Speaker 1
That was a week No time to be meek The goal is to seek The next big thing That was a week That was a week Stand back,
Words and timings
ThatwasaweekNotimetobemeekThegoalistoseekThenextbigthingThatwasaweekThatwasaweekStandback,
Transcript Viewer
Is There an AI Bubble?
Jun 30, 2024 ยท 2024 #23. Read the transcript grouped by speaker, inspect word-level timecodes, and optionally turn subtitles on for direct video playback
Speaker Labels
Name the speakers
Edit labels for this show, save them in this browser, or download a JSON override for the production folder.
Transcript Playback
Is There an AI Bubble?
Human Transcript
Timed transcript
Blocks are grouped by speaker for readability. Expand a block to inspect word-level timing.
Speaker
think big,
Words and timings
thinkbig,
Speaker 1
dig deep
Words and timings
digdeep
Speaker 2
Hello, everybody. The last day of June, a Sunday, June the 30th, 2024. Not a day that we normally do. That was the week, our summary of tech news. But it's a special Sunday, at least according to Keith Teer, who's my partner in the tech roundup of the week. I'm quoting Keith's editorial this week. It's Sunday, two days later than I usually send this out. Two excuses. I was in recovery from PTSD after the quote unquote debate. And then I almost had a relapse watching England in the law in the Euro last 16 game against Slovakia. I'm unsure of my mental state now. We won 2-1. We, I guess, being England in extra time. But the other more important quote-unquote game is still decided. And that game, Keith, is AI, right? Correct. It's a big game as well. Well, your England team, they should be really proud of themselves. They beat Slovakia. That's quite an achievement, isn't it?
Words and timings
Hello,everybody.ThelastdayofJune,aSunday,Junethe30th,2024.Notadaythatwenormallydo.Thatwastheweek,oursummaryoftechnews.Butit'saspecialSunday,atleastaccordingtoKeithTeer,who'smypartnerinthetechroundupoftheweek.I'mquotingKeith'seditorialthisweek.It'sSunday,twodayslaterthanIusuallysendthisout.Twoexcuses.IwasinrecoveryfromPTSDafterthequoteunquotedebate.AndthenIalmosthadarelapsewatchingEnglandinthelawintheEurolast16gameagainstSlovakia.I'munsureofmymentalstatenow.Wewon2-1.We,Iguess,beingEnglandinextratime.Buttheothermoreimportantquote-unquotegameisstilldecided.Andthatgame,Keith,isAI,right?Correct.It'sabiggameaswell.Well,yourEnglandteam,theyshouldbereallyproudofthemselves.TheybeatSlovakia.That'squiteanachievement,isn'tit?
Speaker 3
Well, given that they were losing 1-0 until the 94th minute, I think I'm going to ignore your implied joke and say, yes, it is an achievement. And it was a great bicycle kick from Bellingham. I always want to call him Sheringham.
Words and timings
Well,giventhattheywerelosing1-0untilthe94thminute,IthinkI'mgoingtoignoreyourimpliedjokeandsay,yes,itisanachievement.AnditwasagreatbicyclekickfromBellingham.IalwayswanttocallhimSheringham.
Speaker 2
Faringham played for both Manchester United and Tottenham. I have to admit, I'm a former Englishman when it comes to football, but you've clearly clung to your identity, Keith. I'm surprised that you've been living in Palo Alto for so long and you haven't emancipated yourself of this tribalism. You're the most internationalist fellow I know. How can you still care about how England does in football?
Words and timings
FaringhamplayedforbothManchesterUnitedandTottenham.Ihavetoadmit,I'maformerEnglishmanwhenitcomestofootball,butyou'veclearlyclungtoyouridentity,Keith.I'msurprisedthatyou'vebeenlivinginPaloAltoforsolongandyouhaven'temancipatedyourselfofthistribalism.You'rethemostinternationalistfellowIknow.HowcanyoustillcareabouthowEnglanddoesinfootball?
Speaker 3
It's ironic, isn't it? I tell my Argentinian friends that... I'll agree with them that Las Malvinas Argentinas, but when it comes to soccer, I'll spill every last bit of blood to try to beat them.
Words and timings
It'sironic,isn'tit?ItellmyArgentinianfriendsthat...I'llagreewiththemthatLasMalvinasArgentinas,butwhenitcomestosoccer,I'llspilleverylastbitofbloodtotrytobeatthem.
Speaker 2
Well, the other thing you brought up before we get to tech and AI is the debate. What did you make of that? Was it a debate or a debacle?
Words and timings
Well,theotherthingyoubroughtupbeforewegettotechandAIisthedebate.Whatdidyoumakeofthat?Wasitadebateoradebacle?
Speaker 3
You know, my wife kind of shaped my thinking because she said something very prescient. She was super angry watching it and I said, you know, why so angry? And she said, I can't believe that the senior Democratic Party people allowed Biden to stand when they must know for sure this is how he is, or at least can be. And so you've just got to wonder who's running the Democratic Party and how are decisions made? Because this seems like, even from the point of view of the Biden family, setting him up for that is terrible.
Words and timings
Youknow,mywifekindofshapedmythinkingbecauseshesaidsomethingveryprescient.ShewassuperangrywatchingitandIsaid,youknow,whysoangry?Andshesaid,Ican'tbelievethattheseniorDemocraticPartypeopleallowedBidentostandwhentheymustknowforsurethisishowheis,oratleastcanbe.Andsoyou'vejustgottowonderwho'srunningtheDemocraticPartyandhowaredecisionsmade?Becausethisseemslike,evenfromthepointofviewoftheBidenfamily,settinghimupforthatisterrible.
Speaker 2
Actually, not that I'm a particularly big fan of Trump, but it definitely gives their argument that the country is run by these unaccountable elites some credibility. Clearly, having the most powerful blah-blah country in the world being run by a man who can barely put a census together, who wouldn't be out of place in an old-age home, is embarrassing.
Words and timings
Actually,notthatI'maparticularlybigfanofTrump,butitdefinitelygivestheirargumentthatthecountryisrunbytheseunaccountableelitessomecredibility.Clearly,havingthemostpowerfulblah-blahcountryintheworldbeingrunbyamanwhocanbarelyputacensustogether,whowouldn'tbeoutofplaceinanold-agehome,isembarrassing.
Speaker 3
It's super embarrassing. And, you know, our normal narrative on the show, I try to persuade you that America's in relative decline. If you ever needed evidence, it's a country that has these two as candidates.
Words and timings
It'ssuperembarrassing.And,youknow,ournormalnarrativeontheshow,ItrytopersuadeyouthatAmerica'sinrelativedecline.Ifyoueverneededevidence,it'sacountrythathasthesetwoascandidates.
Speaker 2
What's worse, Keith, the English football team or Joe Biden?
Words and timings
What'sworse,Keith,theEnglishfootballteamorJoeBiden?
Speaker 3
I say the English football team and the American establishment in general, it's not just Biden, it's across both parties. It's the elite that run politics and to some extent economics, have no ambition, no belief, no vision, And it's really just...
Words and timings
IsaytheEnglishfootballteamandtheAmericanestablishmentingeneral,it'snotjustBiden,it'sacrossbothparties.It'stheelitethatrunpoliticsandtosomeextenteconomics,havenoambition,nobelief,novision,Andit'sreallyjust...
Speaker 2
So if you put Joe Biden, if you replaced Harry Kane or Bellingham with Joe Biden, would it make much of a difference to the English team? Or in reverse, if Harry Kane became president of the US, I don't think it would make much difference, would it?
Words and timings
SoifyouputJoeBiden,ifyoureplacedHarryKaneorBellinghamwithJoeBiden,woulditmakemuchofadifferencetotheEnglishteam?Orinreverse,ifHarryKanebecamepresidentoftheUS,Idon'tthinkitwouldmakemuchdifference,wouldit?
Speaker 3
No, but if you do it the other way around, there's no way Biden's doing a bicycle kick to score a goal in the 94th minute. At least Bellingham could do that.
Words and timings
No,butifyoudoittheotherwayaround,there'snowayBiden'sdoingabicyclekicktoscoreagoalinthe94thminute.AtleastBellinghamcoulddothat.
Speaker 2
Well, that could be virtual. And that's perhaps where AI comes in. Let's get beyond Trump and the English football team, two rather boring subjects. Is there an AI bubble, Keith? That's the title. Judging from the... poor quality of this AI art. There is one, you always have this AI art and it seems to get worse every week. So we've touched on this many times over the last couple of years since the open AI eruption. Why this week? Why do you ask, is there an AI bubble?
Words and timings
Well,thatcouldbevirtual.Andthat'sperhapswhereAIcomesin.Let'sgetbeyondTrumpandtheEnglishfootballteam,tworatherboringsubjects.IsthereanAIbubble,Keith?That'sthetitle.Judgingfromthe...poorqualityofthisAIart.Thereisone,youalwayshavethisAIartanditseemstogetworseeveryweek.Sowe'vetouchedonthismanytimesoverthelastcoupleofyearssincetheopenAIeruption.Whythisweek?Whydoyouask,isthereanAIbubble?
Speaker 3
Well, there's a really good essay by Peter Walker from Carter that pretty much asks that question. Carter is the company that manages the share tables for a lot of private companies. I mean, a real lot of private companies, tens of thousands. And so Peter has access to anonymous data about things like the valuations that funding is paid.
Words and timings
Well,there'sareallygoodessaybyPeterWalkerfromCarterthatprettymuchasksthatquestion.Carteristhecompanythatmanagesthesharetablesforalotofprivatecompanies.Imean,areallotofprivatecompanies,tensofthousands.AndsoPeterhasaccesstoanonymousdataaboutthingslikethevaluationsthatfundingispaid.
Speaker 2
And he described this character, Peter Walker, describes himself as a data storyteller, whatever that means.
Words and timings
Andhedescribedthischaracter,PeterWalker,describeshimselfasadatastoryteller,whateverthatmeans.
Speaker 3
Yeah. And by the way, he's super good. I mean, there are a bunch of people. Andrew Rattrath is in this week's who runs Data Driven VC. There's a bunch of people who are trying hard to be the font of wisdom when it comes to data and venture capital. I think Peter Walker is first amongst peers just because of his access to data. He just has fantastic data.
Words and timings
Yeah.Andbytheway,he'ssupergood.Imean,thereareabunchofpeople.AndrewRattrathisinthisweek'swhorunsDataDrivenVC.There'sabunchofpeoplewhoaretryinghardtobethefontofwisdomwhenitcomestodataandventurecapital.IthinkPeterWalkerisfirstamongstpeersjustbecauseofhisaccesstodata.Hejusthasfantasticdata.
Speaker 2
What is Walker doing? The link in the newsletter this week is to his LinkedIn posting. What is Walker doing that hasn't been done before in terms of asking or answering whether or not there's an AI bubble? Because it seems to many people, including myself, that there is one, but maybe that's meaningful, maybe it isn't.
Words and timings
WhatisWalkerdoing?ThelinkinthenewsletterthisweekistohisLinkedInposting.WhatisWalkerdoingthathasn'tbeendonebeforeintermsofaskingoransweringwhetherornotthere'sanAIbubble?Becauseitseemstomanypeople,includingmyself,thatthereisone,butmaybethat'smeaningful,maybeitisn't.
Speaker 3
I think he implies that there is without saying so. What he's doing is he measures the valuation of companies at their A round. And an A round used to be the first time a company raised their money. Back in the day, your first funding was called an A round. These days, an A round is usually about the third time a company has raised money. And the valuation ranges, he breaks them down into percentiles, that is to say clusters, if you will. And the 50th percentile companies are valued at around $40 million. 44, I think it is. And at the 95th percentile, they're valued over $200 million. So that's the spread between the middle and the highest. And he says that's the widest spread ever. He looks back in history five years, that's what he's got data for, and says even in the crazy 2021 period and 2022,
Words and timings
Ithinkheimpliesthatthereiswithoutsayingso.Whathe'sdoingishemeasuresthevaluationofcompaniesattheirAround.AndanAroundusedtobethefirsttimeacompanyraisedtheirmoney.Backintheday,yourfirstfundingwascalledanAround.Thesedays,anAroundisusuallyaboutthethirdtimeacompanyhasraisedmoney.Andthevaluationranges,hebreaksthemdownintopercentiles,thatistosayclusters,ifyouwill.Andthe50thpercentilecompaniesarevaluedataround$40million.44,Ithinkitis.Andatthe95thpercentile,they'revaluedover$200million.Sothat'sthespreadbetweenthemiddleandthehighest.Andhesaysthat'sthewidestspreadever.Helooksbackinhistoryfiveyears,that'swhathe'sgotdatafor,andsayseveninthecrazy2021periodand2022,
Speaker 3
It wasn't as wide as this.
Words and timings
Itwasn'taswideasthis.
Speaker 2
So what number there is significant, the ceiling or the floor? The ceiling. So at the high end, the average Series A is $200 million? Valued. Valuation. And for him, how does that compare historically?
Words and timings
Sowhatnumberthereissignificant,theceilingorthefloor?Theceiling.Soatthehighend,theaverageSeriesAis$200million?Valued.Valuation.Andforhim,howdoesthatcomparehistorically?
Speaker 3
Historically, you can see it in the chart. I don't know if you have it, but in the chart, you'll see there's a range. If you scroll down, you're not showing a browser, are you? So you can't.
Words and timings
Historically,youcanseeitinthechart.Idon'tknowifyouhaveit,butinthechart,you'llseethere'sarange.Ifyouscrolldown,you'renotshowingabrowser,areyou?Soyoucan't.
Speaker 2
No, but it's in the newsletter.
Words and timings
No,butit'sinthenewsletter.
Speaker 3
Yeah, it's in the newsletter. Basically, the range is usually far narrower. Firstly, the gap between 44 and 200 is very big. Normally, it's more likely between 30 and 60, something like that. And why is the gap so wide? Because there's a small number of companies at the high end raising huge sums, and the implication is those are mainly AI companies. And that tells you that insofar as there are high valuations, and there are, the highest ever. So this correction doesn't exist in the 95th percentile.
Words and timings
Yeah,it'sinthenewsletter.Basically,therangeisusuallyfarnarrower.Firstly,thegapbetween44and200isverybig.Normally,it'smorelikelybetween30and60,somethinglikethat.Andwhyisthegapsowide?Becausethere'sasmallnumberofcompaniesatthehighendraisinghugesums,andtheimplicationisthosearemainlyAIcompanies.Andthattellsyouthatinsofarastherearehighvaluations,andthereare,thehighestever.Sothiscorrectiondoesn'texistinthe95thpercentile.
Speaker 2
But also for AI, and we've talked about this before, to be an AI company, especially... one like Perplexity or one of these other more ambitious companies, Anthropic, you have to be valued in the billions to even begin to play in this space. So why does that suggest there's a bubble?
Words and timings
ButalsoforAI,andwe'vetalkedaboutthisbefore,tobeanAIcompany,especially...onelikePerplexityoroneoftheseothermoreambitiouscompanies,Anthropic,youhavetobevaluedinthebillionstoevenbegintoplayinthisspace.Sowhydoesthatsuggestthere'sabubble?
Speaker 3
Well, so let's come back to the bubble question because I don't think there is a bubble in some ways as well. But I do think the facts are important. Remember, these are Series A. So companies like Perplexity and the names that everyone would be familiar with, they're already doing Series B, C, and Ds.
Words and timings
Well,solet'scomebacktothebubblequestionbecauseIdon'tthinkthereisabubbleinsomewaysaswell.ButIdothinkthefactsareimportant.Remember,theseareSeriesA.SocompanieslikePerplexityandthenamesthateveryonewouldbefamiliarwith,they'realreadydoingSeriesB,C,andDs.
Speaker 2
Right.
Words and timings
Right.
Speaker 3
But these are much earlier, these companies. You've probably never heard of most of them yet. These are the companies, some of whom will be big names next year. And so to be valued at 200 million when no one's even ever heard of you is all about the promise of what might be.
Words and timings
Butthesearemuchearlier,thesecompanies.You'veprobablyneverheardofmostofthemyet.Thesearethecompanies,someofwhomwillbebignamesnextyear.Andsotobevaluedat200millionwhennoone'seveneverheardofyouisallaboutthepromiseofwhatmightbe.
Speaker 2
And we know that those are the kind of companies that are founded by entrepreneurs with track records. So it's not just a couple of kids in a garage somewhere who are being valued at 200 million, is it?
Words and timings
Andweknowthatthosearethekindofcompaniesthatarefoundedbyentrepreneurswithtrackrecords.Soit'snotjustacoupleofkidsinagaragesomewherewhoarebeingvaluedat200million,isit?
Speaker 3
There'll be some of those, Andrew. I mean, I'm not so cynical that the value doesn't still back outliers. I think it does. So there will be some of those. But yeah, you're right. A lot of them will be second or third time entrepreneurs as well who are known quantities.
Words and timings
There'llbesomeofthose,Andrew.Imean,I'mnotsocynicalthatthevaluedoesn'tstillbackoutliers.Ithinkitdoes.Sotherewillbesomeofthose.Butyeah,you'reright.Alotofthemwillbesecondorthirdtimeentrepreneursaswellwhoareknownquantities.
Speaker 2
And does this reflect competition amongst VCs? The reason why we have a ceiling at 200 million valuation for Series A is because there's a feeding frenzy amongst VCs?
Words and timings
AnddoesthisreflectcompetitionamongstVCs?Thereasonwhywehaveaceilingat200millionvaluationforSeriesAisbecausethere'safeedingfrenzyamongstVCs?
Speaker 3
You could call it a feeding frenzy, but the normal way to talk about it would be to acknowledge that some of these companies are going to be worth at least billions, if not more. So what they're really doing is competing to earn equity in possible future outbreaks. And there's a price of entry set by everyone. And if you're not prepared to pay the price or don't have the capital to pay the price, you're not going to end up owning that company's equity. So, yes, in a way, you're right. On the other hand, what else could they do in a competitive... One of the things you've always...
Words and timings
Youcouldcallitafeedingfrenzy,butthenormalwaytotalkaboutitwouldbetoacknowledgethatsomeofthesecompaniesaregoingtobeworthatleastbillions,ifnotmore.Sowhatthey'rereallydoingiscompetingtoearnequityinpossiblefutureoutbreaks.Andthere'sapriceofentrysetbyeveryone.Andifyou'renotpreparedtopaythepriceordon'thavethecapitaltopaytheprice,you'renotgoingtoendupowningthatcompany'sequity.So,yes,inaway,you'reright.Ontheotherhand,whatelsecouldtheydoinacompetitive...Oneofthethingsyou'vealways...
Speaker 2
educated me on, Keith, when it comes, because you really have become an expert in this area, is if there's a lot of cash around, the fact that you're valuing Series A's at a high point at $200 million over $150 or $100 million doesn't really matter, does it?
Words and timings
educatedmeon,Keith,whenitcomes,becauseyoureallyhavebecomeanexpertinthisarea,isifthere'salotofcasharound,thefactthatyou'revaluingSeriesA'satahighpointat$200millionover$150or$100milliondoesn'treallymatter,doesit?
Speaker 3
What matters more is what percentage of a company you are. It's a lot better to earn 20% than 10%, and it's a lot better to earn 10% than 2%. So Y Combinator, for example, always owns 7% initially, which is a small number. So some of the more aggressive early stage investors like to target 15% to 20% ownership. So the minute you're raising $200 million, you're almost a unicorn right there, by the way. If you're raising $200 million, you're probably a unicorn.
Words and timings
Whatmattersmoreiswhatpercentageofacompanyyouare.It'salotbettertoearn20%than10%,andit'salotbettertoearn10%than2%.SoYCombinator,forexample,alwaysowns7%initially,whichisasmallnumber.Sosomeofthemoreaggressiveearlystageinvestorsliketotarget15%to20%ownership.Sotheminuteyou'reraising$200million,you'realmostaunicornrightthere,bytheway.Ifyou'reraising$200million,you'reprobablyaunicorn.
Speaker 2
A unicorn on paper, that's... We've had this conversation.
Words and timings
Aunicornonpaper,that's...We'vehadthisconversation.
Speaker 3
All the unicorns are on paper. There is no such thing as a unicorn not on paper.
Words and timings
Alltheunicornsareonpaper.Thereisnosuchthingasaunicornnotonpaper.
Speaker 2
So OpenAI isn't a unicorn as well, but we know it's more than on paper. It actually is worth more than a billion dollars, even if they haven't gone public.
Words and timings
SoOpenAIisn'taunicornaswell,butweknowit'smorethanonpaper.Itactuallyisworthmorethanabilliondollars,eveniftheyhaven'tgonepublic.
Speaker 3
They've got enough revenue to suggest that they're worth 20 to 30 billion.
Words and timings
They'vegotenoughrevenuetosuggestthatthey'reworth20to30billion.
Speaker 2
At least you've suggested more so. There are other articles along with the Peter Walker piece. MG Siegler made his name at TechCrunch. He was hired by Mike Harrington. You note that in your editorial. He writes about the Surface Laptop Review, Microsoft's new product. What's this got to do with AI and the bubble?
Words and timings
Atleastyou'vesuggestedmoreso.ThereareotherarticlesalongwiththePeterWalkerpiece.MGSieglermadehisnameatTechCrunch.HewashiredbyMikeHarrington.Younotethatinyoureditorial.HewritesabouttheSurfaceLaptopReview,Microsoft'snewproduct.What'sthisgottodowithAIandthebubble?
Speaker 3
Well, the best quote in this piece is Microsoft shit their pants in bed or something like that.
Words and timings
Well,thebestquoteinthispieceisMicrosoftshittheirpantsinbedorsomethinglikethat.
Speaker 2
This is a family show, Keith. Are we allowed to say that?
Words and timings
Thisisafamilyshow,Keith.Areweallowedtosaythat?
Speaker 3
He said Microsoft shit the bed. And that was correlated to the AI component of this new Surface Pro. that just to give people who don't know context the surface pro is that microsoft laptop with a kind of a cloth keyboard that falls up or you can take it off and use it as a tablet as well it looks very nice but it has been a underperformer in terms of compared to apple put it politely this new one apparently gets close to being as good as a an apple m1 And it does it using a new chip from Qualcomm called Snapdragon. I forget the number, but it's the most recent in the Snapdragon family. And it's got a separate processor for neural networks, just like Apple does, and so on. So the reputation of this new one is it's caught up with Apple. And then what Ziegler does, he explains how it really hasn't, because Microsoft don't have the confidence to ship the AI software that would make it interesting, which is called co-pilot. Not as Apple, really, does it? They don't either. And in fact, both of them are running scared of regulation, either in Europe or here.
Words and timings
HesaidMicrosoftshitthebed.AndthatwascorrelatedtotheAIcomponentofthisnewSurfacePro.thatjusttogivepeoplewhodon'tknowcontextthesurfaceproisthatmicrosoftlaptopwithakindofaclothkeyboardthatfallsuporyoucantakeitoffanduseitasatabletaswellitlooksverynicebutithasbeenaunderperformerintermsofcomparedtoappleputitpolitelythisnewoneapparentlygetsclosetobeingasgoodasaanapplem1AnditdoesitusinganewchipfromQualcommcalledSnapdragon.Iforgetthenumber,butit'sthemostrecentintheSnapdragonfamily.Andit'sgotaseparateprocessorforneuralnetworks,justlikeAppledoes,andsoon.Sothereputationofthisnewoneisit'scaughtupwithApple.AndthenwhatZieglerdoes,heexplainshowitreallyhasn't,becauseMicrosoftdon'thavetheconfidencetoshiptheAIsoftwarethatwouldmakeitinteresting,whichiscalledco-pilot.NotasApple,really,doesit?Theydon'teither.Andinfact,bothofthemarerunningscaredofregulation,eitherinEuropeorhere.
Speaker 2
Understandably, given how vulnerable they are.
Words and timings
Understandably,givenhowvulnerabletheyare.
Speaker 3
Right. So now you know for sure people who are scared of regulation aren't going to be the ones who change the world. In a way, it's good for OpenAI that these two are so sheepish because in order to make things happen, you really have to behave like Uber did in the early days.
Words and timings
Right.Sonowyouknowforsurepeoplewhoarescaredofregulationaren'tgoingtobetheoneswhochangetheworld.Inaway,it'sgoodforOpenAIthatthesetwoaresosheepishbecauseinordertomakethingshappen,youreallyhavetobehavelikeUberdidintheearlydays.
Speaker 2
What about that link, Keith, I sent you earlier this week from Mustafa? Is it... the new head of ai at microsoft who suggests that anything on the web is free if it's on the web it can be included in in ai isn't that that that doesn't sound very
Words and timings
Whataboutthatlink,Keith,IsentyouearlierthisweekfromMustafa?Isit...thenewheadofaiatmicrosoftwhosuggeststhatanythingonthewebisfreeifit'sonthewebitcanbeincludedininaiisn'tthatthatthatdoesn'tsoundvery
Speaker 3
sheepish yeah well he's not you know if microsoft is the borg uh mustafa is still 90 in the starship enterprise He hasn't fully been incorporated yet into the Microsoft way of thinking. So he's a bit of an outlier, but I will guarantee his bosses won't like the fact that he said...
Words and timings
sheepishyeahwellhe'snotyouknowifmicrosoftistheborguhmustafaisstill90inthestarshipenterpriseHehasn'tfullybeenincorporatedyetintotheMicrosoftwayofthinking.Sohe'sabitofanoutlier,butIwillguaranteehisbosseswon'tlikethefactthathesaid...
Speaker 2
His boss is the CEO.
Words and timings
HisbossistheCEO.
Speaker 3
Yeah, and his boss won't like the fact that he said that because...
Words and timings
Yeah,andhisbosswon'tlikethefactthathesaidthatbecause...
Speaker 2
So basically what Ziegler is saying is that for all this talk of AI and all these new pieces of hardware, it's all, again, if not a... A bubble, there's just nothing there. The emperor has no clothes.
Words and timings
SobasicallywhatZieglerissayingisthatforallthistalkofAIandallthesenewpiecesofhardware,it'sall,again,ifnota...Abubble,there'sjustnothingthere.Theemperorhasnoclothes.
Speaker 3
Well, no, there is a bubble in early stage funding, although I always make the point, and I think you agree with this, that a bubble is a natural reaction to opportunity.
Words and timings
Well,no,thereisabubbleinearlystagefunding,althoughIalwaysmakethepoint,andIthinkyouagreewiththis,thatabubbleisanaturalreactiontoopportunity.
Speaker 2
It's inevitable. I mean, it's unavoidable. And lots of economists suggest that the real foundations of technological innovation lies in bubbles because when they pop, some companies survive. And that's when you get the real interesting building. The real issue is, are we sick of AI? Since this isn't a family show, since you've been swearing, Keith. There's another piece I'm quoting. It's not my words. I will fucking pile drive you if you mention AI again. Do we have to keep on hearing about AI? What does this piece suggest?
Words and timings
It'sinevitable.Imean,it'sunavoidable.Andlotsofeconomistssuggestthattherealfoundationsoftechnologicalinnovationliesinbubblesbecausewhentheypop,somecompaniessurvive.Andthat'swhenyougettherealinterestingbuilding.Therealissueis,arewesickofAI?Sincethisisn'tafamilyshow,sinceyou'vebeenswearing,Keith.There'sanotherpieceI'mquoting.It'snotmywords.IwillfuckingpiledriveyouifyoumentionAIagain.DowehavetokeeponhearingaboutAI?Whatdoesthispiecesuggest?
Speaker 3
So this is an Australian techie who works in AI sounding off from the point of view of a techie. about how when uneducated people who don't really know what they're talking about keep proselytizing AI without even knowing what it is. And a pile drive, for those who don't know, is when you turn someone upside down in a wrestling match and bang their head on the floor.
Words and timings
SothisisanAustraliantechiewhoworksinAIsoundingofffromthepointofviewofatechie.abouthowwhenuneducatedpeoplewhodon'treallyknowwhatthey'retalkingaboutkeepproselytizingAIwithoutevenknowingwhatitis.Andapiledrive,forthosewhodon'tknow,iswhenyouturnsomeoneupsidedowninawrestlingmatchandbangtheirheadonthefloor.
Speaker 2
So it's not a nice... That seems to be the normal behavior in Australia, isn't it? After a few fosters, maybe. Maybe.
Words and timings
Soit'snotanice...ThatseemstobethenormalbehaviorinAustralia,isn'tit?Afterafewfosters,maybe.Maybe.
Speaker 3
So the article is... saying, please shut up. It's not a very good article, by the way. I put it in because it got a lot of attention and was reposted and discussed a lot. So I put it in just so my readers are aware that it exists. It's not really very intelligent. It's just sounding off.
Words and timings
Sothearticleis...saying,pleaseshutup.It'snotaverygoodarticle,bytheway.Iputitinbecauseitgotalotofattentionandwasrepostedanddiscussedalot.SoIputitinjustsomyreadersareawarethatitexists.It'snotreallyveryintelligent.It'sjustsoundingoff.
Speaker 2
It is a bit boring. But on the other hand, these questions keep on coming up. There's one piece you linked to from Wired, which is a credible magazine that Perplexity is a bullshit machine. You've always spoken up perplexity. It's the most credible AI search engine. Is perplexity just scraping the web? Is it a bullshit machine?
Words and timings
Itisabitboring.Butontheotherhand,thesequestionskeeponcomingup.There'sonepieceyoulinkedtofromWired,whichisacrediblemagazinethatPerplexityisabullshitmachine.You'vealwaysspokenupperplexity.It'sthemostcredibleAIsearchengine.Isperplexityjustscrapingtheweb?Isitabullshitmachine?
Speaker 3
You know, this goes along with a second article about is there such a thing as an AI search engine? I can't remember the exact title.
Words and timings
Youknow,thisgoesalongwithasecondarticleaboutistheresuchathingasanAIsearchengine?Ican'tremembertheexacttitle.
Speaker 2
Yeah, what if anything is AI search good for by Adam Clark Estes in Vox? I tend to agree. I mean, search for me is quite adequate. I don't need AI. And anyway, even Google AI, even Google search is built on a kind of AI. AI...
Words and timings
Yeah,whatifanythingisAIsearchgoodforbyAdamClarkEstesinVox?Itendtoagree.Imean,searchformeisquiteadequate.Idon'tneedAI.Andanyway,evenGoogleAI,evenGooglesearchisbuiltonakindofAI.AI...
Speaker 3
is this word we use for large language models. It's kind of almost become reduced down to large language models. And large language models are really no good for search. They're not. They're pretty good for knowledge, as long as you're aware that they sometimes hallucinate and you're critical enough to read what's said and think about it. They're really, really good at knowledge. They're actually good at stopping you searching. So if you try to apply a large damage model to search, it probably isn't going to work. And the Wired story is exposing the fact that perplexity actually isn't really applying AI to search. What it's actually doing, in the words of Wired, is plagiarizing articles.
Words and timings
isthiswordweuseforlargelanguagemodels.It'skindofalmostbecomereduceddowntolargelanguagemodels.Andlargelanguagemodelsarereallynogoodforsearch.They'renot.They'reprettygoodforknowledge,aslongasyou'reawarethattheysometimeshallucinateandyou'recriticalenoughtoreadwhat'ssaidandthinkaboutit.They'rereally,reallygoodatknowledge.They'reactuallygoodatstoppingyousearching.Soifyoutrytoapplyalargedamagemodeltosearch,itprobablyisn'tgoingtowork.AndtheWiredstoryisexposingthefactthatperplexityactuallyisn'treallyapplyingAItosearch.Whatit'sactuallydoing,inthewordsofWired,isplagiarizingarticles.
Speaker 2
Yeah, I quote Wired surreptitiously, or maybe not so surreptitiously, scraping, stealing content.
Words and timings
Yeah,IquoteWiredsurreptitiously,ormaybenotsosurreptitiously,scraping,stealingcontent.
Speaker 3
Scraping and stealing. Now... You know, you need to have a nuanced discussion about this because Google's index is built from crawling the web and indexing every single word. So that is a form of scraping and it indexes the scraping. Are you suggesting that Google is a big thief, Keith? I'm going to report you to my wife. Well, Google is both a thief and a utility at the same time because it has enough users that the traffic it sends back to you is worth giving it access to your content.
Words and timings
Scrapingandstealing.Now...Youknow,youneedtohaveanuanceddiscussionaboutthisbecauseGoogle'sindexisbuiltfromcrawlingthewebandindexingeverysingleword.Sothatisaformofscrapinganditindexesthescraping.AreyousuggestingthatGoogleisabigthief,Keith?I'mgoingtoreportyoutomywife.Well,Googleisbothathiefandautilityatthesametimebecauseithasenoughusersthatthetrafficitsendsbacktoyouisworthgivingitaccesstoyourcontent.
Speaker 2
Google is much less vulnerable in court than perplexity, which is clearly... stealing content to add to its... I don't know.
Words and timings
Googleismuchlessvulnerableincourtthanperplexity,whichisclearly...stealingcontenttoaddtoits...Idon'tknow.
Speaker 3
That goes back to Mustafa's point, which is, is it really stealing if what it's actually doing is indexing and then feeding it back in a search context? By the way, and linking to you. Is there really any difference between Bing Google and Perplexity. Perplexity has got a nice user interface. It does actually plug in ChatGPT and other ones into the interface, so there's more choice into how to deal with the results. But at the level of how does it get its data and what does it do with it, surely plagiarizing is better than obfuscating.
Words and timings
ThatgoesbacktoMustafa'spoint,whichis,isitreallystealingifwhatit'sactuallydoingisindexingandthenfeedingitbackinasearchcontext?Bytheway,andlinkingtoyou.IstherereallyanydifferencebetweenBingGoogleandPerplexity.Perplexityhasgotaniceuserinterface.ItdoesactuallypluginChatGPTandotheronesintotheinterface,sothere'smorechoiceintohowtodealwiththeresults.Butatthelevelofhowdoesitgetitsdataandwhatdoesitdowithit,surelyplagiarizingisbetterthanobfuscating.
Speaker 2
So there's two issues here. On the one hand, you said earlier that that the most innovative companies are pirates and that they break things and break the law. So in a sense, you're suggesting that perplexity is more innovative, certainly than Microsoft or Apple. On the other hand, you're suggesting that perplexity's attempt to turn AI into search is rather like the pioneers of television who tried to put radio onto television. So which is true? Are they both simultaneously true about perplexity?
Words and timings
Sothere'stwoissueshere.Ontheonehand,yousaidearlierthatthatthemostinnovativecompaniesarepiratesandthattheybreakthingsandbreakthelaw.Soinasense,you'resuggestingthatperplexityismoreinnovative,certainlythanMicrosoftorApple.Ontheotherhand,you'resuggestingthatperplexity'sattempttoturnAIintosearchisratherlikethepioneersoftelevisionwhotriedtoputradioontotelevision.Sowhichistrue?Aretheybothsimultaneouslytrueaboutperplexity?
Speaker 3
I actually don't really rate perplexity in the long run. I think it's opened people's eyes to some possibilities. And I do expect that some of the things you see on perplexity will survive as user interface.
Words and timings
Iactuallydon'treallyrateperplexityinthelongrun.Ithinkit'sopenedpeople'seyestosomepossibilities.AndIdoexpectthatsomeofthethingsyouseeonperplexitywillsurviveasuserinterface.
Speaker 2
You don't use perplexity as search. Rarely. You and I have talked about this a lot. Do you basically use AI now for programming? Is that the core value for you?
Words and timings
Youdon'tuseperplexityassearch.Rarely.YouandIhavetalkedaboutthisalot.DoyoubasicallyuseAInowforprogramming?Isthatthecorevalueforyou?
Speaker 3
You know, yes. For those who are listening and want to, if you go to search.signalrank.ai, you'll find a website that's a search engine. And I built it this week using ChatGPT and Claude.
Words and timings
Youknow,yes.Forthosewhoarelisteningandwantto,ifyougotosearch.signalrank.ai,you'llfindawebsitethat'sasearchengine.AndIbuiltitthisweekusingChatGPTandClaude.
Speaker 2
Okay, so it has some value maybe in that sense. I can't comment. I'm not a programmer. But elsewhere, I don't personally see any evidence that this thing has value. very much value we'll come to it at the end where uh oh malik is using it to to write and take photographs but i i have to admit i'm not convinced meanwhile keith let's talk about some other stuff um it's not just ai this week um interesting piece by alex wilhelm you've referenced him linked with him before Why the IPO I'm most excited about isn't Databricks, who we've talked about, or Stripe, we've talked about them many times. But Circle, we haven't talked about Circle much. Do you share Wilhelm's optimism about Circle?
Words and timings
Okay,soithassomevaluemaybeinthatsense.Ican'tcomment.I'mnotaprogrammer.Butelsewhere,Idon'tpersonallyseeanyevidencethatthisthinghasvalue.verymuchvaluewe'llcometoitattheendwhereuhohmalikisusingittotowriteandtakephotographsbutiihavetoadmiti'mnotconvincedmeanwhilekeithlet'stalkaboutsomeotherstuffumit'snotjustaithisweekuminterestingpiecebyalexwilhelmyou'vereferencedhimlinkedwithhimbeforeWhytheIPOI'mmostexcitedaboutisn'tDatabricks,whowe'vetalkedabout,orStripe,we'vetalkedaboutthemmanytimes.ButCircle,wehaven'ttalkedaboutCirclemuch.DoyoushareWilhelm'soptimismaboutCircle?
Speaker 3
I know why he likes them. Circle is run by a gentleman called Jeremy Allaire.
Words and timings
Iknowwhyhelikesthem.CircleisrunbyagentlemancalledJeremyAllaire.
Speaker 2
Yeah, he's been around a while. He's been around the block many times.
Words and timings
Yeah,he'sbeenaroundawhile.He'sbeenaroundtheblockmanytimes.
Speaker 3
He was the original creator of Macromedia. Macromedia director, which was the first kind of animated web and led to Flash, for example. So, yeah, he's been around a while. He then did a video startup and Circle is the company that brought what's called USDC into the world. USDC is a stable coin in the crypto world that is backed by actual assets, billions of them, and makes interest on those assets, a lot of interest. And so has a very big revenue stream simply by being, if you like, the central bank that everyone can cash into and out of when they buy crypto.
Words and timings
HewastheoriginalcreatorofMacromedia.Macromediadirector,whichwasthefirstkindofanimatedwebandledtoFlash,forexample.So,yeah,he'sbeenaroundawhile.HethendidavideostartupandCircleisthecompanythatbroughtwhat'scalledUSDCintotheworld.USDCisastablecoininthecryptoworldthatisbackedbyactualassets,billionsofthem,andmakesinterestonthoseassets,alotofinterest.Andsohasaverybigrevenuestreamsimplybybeing,ifyoulike,thecentralbankthateveryonecancashintoandoutofwhentheybuycrypto.
Speaker 2
And they filed privately to go public earlier this year. Does that suggest that the Web3 crypto economy is still alive?
Words and timings
Andtheyfiledprivatelytogopublicearlierthisyear.DoesthatsuggestthattheWeb3cryptoeconomyisstillalive?
Speaker 3
Oh, it's very much alive. It's a parallel economy. It hasn't really... broken into the mainstream yet, except as an investment asset. So you can now buy Bitcoin and pretty soon Ethereum on the stock exchange with the approval of the SEC through exchange traded funds and the like. So it's broken through as an investment asset, but Hasn't really broken through product-wise. And I think the jury's out on whether it will. You need some super smart product people to build things that people will use.
Words and timings
Oh,it'sverymuchalive.It'saparalleleconomy.Ithasn'treally...brokenintothemainstreamyet,exceptasaninvestmentasset.SoyoucannowbuyBitcoinandprettysoonEthereumonthestockexchangewiththeapprovaloftheSECthroughexchangetradedfundsandthelike.Soit'sbrokenthroughasaninvestmentasset,butHasn'treallybrokenthroughproduct-wise.AndIthinkthejury'soutonwhetheritwill.Youneedsomesupersmartproductpeopletobuildthingsthatpeoplewilluse.
Speaker 2
So go back to Walker's piece about IPR, about Series A. Did he break down crypto versus AI? I mean, is the smart money perhaps now going back into crypto?
Words and timings
SogobacktoWalker'spieceaboutIPR,aboutSeriesA.DidhebreakdowncryptoversusAI?Imean,isthesmartmoneyperhapsnowgoingbackintocrypto?
Speaker 3
Some of it is. He includes crypto in SaaS, and this article is all about SaaS. So it is in there, but he didn't break it out. If you go and look at his LinkedIn, he does actually break out the different sectors in some of his earlier posts. So you can go and find it there. But crypto is the poor cousin of AI right now. AI is attracting most of the capital.
Words and timings
Someofitis.HeincludescryptoinSaaS,andthisarticleisallaboutSaaS.Soitisinthere,buthedidn'tbreakitout.IfyougoandlookathisLinkedIn,hedoesactuallybreakoutthedifferentsectorsinsomeofhisearlierposts.Soyoucangoandfinditthere.ButcryptoisthepoorcousinofAIrightnow.AIisattractingmostofthecapital.
Speaker 2
Well, speaking of capital and valuations. Your startup of the week is Webtoon. It rose modestly in its IPO debut. Tell us about Webtoon. Is this something we should be excited about?
Words and timings
Well,speakingofcapitalandvaluations.YourstartupoftheweekisWebtoon.ItrosemodestlyinitsIPOdebut.TellusaboutWebtoon.Isthissomethingweshouldbeexcitedabout?
Speaker 3
Well, Webtoon is like Substack, but for cartoonists. And it's got a long history, by the way. It was founded in Korea in uh... way back in its uh... seventy percent of it is owned by uh... now which is its korean parent uh... company and uh... of course asia and manga go together very closely in korea is a country where uh... cartoon as an art form is is culturally important so it's not surprising that it came out of korea it could have come out of japan just as easily or even China. It's Korean and it is the world's marketplace for anyone who is a cartoonist who wants to sell their cartoons or have people look at them and earn money from them being looked at. So it's a company that could really only exist in the age of the Internet. Its revenues are very high, around about $300 million to $400 million a year. So it's not a small company.
Words and timings
Well,WebtoonislikeSubstack,butforcartoonists.Andit'sgotalonghistory,bytheway.ItwasfoundedinKoreainuh...waybackinitsuh...seventypercentofitisownedbyuh...nowwhichisitskoreanparentuh...companyanduh...ofcourseasiaandmangagotogetherverycloselyinkoreaisacountrywhereuh...cartoonasanartformisisculturallyimportantsoit'snotsurprisingthatitcameoutofkoreaitcouldhavecomeoutofjapanjustaseasilyorevenChina.It'sKoreananditistheworld'smarketplaceforanyonewhoisacartoonistwhowantstoselltheircartoonsorhavepeoplelookatthemandearnmoneyfromthembeinglookedat.Soit'sacompanythatcouldreallyonlyexistintheageoftheInternet.Itsrevenuesareveryhigh,aroundabout$300millionto$400millionayear.Soit'snotasmallcompany.
Speaker 2
It sounds almost like a Web1 company.
Words and timings
ItsoundsalmostlikeaWeb1company.
Speaker 3
It's close to a YouTube for cartoons, yeah.
Words and timings
It'sclosetoaYouTubeforcartoons,yeah.
Speaker 2
And refreshingly, there's no AI there and there's no crypto. This is a real company, real product, real demand.
Words and timings
Andrefreshingly,there'snoAIthereandthere'snocrypto.Thisisarealcompany,realproduct,realdemand.
Speaker 3
Yeah, it's a marketplace, which is one of the categories of the internet that does very well.
Words and timings
Yeah,it'samarketplace,whichisoneofthecategoriesoftheinternetthatdoesverywell.
Speaker 2
So we should celebrate this? I mean, no one's going to get seriously rich with Webtoon, are they?
Words and timings
Soweshouldcelebratethis?Imean,noone'sgoingtogetseriouslyrichwithWebtoon,arethey?
Speaker 3
Well, I think NAVA, the parent company, is, but they're already seriously rich, so we can't get too excited about that. So why'd you choose it as your startup of the week? I think it's got a story that shows you that good companies can still go public. They don't have to be the coolest thing on the block or new.
Words and timings
Well,IthinkNAVA,theparentcompany,is,butthey'realreadyseriouslyrich,sowecan'tgettooexcitedaboutthat.Sowhy'dyouchooseitasyourstartupoftheweek?Ithinkit'sgotastorythatshowsyouthatgoodcompaniescanstillgopublic.Theydon'thavetobethecoolestthingontheblockornew.
Speaker 2
Well, finally, the X of the week by one of your favorite tech commentators, a man who needs no introduction in our show, Oumalek, at Oum. And we have a photo of a camera, odd-looking camera. He writes, serendipity is what makes San Francisco special. Technologies, new ideas, and people collide randomly in the oddest of places at oddest of times. I was reminded of this when walking in the park, playing with poetry camera. What's poetry camera? Is this the device that we see?
Words and timings
Well,finally,theXoftheweekbyoneofyourfavoritetechcommentators,amanwhoneedsnointroductioninourshow,Oumalek,atOum.Andwehaveaphotoofacamera,odd-lookingcamera.Hewrites,serendipityiswhatmakesSanFranciscospecial.Technologies,newideas,andpeoplecolliderandomlyintheoddestofplacesatoddestoftimes.Iwasremindedofthiswhenwalkinginthepark,playingwithpoetrycamera.What'spoetrycamera?Isthisthedevicethatwesee?
Speaker 3
Yeah, so this is a 3D printed handmade device. The person who did it is a friend of Oam's. And what this guy has done is he's given the camera the ability to write a poem based on what it sees through the lens. And so it doesn't actually take pictures. It writes poems from what it sees. So hence the poetry camera. And it prints them out. It's got a little printer inside like a till receipt. So it's completely weird. But I thought it was really interesting just because the imagination puts things together that no product team would ever come up with.
Words and timings
Yeah,sothisisa3Dprintedhandmadedevice.ThepersonwhodiditisafriendofOam's.Andwhatthisguyhasdoneishe'sgiventhecameratheabilitytowriteapoembasedonwhatitseesthroughthelens.Andsoitdoesn'tactuallytakepictures.Itwritespoemsfromwhatitsees.Sohencethepoetrycamera.Anditprintsthemout.It'sgotalittleprinterinsidelikeatillreceipt.Soit'scompletelyweird.ButIthoughtitwasreallyinterestingjustbecausetheimaginationputsthingstogetherthatnoproductteamwouldevercomeupwith.
Speaker 2
Although Canon's been talking and the other large camera manufacturers have been promising, threatening to put AI into their products. Is this the kind of way in which AI, and we've used that word carefully, could be included in hardware from Canon or Sony?
Words and timings
AlthoughCanon'sbeentalkingandtheotherlargecameramanufacturershavebeenpromising,threateningtoputAIintotheirproducts.IsthisthekindofwayinwhichAI,andwe'veusedthatwordcarefully,couldbeincludedinhardwarefromCanonorSony?
Speaker 3
Well, they already are included in things like this camera I'm using now. It's got a little square that I can see that's focused on my eyes. And it's using AI to make sure I keep in focus because it knows where my eyes are. And if I duck down, it'll focus on the background. And if I stand up again, it'll focus on me. And that's AI. But this is a different kind of AI. This is more like artistic or creative AI in the world, in the physical world. Imagine if that camera had legs and arms and could move around and had a microphone and a speaker and it was a robot speaking the poems, which is not very far-fetched. There's an article in this week's newsletter about robotics investments, which are on the up and up. We're approaching the age of robotics very rapidly now.
Words and timings
Well,theyalreadyareincludedinthingslikethiscameraI'musingnow.It'sgotalittlesquarethatIcanseethat'sfocusedonmyeyes.Andit'susingAItomakesureIkeepinfocusbecauseitknowswheremyeyesare.AndifIduckdown,it'llfocusonthebackground.AndifIstandupagain,it'llfocusonme.Andthat'sAI.ButthisisadifferentkindofAI.ThisismorelikeartisticorcreativeAIintheworld,inthephysicalworld.Imagineifthatcamerahadlegsandarmsandcouldmovearoundandhadamicrophoneandaspeakeranditwasarobotspeakingthepoems,whichisnotveryfar-fetched.There'sanarticleinthisweek'snewsletteraboutroboticsinvestments,whichareontheupandup.We'reapproachingtheageofroboticsveryrapidlynow.
Speaker 2
This could go the other way. You could have products that take text and turn them into photographs.
Words and timings
Thiscouldgotheotherway.Youcouldhaveproductsthattaketextandturnthemintophotographs.
Speaker 3
Well, yeah, as you've noted, Andrew, the results are...
Words and timings
Well,yeah,asyou'venoted,Andrew,theresultsare...
Speaker 2
Well, I think that's probably a bad reflection on you. I can't believe that AI art is always quite as bad as yours, but that...
Words and timings
Well,Ithinkthat'sprobablyabadreflectiononyou.Ican'tbelievethatAIartisalwaysquiteasbadasyours,butthat...
Speaker 3
I thought this one was quite good. I mean, the fact that it spelled the words correctly.
Words and timings
Ithoughtthisonewasquitegood.Imean,thefactthatitspelledthewordscorrectly.
Speaker 2
It didn't spell the words. Natural networks? What's a natural network?
Words and timings
Itdidn'tspellthewords.Naturalnetworks?What'sanaturalnetwork?
Speaker 3
You can't see that one, but is there an AI bubble is properly rendered.
Words and timings
Youcan'tseethatone,butisthereanAIbubbleisproperlyrendered.
Speaker 2
Well, that's quite an achievement. Is there an AI bubble? Remarkable achievement.
Words and timings
Well,that'squiteanachievement.IsthereanAIbubble?Remarkableachievement.
Speaker 3
Well, any of you who use these things know that the text is almost always garbled.
Words and timings
Well,anyofyouwhousethesethingsknowthatthetextisalmostalwaysgarbled.
Speaker 2
There is a bubble, Keith. The more I talk to you, the more I think there's a bubble. What does Ohm think? He writes about, as a writer, borrowing from AI to create poetry. Does he find it useful? He's hard-headed, somewhat sceptical.
Words and timings
Thereisabubble,Keith.ThemoreItalktoyou,themoreIthinkthere'sabubble.WhatdoesOhmthink?Hewritesabout,asawriter,borrowingfromAItocreatepoetry.Doeshefindituseful?He'shard-headed,somewhatsceptical.
Speaker 3
I think the answer is both. The subtitle of this week's newsletter is Yes and No. It says answer, colon, yes and no. So there is a bubble, but also there isn't insofar as saying there's a bubble in Silicon Valley is like saying there's an opportunity. it's not a negative statement. There's no pejorative implied.
Words and timings
Ithinktheanswerisboth.Thesubtitleofthisweek'snewsletterisYesandNo.Itsaysanswer,colon,yesandno.Sothereisabubble,butalsothereisn'tinsofarassayingthere'sabubbleinSiliconValleyislikesayingthere'sanopportunity.it'snotanegativestatement.There'snopejorativeimplied.
Speaker 2
I mean, the reality is in Silicon Valley, there's always a bubble. And when there isn't one, then it's no longer Silicon Valley.
Words and timings
Imean,therealityisinSiliconValley,there'salwaysabubble.Andwhenthereisn'tone,thenit'snolongerSiliconValley.
Speaker 3
Exactly. It means nothing interesting is happening. So there you go. And I think the people who read the newsletter know that. So there's no news there. But sometimes you have to hold the line in the face of uh... you know that being a manager to do this in the last uh... england played three games in the euros and everyone's going to manage to be fired and he's been saying i don't care and today in the ninety fourth minute He went from the last 16 to the quarterfinals with an overhead kick and a Harry Kane header. And funnily enough, the commentators before the goal were saying, this manager sucks. Why doesn't he put on his substitutes? As soon as the goal was scored, the commentator said, he's a genius. He still has three substitutes left. So now you know that you've got to stand firm in the face of of the BS, even if people threaten to pile drive you.
Words and timings
Exactly.Itmeansnothinginterestingishappening.Sothereyougo.AndIthinkthepeoplewhoreadthenewsletterknowthat.Sothere'snonewsthere.Butsometimesyouhavetoholdthelineinthefaceofuh...youknowthatbeingamanagertodothisinthelastuh...englandplayedthreegamesintheeurosandeveryone'sgoingtomanagetobefiredandhe'sbeensayingidon'tcareandtodayintheninetyfourthminuteHewentfromthelast16tothequarterfinalswithanoverheadkickandaHarryKaneheader.Andfunnilyenough,thecommentatorsbeforethegoalweresaying,thismanagersucks.Whydoesn'theputonhissubstitutes?Assoonasthegoalwasscored,thecommentatorsaid,he'sagenius.Hestillhasthreesubstitutesleft.Sonowyouknowthatyou'vegottostandfirminthefaceofoftheBS,evenifpeoplethreatentopiledriveyou.
Speaker 2
Yeah, which is an unpleasant Australian term. Finally, Keith, to go back to the beginning, we'll talk again next week. What is more likely, England win the Euros or Joe Biden decides not to run?
Words and timings
Yeah,whichisanunpleasantAustralianterm.Finally,Keith,togobacktothebeginning,we'lltalkagainnextweek.Whatismorelikely,EnglandwintheEurosorJoeBidendecidesnottorun?
Speaker 3
Ooh, I think England winning the Euros is more likely. I don't believe Joe Biden has it in him to decide not to run. And I don't believe the Democratic Party has it in them to persuade him not to run. But he shouldn't run.
Words and timings
Ooh,IthinkEnglandwinningtheEurosismorelikely.Idon'tbelieveJoeBidenhasitinhimtodecidenottorun.AndIdon'tbelievetheDemocraticPartyhasitinthemtopersuadehimnottorun.Butheshouldn'trun.
Speaker 1
That was a week No time to be meek The goal is to seek The next big thing
Words and timings
ThatwasaweekNotimetobemeekThegoalistoseekThenextbigthing
Speaker 1
AD.
Words and timings
AD.