial task and so and they're very expensive even if you can't get a hold of them so i think what we're talking about is basically that these rounds have to go up my and because if you notice like the post you had to raise the post so that it wasn't so utterly dilutive as to completely disincentivize the employees but let's be honest here nobody's writing 105 million dollar seat check that 105 million is chopped up probably 10 ways to sunday so there's a bunch of people putting in fives and tens and fifteens i'll just be honest in the round and they said it was massively oversubscribed yeah so i so it's everybody taking a little teaser bet the problem with these teaser bets is they never hit in the way that you think maybe you'll get

SPEAKER_10: your money back with all the dilution that's going to happen etc etc this is not the way to make money guys i'm just going to be honest with you so whoever's putting money in thinking they know what the they're doing you might as well just light it on fire go to vegas and have some fun with it because

SPEAKER_39: you will make more you'll get more enjoyment from that than you will for making these kinds of

SPEAKER_05: investments stupid stupid bat brilliant bat well by the way this has nothing to do with the

SPEAKER_08: company nothing to do with the idea just talking about when you're betting 4 million in 105 million dollar round at 240 plus you do not know what the you're doing that is not the job so again mr there

SPEAKER_106: are going to be a couple monster rounds i think announced next week like that are going to make this one look like kids play um so a lot more of this is coming you're exactly right this is about

SPEAKER_376: buying h100s and compute everything we're talking about right an essential ingredient is compute and

SPEAKER_273: it's a scarce resource my god these people that put the money in the seed round should have just

Chamath Palihapitiya: bought in video buy some call options like you'd make more money well and there's downside protection video has a floor and by the way because in just in case mistral doesn't work

SPEAKER_434: nvidia will sell those h100s to somebody else to resell about yeah they probably have the option to

SPEAKER_106: resell crazy i do think you're right on these teaser bets this is a power law business there's a massive pressure on young junior partners principles within these firms to do something fire them it's not just teaser bets what happens in these fire them fire them they want to get the

SPEAKER_139: logos because they need to explain to them sorry no actually no that's even worse the gp that can't

SPEAKER_10: manage their principle should be fired right so those people should those the young people should shut the up okay be lucky you have a job learn the craft it'll take you a decade and if you are proposing stupid bets like this again sizing matters again i'm not talking about the company here i'm just saying when you make a five million dollar three million dollar deal memo for 105 million dollar round that is stupid okay and so if you're the partnership that allows those kinds of things to leak through you don't know what you're doing so at some point somebody should be held accountable for this and i guess what's going to happen is the returns of calpers are going to cascade through everybody else thinking well at least i own some you know h100s for a minute so let me stipulate i think

SPEAKER_357: this could you want to take the other side brad of that no this is just an incredible team that's

SPEAKER_106: going to do you know that this may turn out to be a fantastic bet or i don't know light speed or whoever

SPEAKER_96: led this round let me let me say something different in in 1997 98 we had a similar phenomenon everybody thought internet search was going to be huge there was massive fomo and chasing and everybody scrambled to get a search logo altavista infoseek lycos go planet all you know geo cities just go through the lists you know that people were scrambling after and the truth of the matter is almost all those companies went to zero even though you got a couple bets right you got the internet right you got search

SPEAKER_218: right but you didn't have to invest a dollar in search until 2003 and you would have captured 98 99

SPEAKER_446: do it again now for social networking same thing do it again for social networking brad say all the names

SPEAKER_96: say all the names it's the same thing same thing and so you know when you when i look at it today we have a huge anti-portfolio for ai it's painful we've said no to over 60 companies right we you know but when we look around i see a lot of our competitors doing a lot of these deals maybe they're teaser bets i don't really know the size they're putting into those companies but i suspect that if we believe this is as big as is is it's going to be and going to play out over decades then putting a bunch of really small bets in order to buy a network or buy relationships or buy logos etc i don't think it's going to work any better this time than it worked then around social networking but to be clear there are going to be some people who lead these deals who help these companies build incredible businesses and those will you know there are going to be some winners here i think there's time to participate in the winners right now a lot of this is unknown and unknowable it will become more clear in the one

SPEAKER_10: two three years ahead the problem is for the lps like if you are the lp i am an lp in a bunch of venture funds this stuff really turns my stomach because i'm like wow i am losing money every day when i see these things that's why i get so emotional about this i think to myself if the gps that i've given money to are doing these kinds of deals i'm screwed at best maybe i'll get back 50 or 60 cents on the dollar and i immediately start thinking to myself i really need to write this down and i'm re-underwriting that person and that organization because i'm wondering how can you let these things happen because if you just look back in history you have to be really really negligent to not learn that these things never work out the way that you think they are and especially these kinds of rounds and this nominal ownership the inability to defend ownership it's just not a path to success i mean

SPEAKER_48: also if you think about this shama you know what company in recent history that got overfunded actually

SPEAKER_12: use that money logically the magic of silicon valley is the milestone based funding system and whenever you short circuit that this money becomes a huge distraction to founders if they were to receive 10 million 25 million work for a year or two then raise another 25 or 50 million they don't need to raise all this money at once this is like taking your abc round putting it all in 100 and all it does is distract founders and then everybody coming for a salary says well you got 105 million in the bank i want three million dollars i want five million dollars what's the best example guys of a huge

SPEAKER_00: financial winner that raised these ginormous amounts of money what's the best free launch of a

SPEAKER_230: product jama there's no product here what's the best example magic leap what is it

SPEAKER_30: oh of disasters that actually no no no what is the best example of a great company quote i'm putting great in quotes that raised these kinds of amounts so early in the cycle quibi

SPEAKER_218: no brad do you have an example what's me what's made i think there are people who raised a lot of money over the life cycle like an uber no no i'm not talking about that three rounds at once what was

SPEAKER_02: your first seed check into uber what was the total seed round i can tell you what facebook's was it

SPEAKER_464: was five hundred thousand dollars it was one point two five i think okay facebook's was five hundred

SPEAKER_10: thousand dollars yeah we need a lot of money later we needed a lot of money later but no it's five million sorry five hundred thousand safe at five five posts here's another five million dollar

SPEAKER_106: valuation for uber 1.25 and they went here's another thing everybody walks in they said well i have to

SPEAKER_96: raise this much money and as a it's it's good this is a circular logic i have to raise this money because i have to have all this compute and i say well okay you got to train it you know we want to have a vertically integrated model we want to train a model i say okay so there's a huge upfront cost and they're like but i don't want to take a lot of dilution so i have to raise it at a really high price and and so you say okay well um that's a that's a challenge for you not necessarily a good and then they say oh and i if you ask the question is this an ongoing expense they're like oh yeah we're gonna have to retrain like we're gonna have to continue to spend this money and i'm like well so if if you're a software company for example and you say well what's my cogs if all of a sudden you have an embedded cogs that's massive and recurring right for your compute costs that we haven't had in the past then the revenue on the other side of that's got to be a multiple of what a traditional software company might have in order to get back to that set of economics so i think there's this there's a scramble and understandably so if you think the big win agi or you know this autonomous agent that's going to you know be the top of the next funnel that thing is going to be worth a lot but to me those are those are lottery tickets at this at this stage can i tell you a little secret

SPEAKER_151: here's a little secret when you put in a hundred million dollars into a startup to buy compute you

SPEAKER_10: are not buying whiz bang next generation ip you are subsidizing capex and that is a job that many many other people will do at a very low hurdle rate and so it is a law of capitalism that it could be the most incredibly innovative company in the world but if you are offering money to them to fulfill a low yield thing you are just not going to make a lot of money when you put money into a startup that is their writing code to build groundbreaking ip you own something that's really real but that's because 80 cents on the dollar is going to core critical r d in that point in time and then they raise a lot of money at a lot less dilution when 80 cents on the dollar goes to sales and marketing then they raise a lot of even more money at an even smaller smaller amount of dilution you start to get to scale internationally so you start to see right more dollars less return right you're owning less of the critical differentiation so if if everybody is like oh it's so expensive for compute i need to raise 100 million well buddy you know that's like a leasing function you know you're you're like all of a sudden like the best vcs in the world have become like comerica bank

SPEAKER_89: yeah i mean they could have done it with the same structure right comerica could have given them 50 million to buy these machines and maybe they should and comerica and jp morgan and somebody else

SPEAKER_273: should basically say hey you know what here's a lease line for your h-100s because i know they're

SPEAKER_10: worth so much and i'll just yes write it at 10 percent and and my point is that the fact that people don't understand this is why the money will get torched i would love a critique that says actually

SPEAKER_273: chamath you're an idiot i'm right i know that that's happening but here's why i still see it happening

SPEAKER_460: i don't hear any of that here's the problem with that critique okay so you asked like what are the

SPEAKER_96: biggest projects in history you know uh around startups think about aws i don't know they spent

SPEAKER_106: 400 million probably uh in order to get aws off the ground but it wasn't done by a startup right you think about what zuck's spending on the metaverse it's not being done right by a startup the truth of the

SPEAKER_10: matter is can you go back to ap aws aws was dog fooded on amazon retail of course of course and

SPEAKER_00: oculus was done on kickstarter and the cash flow of amazon retail fed the development of aws correct

SPEAKER_96: my point is that when you think about what these hyperscalers are going to do they're not going to spend a billion they're not going to spend 10 billion they'll spend a hundred billion dollars right in order to be in this race and so if you're backing a startup that says i'm going to build a better chat gpt right just like open ai discovered themselves they sold 51 or 50 of the company to microsoft for a reason they had to they had to have the data and they had to have the compute this is a nuclear arms

SPEAKER_460: race around compute and so but i think it's this is insane it's financially illiterate for someone to

SPEAKER_10: think that they are actually doing anything other than subsidizing capex when you're giving a hundred million dollars to a startup to literally buy chips and servers and give up 40 of their equity

SPEAKER_12: that's the other thing this equity is so valuable why would you want to give 40 of it when you could

SPEAKER_479: get an equipment lease and keep another 20 of your company you can't you probably can't get those you

SPEAKER_12: can't get i mean if the vcs put in 50 million instead of 105 you don't think co-america would come on the back end with 25 million of course they would brad is right this happened in 98

SPEAKER_00: 99 2000 where all of this money was getting flushed down the drain going into buying data center capacity where remember even at facebook like we were racking and stacking our own servers and then

SPEAKER_39: we then we ultimately got big enough where we actually built out our east coast and west coast

SPEAKER_10: data centers and data centers all around the world but it was very expensive and in that moment again all of those companies just lit all that money on fire it they torched it there was no remnant equity value for that capital and so i guess i guess i guess i'm just i'm just questioning like what does a gp think they're actually buying well 80 if 80 of it's going to hardware i mean they're

SPEAKER_482: buying the other 20 buying chips i mean it doesn't make a chip and the chip that's going to get

SPEAKER_98: i want to make one point here and freeberg i want i want to get your input on this as well

SPEAKER_54: constraint is important for founders and the thing that i find really troubling about this is and putting this startup aside because crypto people also went through this for the last three or four

SPEAKER_98: years where they were overfunded it was tens of billions of dollars burnt of lp monies people's retirements and college endowments and it's going to be quite a post-mortem but look at say

SPEAKER_12: you invoked meta it's important for people who don't know this to know that that was a kickstarter shout out to palmer lucky he raised a couple of million dollars in 2011 i think it was on a kickstarter pre-selling the devices right constraint makes for great art constraint makes for great startups you need to have pressure on a startup for them to deliver you cannot give startups five years of runway and expect it's going to work it just doesn't work i've never seen this movie so

SPEAKER_273: many times but now we've gone through 18 months of nobody doing anything i guess in vc land so folks on

SPEAKER_00: sandhill rotor are itchy they want to justify why they should still be drawing two percent on the full fund they want to try to show activities so that they can raise the next fund and continue to stack fees

SPEAKER_10: and all of these sort of leads to these suspension of financial logic but it gets replaced with financial illiteracy which is why there's an optimal fund size right shumath and this

SPEAKER_273: is why the people that pay the price are ultimately the lps and it may be the case that calpers

SPEAKER_10: maybe actually avoids a lot of these pitfalls because by missing yeah they missed all the returns

SPEAKER_00: but then they missed writing the mega follow-on checks for all of these folks that they and then

SPEAKER_106: they would have torched okay i'm going to take the other side of that go ahead brad i think it's nearly impossible to conceive that all of these bets that are currently being made are bad bets

SPEAKER_259: i think it's a major platform disruption but i you know instead i think the right way to think

SPEAKER_96: about it is it's about pacing and if you're trying to if everybody thinks they're going to build the next google they're going to build the next autonomous agent that's going to sit on top of the funnel that's going to be worth a trillion and therefore they can burn a billion dollars training models that's not like we're not going to have 10 of those winners okay but at the same time there's a lot of stuff getting funded that is in the application layer that is in the tool layer and these are not the big headlines that you're reading about but these are really interesting businesses that are solving real problems in the software and tooling layer here you know in the smaller model layer vertically oriented things around life sciences or uh you know targeting you know financial services or things in the application layer you know like character.ai etc so i do think there are a lot of good things getting funded that will deliver real value but i agree with you the problem that there's a there's a second problem to this chamath if you drop a billion or 2 billion or 3 billion into something you have not only a a product challenge you have a distribution challenge right we know all the hyperscalers are going to play google is going to play meta is going to play and so you've got to compete and then beat them and it used to be that you would say well if i get a lot of traction they'll buy me if i'm instagram or whatsapp like they'll buy me well we have such a regulatory nightmare in washington dc today no hyperscaler can spend over a billion dollars to buy

SPEAKER_490: any ai company not even that 400 million gift what is it giphy i bet it wasn't even unfortunately in the

SPEAKER_96: united states today we replace mergers and acquisition right with copy and compete because we've said to hyperscalers you're not allowed to acquire any of these companies so the unintended consequence of the regulation in washington is that entrepreneurs and founders and venture capitalists who might otherwise have had a good idea built something with some traction they can't find a

SPEAKER_218: home for it in the way you know that whatsapp found a home or is that a good thing instagram

SPEAKER_494: is that a good thing because they go public and be independent is it better for the market presupposes

SPEAKER_334: that everything can become a big and profitable business there are a lot of net net can become a

SPEAKER_247: big and profitable business no on its own no chance well it's still not a big and problem okay so that's

SPEAKER_171: what i'm saying so maybe it should have died i mean maybe it's been kept alive but i mean it's

SPEAKER_259: oh i actually don't you think it's better for the market acquired i think it's created it got it

SPEAKER_96: it was it was innovative technology it was yeah it you know you you were able to to back it with some good funding and i think what's coming is going to be really exciting but it took a really long runway a lot of capital a lot of intelligence in order to build unfortunately we've killed that so you have a two-sided problem we're spending more than ever to fund and start these companies and we you know have undermined a lot of the downside protection freeberg your thoughts so if you

SPEAKER_409: look at how the capital is being deployed if it's mostly being deployed to train models

SPEAKER_68: then the question has to be is there really a sustainable advantage that arises by being the first to train the models and then being able to persist an advantage by training new models from that foundational model going forward and the reason that that matters so much is because you have to really have a deep understanding if you're going to invest a lot of capital here you have to have a deep understanding for how quickly model development and training is accelerating and how quickly the costs are reducing so something that costs for like we said open ai spent 400 million dollars training models for for gpt4 if they spent 400 million dollars in the last couple of years you could probably assume that doing the same training exercise could be done for five to ten million dollars 18 months from now to generate the same model that's a you know 100x cost reduction and i'm just ballparking it here but if that's really where things are headed then does the hundred million dollars to train models today really make sense if trading those same models can be done for five million dollars in 18 to 24 months great point and that's where it becomes a really difficult investment exercise and one that you have to really critically understand how cost curves are moving in ai the same thing was true in dna sequencing in the early days and it's following by the way a similar cost curve as dna sequencing which is actually greater than moore's law greater than a 2x cost reduction every 18 months we're seeing something much greater than that in machine learning right now in terms of cost reduction and model training so ultimately the business model has to have some advantage that by being the first to market you can then generate new data that gives you a persisting advantage and no one can catch up with you and my guess is if you get under the hood of the business models it's unlikely going to be the case and it's very likely going to be the case that you don't know when the market advantage will lie when you will be able to kind of create a persisting mode a moat that expands as you get more data and can train more and this is why it's so hard to invest generally in technology it's because you don't know the point at which the market tips relative to the point at which the technology tips so there's a moment where the technology gets so cheap and then the market maybe adopts after the technology gets cheap and at that point it's a

SPEAKER_63: totally different game remember in the mid 2000s where we had memory shortages and we used to have

Chamath Palihapitiya: to buy ram yeah i mean it's just like it's all this stuff and it's like if vcs are funding this stuff just you just like like the equity on fire guys not going to be worth anything brad's point is right

SPEAKER_68: which is the question is what's possible now where can you build a sustaining advantage now rather than go after big model development cycles where the cost curve is going to come down by a hundred fold in some period of time in the near future is there a business model advantage you can build by being in the market first building a customer base accelerating your features getting user feedback and that certainly exists in the application and the tools layer as brad is talking about that seems like such a no-brainer for disruption across many different segments many different verticals many different markets right now versus trying to compete further down the stack where it takes hundreds of millions of dollars of capital and in a couple of months that hundred millions of dollars of capital can be replaced with five million bucks of training

SPEAKER_00: exercise and compute can i take the other side of that yeah all that coordination makes no money today so to your point when you cut the actual input cost by a hundred fold the coordination cost goes from being zero to being worth less than zero right i don't see any money being made there either and

SPEAKER_10: all the people that say i'm sure there's going to be some genius in the comments but what about open source it's like what about it open ai also just gave an update on their cost structure for their

SPEAKER_98: api and they just dropped it 25 to 75 again this is after the tenfold drop they did last year play

SPEAKER_68: this out a hundred million dollars of capital spent trading today is a million dollars spent doing trading in 18 years yeah three years 18 18 to 36 months somewhere in that time frame is likely the time frame so why would you spend all this money today when an 18 to 24 month things are going to get so much cheaper yeah i think the further up the value stack you go the more of an opportunity to

SPEAKER_187: truly kind of innovate disrupt and make what capital is possible what happens in 24 months though when

SPEAKER_58: you've made a bunch of hundred million dollar investments and they're all zeros uh you uh get fired

SPEAKER_71: unfortunately as an investor chamath is that what you're asking at some point you don't get invited to

SPEAKER_68: join the next fund think about the alternative investment strategy with lots of capital where the cost curve is not coming down as quickly in terms of where that capital is being deployed for example building rockets to go to space or building uh infrastructure to transport power or building roads you could raise a billion dollars to build a toll booth system or or a port let's use a port a shipping port is a good example you could spend a billion dollars to build a shipping port it's not that the cost of building a shipping port is going to come down by 10x in 18 months so it makes sense to raise a billion dollars and build a friggin shipping port and charge people right

Chamath Palihapitiya: to coming out vcs thought they were underwriting ip instead they're just actually subsidizing capex

SPEAKER_12: it's the biggest thing it's the craziest thing brad is this a problem of the optimal venture fund size that fred wilson talks about that bill gurley talks about a lot of the ogs say hey 250 400 600 million there's an optimal size here for four or five partners in a venture fund to put money to work

SPEAKER_98: is this part of the problem right now which also happened in crypto is you had billion dollar two billion dollar funds sitting around and and different venture brand names having four or five of

SPEAKER_101: these multi-billion dollar funds burning a hole in their pocket and getting frisky uh over this you know 18 month pause and is this about optimal fund size making bad you know we have over a billion

SPEAKER_44: dollar fund so if i if i if i take the other side of that then people are going to say you're talking

SPEAKER_106: your book but i don't think this is about large funds i think this is about good and bad decisions at the end of the day some of these decisions will pay off like for example what's the largest bet

SPEAKER_519: size from a billion dollar fund that you've made or will make i think out of that fund would be a

SPEAKER_45: hundred million dollars most likely uh but we may cross it over other funds and have you know bigger

SPEAKER_54: bets um certainly would that happen two or three times though would you do it all at once like this or might it happen over you know a series abc kind of situation where you build a position i would say

SPEAKER_96: you know we're not writing i you know we haven't written a hundred million dollar check into a series a and i don't think most of people to chama's point that you're talking about are writing a hundred million dollar checks into those series days the mr all-round that you reference i imagine the lead check into that was maybe 50 maybe 50. so listen larger fund sizes enable you to participate in companies that require more capital and and these companies do require more capital so you may take the position that all these companies are going to zero that's not my position my position simply is that i do believe there is a bit of over exuberance that too many things are getting funded right and that some you know like the margin of safety the margin of return being required is probably lower than it should be but there's no doubt out of this vintage in my mind that you're going to have some epic companies now i don't know if what mustafa and reid hoffman are doing at inflection is you know building pi to take on chat gpt and to take on bar to be the intelligent assistant the ambition is extraordinary the cost of compute is high but the first mover advantage is also high right because whoever secures this position you know bill gates said he's been playing around with it it's one of his favorite agents or whatever that's an interesting comment i think it's pretty good but i think chat gpt is out in front in this regard i think a lot of people are going to try to compete for that space but you know i can't imagine that all these researchers leaving deep mind right are going to be able to compete for the most sophisticated model to answer general purpose questions so i don't think it's the large fund size i think it's just a lot of exuberance to participate in what everybody perceives to be a massive platform disruption i think that can be true just like the internet was in 1997 and a lot of these bets can you know can and will likely go to

SPEAKER_10: zero how great is america you just rent other people's money they pay you two percent and then you're allowed to get exuberant yet keep your job when you lose it god bless america well i mean the

SPEAKER_101: the issue chamath is it takes 10 years to prove you're bad at this job no no no no it takes 10

SPEAKER_10: years to prove you're good and it takes 20 to prove to prove that you can be consistently good and didn't get lucky but it in a few years you can tell that somebody sucks can lps tell well i'm not sure that they get the visibility because when lps interact with gps they're grin for the most part so i don't know probably not but when i interact with them just as a peer-to-peer level and i see the deals that get done it's pretty easy to understand that some people just suck they don't know how to make money i guess is the point which in the job that's the job is to make money to

SPEAKER_54: consistently make money okay brad take care and uh let's go to science corner bill gates wants to genetically modify mosquitoes is this fake news or real incorrect i see it's fake news okay explain

SPEAKER_68: the reference is from rfk jr's tweet that he sent out where he retweeted someone talking about this mosquito factory in colombia and this guy basically put out a tweet saying oh look bill gates has a

SPEAKER_60: mosquito factory in colombia it's the largest in the world 30 million genetically modified mosquitoes are released every week into 11 countries because bill knows better than nature what could possibly go wrong rfk jr then took it upon himself to retweet and say should bill gates be releasing 30 million genetically modified mosquitoes into the wild part of the mentality of earth as engineering object what

SPEAKER_68: could possibly go wrong so i really wanted to take issue with this because i do think that this is the sort of misinformation that both creates scientific illiteracy and damages and impacts negatively some of the significant progress that can be made in medicine and in science so i want to speak very clearly as to what is going on what the science is behind it why this is super important and then we can speak philosophically if you guys are interested on kind of should we be doing this

SPEAKER_54: sort of stuff and why so what's real here like what what are actual facts versus fake news maybe that's a

SPEAKER_68: good place to start the most common disease carrying mosquitoes are called uh edis uh aegypti uh gyptia it's a species of mosquito that carry yellow fever dengue uh zika a number of viruses that obviously are pretty adverse to human health each year about 400 million people are infected with dengue virus via this mosquito vector 100 million become ill and 21 000 deaths are attributed to dengue globally 200 000 cases of yellow fever each year causing 30 000 deaths these are pretty significant health concerns and it turns out that in mosquito populations not in this particular species that carry these viruses but in other species of mosquito there's a bacteria a natural bacteria called walbachia and this bacteria exists in nature and sometimes these mosquitoes get infected with this bacteria so they carry this this bacterial bug and this bacteria is really interesting because it causes what's called cytoplasmic instability in the mosquito cells which actually makes the mosquito largely resistant to a lot of viruses and there's a bunch of theories for this mechanism and why this is the case but it causes the mosquito to not be able to carry and spread these viruses that are super adverse to human health so number one there's a natural bacteria that's found in nature and up to 40 percent of mosquitoes are already infected with it number two is it's not common in the mosquito species that is common in these areas that are spreading these awful viruses to humans and so there's been a project that's been going on now for 12 plus years where they're taking large amounts of mosquitoes and breeding them specifically to have this bacteria in the mosquitoes and then they release those mosquitoes into the wild and over time the bacterial infected mosquitoes start to become a larger percentage of the population and as a result the vector of carrying these awful viruses into humans goes way down there was a study done in indonesia where they took these walbachia infected mosquitoes and they released them into the wild and they looked at a population that was in a region where they released them in a population that they didn't in the relation region where they didn't release them 9.4 percent of people ended up getting infected with dengue fever and where they did release them only 2.3 percent were infected so it was an amazing 75 reduction in the infection of people by these mosquitoes and so the whole point is just to kind of you know move the mosquito populations in a way without doing any sort of genetic modification but by exposing them to this bacteria so that they don't carry these viruses into

SPEAKER_98: people but you know because so that's a nuanced point there friedberg the it's not genetic modification

SPEAKER_165: or it is it's not i just explained it's a bacteria and they're just exposing mosquitoes they're exposing

SPEAKER_68: the bit so that they end up getting infected with this bacteria and then as they breed the mosquitoes breed in this facility you have you have to have two an infected male and an infected female for the offspring to have the bacteria if you have an infected male they're actually infertile they can only fertilize an infected female so unless you do this breeding work you don't end up seeing this happen naturally in the environment where the wolbachia starts to spread where is the genetic modification

SPEAKER_187: misinformation coming from then where is that also occurring from from from the the presidential candidate rfk jr who just propagated it and so this is why i want to make this point because this

SPEAKER_68: whole idea that oh should we be engineering the earth let me just say something about engineering the earth humans used to wander around the earth or proto humans did without access to food and until we realized that we could plant a seed in the ground and grow crops and started to engineer the earth in the form of farming we did not have access to a reliable source of calories human ingenuity human engineering gave us the ability to do this gave us the ability to feed ourselves similarly humans got infected by viruses by bacteria by fungus and died at a young age over and over again and when humans began to engineer medicine and engineer unnatural substances because he makes this point oh should we be interfering in the natural world what is natural is for people to get infected with viruses or bacteria and die and if not for the advent of our engineering and our ingenuity and our ability as a species to create solutions through science which allows us to do discovery and then through engineering which allows us to make solutions that solve problems that humans face we would all be dead at a young age and we would not have realized the progress that we've had as a species so i really get ticked off when i see guys like rfk jr and others not just propagate this this bs misinformation spiel about oh genetically modified this and that science is bad but to then say should we be messing with the natural world because i would say to him what about when your kid got infected and you gave your kid antibiotics maybe you shouldn't have done that and there is a group of people who are saying

SPEAKER_12: that you shouldn't do that right there is a movement to stop taking antibiotics because it's making

SPEAKER_68: because it's it's having a i'll tell you a couple things there are bad pesticides that impact human health and cause damage to our dna there is bad sunscreen that is endocrine disruptors and can damage human health there are plenty of chemical products that are made that we use in everyday products that cause cancer there is an endless string of things that are wrong with the system of engine the systems of engineering that we do use that doesn't mean that they're all bad that doesn't mean that we then say hey you know what let's not do anything let's not do any engineering let's not use any antibiotics net let's not use any technology in food and that's the challenge is you know getting into the details on like i'll tell you i don't use uh any sunscreen products with myself or my kids i only use zinc and i have a similar sort of nuanced approach to understanding what things we should or shouldn't use in our lives because of pause pause pause pause pause can you please explain that

SPEAKER_172: because i didn't know this what am i doing am i getting what explain yeah so there's a number of

SPEAKER_68: substances uh which are known endocrine disruptors that are found in sunscreen i think it's like one of the craziest things that we haven't made these products illegal at this point but the only sunscreen that you should use is natural mineral sunscreen uh this sorry i shouldn't say that you should i should say this is what i chose choose to do based on the data that i've seen what's a good brand what brand

SPEAKER_71: any brand that's just zinc sunscreen just look at the back if there's anything but zinc in it don't use it zinc oxide is that yes totally exactly zinc oxide sunscreen yeah that's the ingredient in it

SPEAKER_10: so i don't know what you guys um huh wow but uh here let me just send you this are you saying that you're not allowed to have any other ingredient like there's no stabilizers there's nothing else

SPEAKER_00: no no that like it's it's it's the stuff principal ingredient has to be zinc oxide versus some other

SPEAKER_68: chemical you're saying yeah so oxybenzone i don't use any products with oxybenzone that's like the most common sunscreen ingredient uh octanoxate is the other one homo salate and the parabens all of those product categories which are the most common products used in sunscreen they're absorbed by your skin they go into your bloodstream and they're endocrine disruptors now the problem with the zinc sunscreen and the mineral sunscreen is it actually stays on your skin so you look like you're wearing it's really hard to rub it in you got to really really rub it in so they're actually not popular from a cosmetic point of view people don't like wearing them because they look like idiots and

SPEAKER_39: it's really hard to yeah that's the stuff i use in the summer the isdin it drives me crazy because i look like this weird ghost shiny ghost thing i know but i do use that that's like really hardcore

SPEAKER_68: about that so that's an example of understanding nuance right it doesn't mean all sunscreens are bad and it doesn't mean that we shouldn't use sunscreen but understanding what the risk factors are that are associated with different ingredients or different engineering that's been done to make sunscreens available is important but that's so many levels deep it's a really difficult thing so then people end up being scientifically illiterate and being wrong because someone like rfk jr comes along and says hey should we really be engineering the earth with genetically modified mosquitoes and then people have this call to action shut those things down shut those things down and they're incredibly beneficial and effective they're not taking any sort of genetic editing to market they're not doing anything that people might consider risky and we could have a separate conversation about the risks and benefits of genome editing that's another topic but how much of this

SPEAKER_98: freeberg orcshima as we get ready to wrap here is a reaction to what happened with covet 19 and people's fear now of and getting sort of educated on gain of function research and you know this sort of recency effect of my lord doing some of this science feels like it's too dangerous certainly too dangerous to do inside a city and what's the point of taking bats out of caves and and doing gain of function research how much of it has to do with that right now and this is sort of the downside to questioning that technology can

SPEAKER_68: can be asymmetric you can have like nuclear weapons can wipe out the world they can wipe out the whole population yeah you know talib makes this point on his argument against gmos which i would argue against him on this point but we could do that another time if he's willing to come on the show i'd be super happy to debate him on this point but the idea being that there's super asymmetric downside and so you know what happens is people see incremental improvements from technology and they don't really praise those incremental improvements they assume that to be the case it's a linear step function but when something goes wrong it's a big step down and then people are like oh wow and then people get scared of technology and then people want to step away from it and this is true in anything that relates to your health to food systems to the environment now um now it doesn't mean that all technology is bad or all engineering is bad but you know as mistakes are made in the system as new things are discovered we have to retrench and change what we're doing but it doesn't mean that we should stop progress and it doesn't mean that the whole system of humans figuring out how to engineer ourselves of the world around us to benefit the health to benefit the planet to benefit other species on the planet isn't a critical

SPEAKER_57: mission and effort that we should be undertaking well the sunscreen thing is really tilting i mean i just i need to make sure i'm pretty sure we have a good one but i don't yeah i'm on this right now i'm

SPEAKER_101: this has got me a little nervous with my kids i i like these um like i have all my kids have long sleeve sun shirts and i try to that i think is like the key thing and because i my family has skin

SPEAKER_548: cancer do i have to do that since i'm dark-skinned freeberg i mean do you want to avoid skin cancer i

SPEAKER_549: mean i i don't know why you wouldn't wear like a rash guard uh thing you know yeah you know it's a

David Friedberg: summertime you're in the bed you know you want to show up yeah you want to show off your typical stuff

SPEAKER_98: revenge body i get it i get it all right everybody on behalf of sax let me just say ukraine ukraine

SPEAKER_48: ukraine biden biden biden and uh francis mayor francis is now in the race so i guess we'll have him on the pod we already talked to him oh okay so uh yeah at the summit last year and also we're doing a survey for the podcast all in podcast dot co slash survey all in podcast dot co slash s-u-r-v-e-y if you got to this point in the podcast please fill out our survey our listener survey

SPEAKER_557: and we will see you all next time on the all in podcast love you boys bye-bye bye-bye love you best of all

David Sacks: open source it to the

SPEAKER_571: get merges

SPEAKER_578: we hope you guys are back